Citation
An Examination of Strategic Group Membership and Technology in the Nursing Home Industry

Material Information

Title:
An Examination of Strategic Group Membership and Technology in the Nursing Home Industry
Creator:
Laberge, Alexandre
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (130 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Health Services Research
Health Services Research, Management, and Policy
Committee Chair:
Weech-Maldonado, Robert
Committee Members:
Harman, Jeffrey S.
Yarbrough, Amy K.
Horgas, Ann L.
Hyer, Kathy
Graduation Date:
5/2/2009

Subjects

Subjects / Keywords:
Focus groups ( jstor )
Gene therapy ( jstor )
Health care costs ( jstor )
Medicaid ( jstor )
Medicare ( jstor )
Nursing ( jstor )
Nursing homes ( jstor )
Pressure ulcer ( jstor )
Rehabilitation facilities ( jstor )
Staffing ( jstor )
Health Services Research, Management, and Policy -- Dissertations, Academic -- UF
cost, financial, nursing, porter, quality, strategic, technology
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Health Services Research thesis, Ph.D.

Notes

Abstract:
Purpose: The purpose of the study was to examine if strategic groups in the nursing home industry can be determined by examining how facilities apply their technology and how facilities commit their resources. The study also determined if these groups can be defined using Porter's generic strategies and whether the different groups have varying strategic-performance relationships. Methodology: The study began by performing a factor analysis using scope (technology) and resource commitment variables from the Minimum Data Set (MDS) and the Online Survey Certification and Reporting (OSCAR) system. The newly defined variables were used in a cluster analysis to identify the different clusters. The clusters were classified using Porter?s generic strategy model. Once the strategic groups were identified, the study analyzed the strategic-performance relationships using negative binomial, ordinary least squares regression (OLS), generalized linear model (GLM) with gamma distribution and ordered logit analytical tools. Quality dependent variables included quality of care deficiencies, long term care (LTC) pressure sore prevalence and post acute pressure ulcer incidence, LTC activities of daily living (ADL) decline, post acute walking improvement, LTC bowel and bladder decline and LTC restraint. Financial measures included cost per resident day, operating margin and total margin. Results: The study identified four strategic groups (differentiator, cost leader, rehab focus, and specialty care focus) and a lack of strategy group. Differentiators, rehab focus and specialty care focus groups had less quality of care deficiencies than the lack of strategy group and cost leaders. Differentiators generally had better long term care quality outcomes compared to other strategic groups and the lack of strategy group. Cost leaders, differentiators and lack of strategy groups had lower cost than the rehab focus and specialty care focus groups. Cost leaders, differentiators and rehab focus group had better financial performance than the specialty care focus group and lack of strategy group. Conclusions: The study indicates that using technology and resources commitments is an effective way of identifying strategies groups. The study also found that some strategic groups have better financial performance than other strategic groups or than nursing homes that have no strategy at all. Nursing homes that provide a high level of technology had better quality, without increased cost or sacrificing financial performance. The study also suggests that nursing homes that service residents with high cost conditions may not be adequately reimbursed. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2009.
Local:
Adviser: Weech-Maldonado, Robert.
Statement of Responsibility:
by Alexandre Laberge.

Record Information

Source Institution:
UFRGP
Rights Management:
Copyright Laberge, Alexandre. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Classification:
LD1780 2009 ( lcc )

Downloads

This item has the following downloads:


Full Text





LIST OF TABLES


Table page

4-1 Internal consistency of the composite variables ..................................... ....... ........... 76

4-2 Individual variables used in cluster analysis that were not included in a factor
v a riab le ................... .......... ............................................. ................ 7 6

4-3 Rank scores scope and resource variables .............................................. ............... 78

4-4 Rank scores of structure and market factors ..................................... ........ ............... 79

4-5 Strategic group quality of care with Differentiator / nursing care focus cluster as the
reference group ............................................................................ 82

4-6 Strategic group quality of care with rehab focus cluster as the reference group ..............83

4-7 Strategic group quality of care with Specialty care focus cluster as the reference
g ro u p ............................................................................................. 8 4

4-8 Strategic groups and costs with cost leader as reference group regression results ............85

4-9 Strategic groups and predicted costs.......... ..... ......... ................... 85

4-10 Odds Ratio of Nursing Hom e Strategy ........................................ ......................... 86

4-11 Probability of nursing homes being in the highest operating margin tier .......................86

4-12 Probability of nursing homes being in the highest total margin tier..............................87

A-i Eigenvalues of scope (technology) factor analysis................................... ........100

A-2 Factor analysis with 10 factors of scope (technology) variables.............. ................ 102

A-3 Factor analysis with 10 factors of scope (technology) variables.............. ................ 104

A-4 Eigenvalues of scope (technology) factor analysis................................... ........106

A-5 Factor analysis of resource staffing intensity with three factors.................... ........ 106

A-6 W hard's minimum variance cluster analysis................................ ........................ 107

A-7 Cluster History ............................. ..................... ... ........ 107

A-8 Determination of rank scores table ......................... ............................ ........... .... 108

A-9 Descriptives of dependent variables, independent variables and control variables.........109







would have higher cumulative rank scores for resources and scope when compared to the cost

leadership group and would have lower cumulative resources and scope scores than the

differentiator. In Figure #2, the centroid best cost cluster would have fallen in the area between

the boundaries set by the cost leadership's centroid and the differentiating cluster's centroid.

Focus: Like any industry, it is possible for a nursing home to focus their strategy on a

smaller market segment. The focus group can be identified by providing a high amount of

technology and resources in a specific area like rehab services. For example, if a cluster had

high rank scores in post acute rehab, long term care rehab and therapy resources, but low scores

in every other category; they would be considered a rehab focus group. There was the possibility

that other focus groups also existed.

Lack of strategy: A facility that uses more resources but provides less technology than a

cost leader was considered as a facility that lacks a strategy.

Analysis of Variance

This study used Analysis of Variance (ANOVA) to determine if the component variables

of the newly identified strategic groups are significantly different from each other. The study

also used Tukey's method to make multiple pair wise comparisons of the means for each

variable. The ANOVA used the F test statistic which is the ratio of the between error estimates

divided by the within error estimates. If Ho is false, the between estimates tend to overestimate

the variance, so it tends to be larger than the within estimates. This can make the F statistic

significantly greater than 1.0 (Agresti and Finlay 1997). The Ho Hypothesis is that the mean of

each variable of interest (for example RNs per resident) was the same for each strategic group;

that is Differentiator = Cost Leadership = Best Cost = Focus = Lack of Strategy. To determine

whether the entire clusters are significant from each other, the study also used the MANOVA test

with Wilks' Lambda. Similar to ANOVA, this statistic evaluates the hypothesis that the











Table 4-12. Probability of nursing homes being in the highest total margin tier
Specialty


Dependent
Variable
Low
Mid
High
# Significant p>0.05


0.089
0.799
0.112


Cost
Leader4,5
0.094
0.799
0.106


Rehab
Focus4,5
0.094
0.799
0.106


care
Focus1,2,3
0.112
0.798
0.090


Lack
of Strategy1,2,3
0.109
0.799
0.092


Differentiator '







CHAPTER 5
DISCUSSION

This study concurred with prior research because it demonstrated the existence of strategic

groups in the nursing home industry. However, this study used a different conceptual framework

to identify them. While earlier studies by Zinn et al. 1994 and Marlin et al. 2002 used resources

and payer mix, this study used the business level decisions of resource allocation and technology

provision. One advantage of using technology over payer mix variables to identify strategic

groups is the applicability of the results for managers. Management has control over how much

technology to provide and how much resource to use. How a nursing home decides to balance

these two factors is an important part of that facility's overall strategy. The results from this study

demonstrate that different strategic groups provide their technology and utilize their resources in a

different way.

Using technology and resource allocation to identify the clusters makes it possible to define

the groups using Porter's generic strategies. There are no nursing home studies that have

attempted to define strategic groups by Porter's generic strategies. However, Marlin et al. 2002

did find the existence of Porter's strategies in the hospital industry. Unlike this study, Marlin et al

2002 explicitly used variables that defined the differentiator strategic group. These variables

included the total number of hospital units, the number high technology units and ratio of

intensive care beds to total beds. They did not use any measures that reflect how the care was

provided. Although a hospital could have a high number of different specialty units, it did not

necessarily provide a high degree of technology within those units. Each specialty unit could have

been poorly staffed and the patients could have received a very low level of services. Marlin et al.

2002 study was examining the governance level of decision making as opposed to the business

level decisions. Business level operations are significantly impacted by the day to day decisions of

how to provide service and how to use resources.







themselves. It is a routine service and does not require the skill of the therapist. However the

goals of the program need to be generated by the therapist.

Restorative amputee intervention: The restorative amputee intervention variable is facility

level and is continuous. It represents the average number of days that a resident receives

restorative amputee intervention while remaining in the facility. It is generated by dividing the

resident level restorative variable that represents the number of days of amputee intervention

provided in the 7 days before the assessment date by the total number of residents in the facility.

Restorative amputee intervention have the nursing staff assist with the donning and doffing of

prosthetic in order to improve the resident's ability to don and doff their prosthetic by

themselves. It is a routine service and does not require the skill of the therapist. However the

goals of the program need to be generated by the therapist.

Restorative communication: The restorative communication variable is facility level and is

continuous. It represents the average number of days that a resident receives restorative

communication while remaining in the facility. It is generated by dividing the resident level

restorative variable that represents the number of days of communication provided in the 7 days

before the assessment date by the total number of residents in the facility. Restorative

communication has the nursing staff assist with the communication in order to improve the

resident's ability to communicate. It is a routine service and does not require the skill of the

therapist. However the goals of the program need to be generated by the therapist.

Restorative other treatment: The restorative other treatment variable is facility level and is

continuous. It represents the average number of days that a resident receives other restorative

while remaining in the facility. It is generated by dividing the resident level restorative variable

that represents the number of days of other provided in the 7 days before the assessment date by

the total number of residents in the facility.









Table A-8. Determination of rank scores table
RN focus n=1634 Rehab Focus n=1816 Differentiator n=990 Cost leader n=2752 Lack of strategy n=2554
V LV TV R V LVTR V LV TV RV LV V LV TV R

Zwages 0.15 -0.10 0.25 5.0 0.10 -0.10 0.20 4.0 0.00 -0.10 0.10 2.0 -0.06 -0.10 0.04 0.8 -0.10 -0.10 0.00 0.0
ZRNperres 1.16 -0.56 1.72 5.0 0.38 -0.56 0.94 2.7 0.02 -0.56 0.58 1.7 -0.56 -0.56 0.00 0.0 -0.42 -0.56 0.14 0.4
ZLPNperres -0.26 -0.26 0.00 0.0 0.57 -0.26 0.83 5.0 -0.10 -0.26 0.16 1.0 -0.08 -0.26 0.18 1.1 -0.12 -0.26 0.14 0.9
ZCNAperres -1.02 -1.02 0.00 0.0 0.46 -1.02 1.48 5.0 0.27 -1.02 1.29 4.4 0.04 -1.02 1.06 3.6 0.18 -1.02 1.20 4.0
Therapy -0.07 -0.81 0.74 1.8 1.31 -0.81 2.12 5.0 -0.39 -0.81 0.42 1.0 -0.81 -0.81 0.00 0.0 0.14 -0.81 0.95 2.2
Therapy assist -0.19 -0.91 0.72 1.6 1.33 -0.91 2.24 5.0 -0.38 -0.91 0.53 1.2 0.20 -0.91 1.11 2.5 -0.91 -0.91 0.00 0.0
ZPT percent 0.06 -0.79 0.85 2.4 -0.24 -0.79 0.55 1.6 0.10 -0.79 0.89 2.5 -0.79 -0.79 0.00 0.0 0.95 -0.79 1.74 5.0
ZOT percent 0.07 -0.76 0.83 2.6 -0.14 -0.76 0.62 1.9 0.07 -0.76 0.83 2.6 -0.76 -0.76 0.00 0.0 0.84 -0.76 1.60 5.0
ZRN percent 1.67 -0.58 2.25 5.0 0.06 -0.58 0.64 1.4 -0.06 -0.58 0.52 1.2 -0.58 -0.58 0.00 0.0 -0.47 -0.58 0.11 0.3
ZLicense perc. 1.47 -0.48 1.95 5.0 0.21 -0.48 0.69 1.8 -0.32 -0.48 0.16 0.4 -0.45 -0.48 0.03 0.1 -0.48 -0.48 0.00 0.0

Total Resource 28 34 18 8 18
zltcskinpr-s 0.05 -0.31 0.36 2.1 0.56 -0.31 0.87 5.0 0.25 -0.31 0.56 3.2 -0.20 -0.31 0.11 0.6 -0.31 -0.31 0.00 0.0
Zrestorative 14d -0.21 -0.43 0.22 0.3 -0.43 -0.43 0.00 0.0 2.93 -0.43 3.36 5.0 -0.34 -0.43 0.09 0.1 -0.34 -0.43 0.09 0.1
Zrestorative Q -0.17 -0.31 0.14 0.3 0.05 -0.31 0.36 0.9 1.73 -0.31 2.04 5.0 -0.27 -0.31 0.04 0.1 -0.31 -0.31 0.00 0.0
Zrelief device -0.05 -0.17 0.12 1.5 0.24 -0.17 0.41 5.0 0.17 -0.17 0.34 4.1 -0.03 -0.17 0.14 1.7 -0.17 -0.17 0.00 0.0
Therapy min -0.08 -0.57 0.49 1.7 0.88 -0.57 1.45 5.0 -0.57 -0.57 0.00 0.0 -0.11 -0.57 0.46 1.6 -0.24 -0.57 0.33 1.1
Zacute skin care 0.07 -0.23 0.30 3.0 0.26 -0.23 0.49 5.0 0.23 -0.23 0.46 4.7 -0.08 -0.23 0.15 1.6 -0.23 -0.23 0.00 0.0
Ztoiletprogram -0.06 -0.07 0.01 0.1 -0.06 -0.07 0.01 0.1 0.29 -0.07 0.36 5.0 0.04 -0.07 0.11 1.5 -0.07 -0.07 0.00 0.0
Weight prog. 0.01 -0.10 0.11 2.5 -0.10 -0.10 0.00 0.0 0.12 -0.10 0.22 5.0 0.06 -0.10 0.16 3.7 -0.05 -0.10 0.05 1.0
Percent Mcare -0.13 -0.39 0.26 0.7 1.37 -0.39 1.76 5.0 -0.36 -0.39 0.03 0.0 -0.34 -0.39 0.05 0.2 -0.39 -0.39 0.00 0.1
Total scope 12 _26 32 11 2.3
V is the actual Z score value
LV is the lowest value for the variable among clusters
TV is the total variance is the difference between the highest value among the clusters and the lowest values among clusters
R is the final rank from 0 to 5. Determined ((V-LV)/TV*5)









Table A-20. OLS regressions of root transformed bowel
Coef. Std. Err.
Cost leader -0.001 0.003
Differentiator -0.009 0.004
Specialty care focus 0.002 0.003
Lack of strategy -0.004 0.003
Ownership 0.005 0.003
Percent Medicaid 0.000 0.000
Percent Medicare 0.001 0.000
Market competition 0.040 0.050
Occupancy 0.013 0.009
Total beds 0.000 0.000
Chain 0.004 0.002
Acuity index 0.011 0.001
ADL index 0.025 0.003
Tube feed care -0.039 0.029
Respiratory care -0.040 0.015
Suctioning care 0.014 0.077
IV therapy care 0.016 0.044
Tracheostomy care -0.101 0.085
Need injection 0.020 0.013
cons 0.186 0.017


decline with rehab focus as ref group
t P>t
-0.29 0.77
-2.22 0.03
0.64 0.52
-1.26 0.21
1.83 0.07
-3.76 0.00
3.82 0.00
0.79 0.43
1.49 0.14
2.94 0.00
1.79 0.07
7.26 0.00
7.43 0.00
-1.34 0.18
-2.74 0.01
0.18 0.86
0.36 0.72
-1.19 0.23
1.55 0.12
10.67 0.00


Table A-21. OLS regressions of root transformed ADL
Coef. Std. Err.
Cost leader 0.006 0.003
Differentiator -0.003 0.003
Specialty care focus 0.002 0.003
Lack of strategy 0.002 0.003
Ownership 0.001 0.002
Percent Medicaid 0.000 0.000
Percent Medicare 0.000 0.000
Market competition 0.083 0.053
Occupancy 0.006 0.007
Total beds 0.000 0.000
Chain -0.001 0.002
Acuity index 0.002 0.001
ADL index -0.003 0.003
Tube feed care -0.028 0.026
Respiratory care -0.005 0.013
Suctioning care -0.054 0.057
IV therapy care -0.013 0.033
Tracheostomy care -0.072 0.066
Need injection -0.008 0.011
cons 0.377 0.016


decline with
t
2.18
-0.86
0.86
0.63
0.50
0.10
2.97
1.57
0.84
-0.71
-0.72
1.46
-1.01
-1.10
-0.42
-0.95
-0.39
-1.10
-0.69
23.89


rehab focus as ref. group
P>t
0.03
0.39
0.39
0.53
0.62
0.92
0.00
0.12
0.40
0.48
0.47
0.14
0.31
0.27
0.67
0.34
0.70
0.27
0.49
0.00







Table 4-9 provides the cost predicted values for each strategic group. The predicted values

are the patient costs per resident day. Specialty care focus and Rehab focus groups had

significantly higher costs than the cost leader by 9 and 13 dollars per resident day respectively.

Table 4-8. Strategic groups and costs with cost leader as reference group regression results
Lack of Differentiation / Specialty Care
Strategy Nursing Care Focus Rehab Focus

Costs per resident day -0.006 (0.60) 0.002 (0.90) 0.039 (0.00)* 0.080 (0.00)*
Represents significant values compared to cost leadership cluster

Table 4-9. Strategic groups and predicted costs
Differentiator Cost Rehab Specialty Lack
/ Nursing Leader Focus Care of Strategy
Care Focus Focus
Costs per resident day 149.15 149.06 161.65* 155.03* 148.12

In dollars

Hypothesis #5

The initial analysis that included all the case mix variables was unable to converge. It was

therefore necessary for the study to use acuindex variable rather than the 7 individual case mix

variables. By dropping the 7 case-mix variables, the ologit analyses were able to converge (A-34

and A-35 in Appendix A).

The ologit for both financial performance measures (table A-34 and A-35 Appendix A) had

variables that were completely determined. 53 of the observations for operating margin and 57

observations for total margin observations were completely determined. There was possibility this

was the result of multicollinearity between some of the county level fixed effects. There was over

1500 county level fixed effects variables with many of them having a very low number of nursing

homes within the county. A correlation matrix was generated to determine if the completely

determined variables were either the independent variables of interest or one of the control

variables other than the county fixed effects variables. The correlation matrix (table A-36

appendix A) illustrated that none of the independent or control variables were completely

determined. As a result of the high number of county level fixed effects, the determined variables

85







variables are correlated and need to be combined, the new variable may be considered a measure

of the construct 'Pressure Sore Preventative Interventions'.

The study used the interaction between the scope and the resource dimensions to classify

the strategic groups. It was therefore necessary to perform separate exploratory factor analysis

for scope commitment and for resource commitment. Similarly, the staffing intensity and staffing

mix variables were also analyzed separately. This is because it is necessary to separate the

intensity and the skill mix of the staff to be able to identify the relationship between licensed and

non licensed staff. Staffing mix variables cannot be used as a substitute for the staffing intensity

(i.e. nurse per resident) variables because they do not explain the volume of nursing care

provided. For example, in the case of nursing, CNA's can be negatively correlated with the RNs

in one cluster while being positively correlated in another cluster. As a result, the three separate

factor analysis include technology variables, staffing intensity variables and skill mix variables.

Variables included in the cluster analysis but t not in the factor analysis were wages and

percent Medicare beds. Although average wage paid is a resource commitment measure and is

related to the staffing mix variables (higher RN mix represents higher average wages), it was not

included in any exploratory factor analysis. Wages are used by nursing homes as an incentive for

hiring new staff and retaining their staff and therefore were included separately in the cluster

analysis. Percent Medicare beds variable is related to scope because it indicates the proportion of

the facility residents getting rehab. It was included in the cluster analysis independently to

identify the total volume of therapy services provided. For example, a facility that provided a lot

of minutes of therapy to a small number of residents could be misclassified as being a rehab

facility if the total amount of residents receiving therapy were not accounted for.

Factor Matrix

In the factor matrix, the columns represent the derived factors while the rows are the

variables. The cells represent the factor loadings which represents the degree that the variables

58










capture cost leadership were measures that capture different types of cost like salaries and labor

hours. The variables used to capture the differentiation group were types of services. Marlin et al.

(2002) found that hospital strategic groups could be defined using Porter's generic strategies

theory and that different strategic groups had different structure-performance relationships.

Strategic Group Model Using Scope (Technology) Commitment, Resources Commitment
and Porter's Generic Strategy Theory

Figure #2 is a strategic group model of the nursing homes industry. The horizontal axis

represents the scope commitment of the nursing home as measured by the amount of technology

provided by the facility.

Upper Scope Boundary

High F\

SFocus


Lack of Strategy I Differentiator
Resource -1

R esorcF Best Cost

Coaster 1, Lower Resource Boundary


Low
Low High
Scope


Figure 2-2. Strategic Group Model Based on Scope and Resource commitment

The vertical axis represents the resources that the facilities commit to provide its

technology. The circles represent clusters as defined by Porter's strategies (Differentiator, Cost

Leadership, Best Cost and Focus). Those facilities that do not have a strategy are defined as

being groups that lack strategy







made up approximately 5% of all the county fixed effects and therefore likely have minimal

impact the overall model.

Hypothesis #5 was partially supported. Table 4-10 highlights that differentiator had a 20%

greater odds of being in the highest tier for operating margin and a 31% greater odds of being in

the highest tier for total margin relative to those nursing homes that lacked a strategy. Cost leaders

had 25% greater odds of being in the highest tier for operating margin and 22% greater odds of

being in the highest tier for total margin relative to those nursing homes that lacked a strategy.

Rehab focus had 46% greater odds of being in the highest tier for operating margin and 27%

greater odds of being in the highest tier for total margin relative to those nursing homes that

lacked a strategy. Differentiators, cost leaders and rehab focus had better total margins than the

specialty care focus strategy but only the rehab focus group and the cost leadership group had

better operating margins than the specialty care focus group. The specialty care focus group did

not have significantly higher odds of being in the highest tier with respect to the nursing homes

that lacked strategy.

Table 4-10. Odds Ratio of Nursing Home Strategy
Differentiation Specialty F test
Dependent Cost care Rehab
Variable Leader Focus Focus
Operating Margin 1.25 (0.01)* 1.20 (0.09)* 1.04(0.66) 1.46(0.00)* F=9.7 (0.05)
Total Margin 1.22 (0.01)* 1.31 (0.01)* 0.98 (0.80) 1.27 (0.01)* F=10.0 (0.03)
* Significantly different from lack of strategy

Table 4-11 and 4-12 illustrate the predicted probabilities of the operating and total margins.

Both demonstrate that the nursing homes with differentiation, cost leadership and rehab focus

strategy perform better than those nursing homes that lack strategy.

Table 4-11. Probability of nursing homes being in the highest operating margin tier
Specialty
Dependent Differentiator5 Cost Rehab care Lack
Variable Leader4'5 Focus4'5 Focus2'3 of Strategy1'2'3
Low 0.099 0.096 0.085 0.110 0.111
Mid 0.799 0.799 0.797 0.798 0.797
High 0.102 0.105 0.118 0.092 0.090
#Significant p>0.10
86







CHAPTER 4
RESULTS

Hypothesis #1

Identification of Different Factors: The study completed three separate factor analyses to

reduce the total number of variables. The newly generated composite variables represented valid

constructs of nursing home's technology and resource commitments. Initially there were 57

technology variables and 11 resource variables that were used in the factor analysis.

Factor analysis #1 (Technology): The first factor analysis included only the variables that

measured the technology of the nursing homes. Of the 57 reported variables, there were 20 factors

that were derived reporting an eigenvalue greater than 1 (table A-i in the appendix), with ten of

those factors having eigenvalues between 1.0 and 1.54. Although the factors just above one would

qualify using the Kaiser (1960) criteria, the amount of variation removed by each subsequent

factor after 10 was marginally inadequate. Therefore it was necessary to use the scree plot to

identify the ideal number of clusters. The scree plot indicated a point of inflection between 9 and

11 factors (Figure A-i in the appendix). Although the eigenvalues could justify the use of 20

factors, the scree plot indicates that the change in the amount of variance contributed by using

additional factors beyond 10 appeared to be non- significant. Therefore, the factor analysis for

technology only used the 10 factors.

All variables with a factor loading above 0.4 were initially included in the each of the 10

factor variables (Table A-2 Appendix A). There were six variables that were included in more

than one factor. Since each of these six variables can only be in one factor, it was necessary to

determine the most appropriate factor for each variable. The six variables included long term care

pressure relieving chairs, long term care pressure relieving beds, long term care and post acute

restorative PROM and AROM. The long term care pressure relieving chairs and beds had loadings

that were similar in both factors (0.6). One of the factors also included post acute pressure

relieving devices. Therefore this factor generated a clear construct of pressure relieving devices.
74








especially in comparison to the differentiator and the lack of strategy group. This group had higher

LTC prevalence and higher PA incidence of pressure ulcers than cost leader and lack of strategy

group.

Table 4-7 (full regression results are available in table A-26 to A-33 in Appendix A)

illustrates that nursing homes in the specialty care focus group had lower quality of care

deficiencies than the cost leaders and the lack of strategy group. They had less pressure ulcer

prevalence than facilities in the rehab focus group. The specialty care focus group had higher

prevalence of pressure ulcers than cost leader and lack of strategy. It had higher bladder decline,

bowel decline, ADL 4 point decline than differentiators. It had lower restraint use than cost

leaders. With respect to post acute incidence of pressure sores, the results were mixed with

specialty care focus having a higher incidence of pressure ulcers than lack of strategy group. And

a higher prevalence of long term care pressure sores that cost leaders and lack of strategy.

Table 4-7. Strategic group quality of care with Specialty care focus cluster as the reference group
Differentiator /
Dependent Variables Cost Nursing care focus Specialty care Lack of Strategy
Quality of care deficiencies 1.06 (0.05)* 0.99 (0.80) 1.00 (0.97) 1.12 (0.00)*
Long term care measures
Bladder Decline -0.003 (0.26) 0.003 (0.39) -0.009 (0.01)y -0.004 (0.12)
Bowel Decline -0.003 (0.27) -0.002 (0.52) -0.012 (O.OO)y -0.006 (0.03)y
ADL 4 pt decline 0.003 (0.16) -0.002 (0.39) -0.005 (0.10)y -0.001 (0.78)
Restraint use 0.005 (0.22) -0.006 (0.19) -0.002 (0.68) -0.001 (0.88)
Pressure Ulcers -0.005 (0.01)y 0.008 (0.00)* -0.004 (0.14) -0.008 (O.OO)y
Post acute measures
Ambulation improvement -0.005(0.74) 0.014 (0.47) -0.037 (0.08)* 0.024 (0.15)
Pressure Ulcers incidence -0.018 (0.25) 0.012 (0.49) 0.027 (0.20) -0.049 (O.OO)y
*Quality of care results are incidence rate ratio
Square root transformation LTC bladder decline, bowel decline pressure ulcer prevalence, LTC 4 point ADL decline, restraints,
Post acute pressure sore incidence are OLS regression coefficients. Post acute walking improvement was glm with gamma
distribution. Represents significant values in the desired direction. y Represents significant values in the negative
direction.

Hypothesis #4

Hypothesis #4 was not supported. Table 4-8 illustrates that cost leaders nursing homes had

significantly lower costs that specialty care and rehab focus groups, but no significant difference

between cost leaders and differentiators or with those facilities that lack a strategy altogether.









implementation of the caps was delayed by a number of moratoriums. Even in 2005, after the

caps were implemented, nursing homes were able to circumvent the caps if a resident had

therapy needs that were deemed medical necessary (APTA 2008).

Post Acute Care Reimbursement

Reimbursement for post-acute care is primarily reimbursed for by Medicare; however it

can also be covered by private insurance and the VA. Approximately 12% of the beds in nursing

homes are occupied by rehabilitation patients paid for by Medicare (GAO 2000). These patients

usually have experienced a decline in their health status and require rehabilitation before they

can return either home or to the next appropriate level of care.

Medicare reimbursement uses a prospective payment system. Similar to Medicaid case mix

reimbursement, residents are classified using the Resource Utilization Groups (RUG). These

groups are commonly determined by the amount of therapy and nursing services that the resident

requires. For example, a facility will receive a higher daily rate if the resident gets more therapy.

Each rehab RUG level has a minimum amount of therapy minutes that have to be provided

within 7 days prior to a given assessment reference date. For example, for a patient to be

classified as an ultra high rehab RUG level, they would have to receive 720 minutes of therapy

during the 7 day assessment period. Patients are assessed a maximum of 5 times depending on

the amount of time they remain at the facility for rehab. The assessments include the 5-day, 14

day, 30 day, 60 day and the 90 day assessment. The 5 day assessment identifies the RUG level

for day 1 to 14. The 14 day assessment identifies the RUG level for day 15-30. Each subsequent

assessment identifies the RUG level for the following period. Although the each resident is

eligible for up to 100 days of rehab per medical incident, most patients are discharged from

subacute services well before the 100 day limit because they go home or are no longer making









Table A-16. GLM with gamma distribution with PA walking improve with differentiators as ref group
Coef. Std. Err. z P>z
Cost leaders 0.031 0.019 1.65 0.10
Rehab focus 0.050 0.022 2.32 0.02
Specialty care focus 0.037 0.021 1.76 0.08
Lack of strategy 0.060 0.019 3.17 0.00
Ownership -0.087 0.014 -6.04 0.00
Percent Medicaid -0.004 0.000 -10.18 0.00
Percent Medicare -0.002 0.001 -2.81 0.01
Market competition -0.425 0.248 -1.71 0.09
Occupancy -0.160 0.043 -3.73 0.00
Total beds 0.000 0.000 -0.53 0.60
Chain -0.032 0.012 -2.74 0.01
Acuity index -0.028 0.007 -4.07 0.00
ADL index -0.110 0.017 -6.49 0.00
Tube feed care -0.666 0.143 -4.66 0.00
Respiratory care 0.139 0.081 1.71 0.09
Suctioning care 0.444 0.419 1.06 0.29
IV therapy care 0.240 0.213 1.13 0.26
Tracheostomy care -0.238 0.449 -0.53 0.60
Need injection 0.038 0.068 0.56 0.58
cons 0.38 0.283 0.77 0.44


Table A-17. OLS regressions of root transformed PA skin ulcers with
Coef. Std. Err.
Cost leaders -0.045 0.020
Rehab focus -0.015 0.021
Specialty care focus -0.027 0.021
Lack of strategy -0.076 0.020
Ownership 0.034 0.014
Percent Medicaid 0.001 0.000
Percent Medicare 0.001 0.001
Market competition -0.031 0.308
Occupancy 0.014 0.046
Total beds 0.000 0.000
Chain -0.021 0.012
Acuity index 0.034 0.007
ADL index 0.027 0.019
Tube feed care 0.568 0.156
Respiratory care 0.088 0.082
Suctioning care 0.535 0.413
IV therapy care 0.270 0.194
Tracheostomy care -0.160 0.431
Need injection 0.183 0.069
cons -2.513 0.098


differentiators as
t
-2.30
-0.74
-1.28
-3.81
2.38
3.39
0.70
-0.10
0.31
5.08
-1.76
4.88
1.42
3.64
1.07
1.30
1.40
-0.37
2.63
-25.77


ref. group
P>t
0.02
0.46
0.20
0.00
0.02
0.00
0.48
0.92
0.75
0.00
0.08
0.00
0.16
0.00
0.28
0.20
0.16
0.71
0.01
0.00









maximization are needed (Scanlon, 1980). Finally, health care organizations (like nursing

homes) are often charged with the responsibility of providing care for one of the most vulnerable

populations. The direness of poor quality health care is so great that using financial performance

as the sole measure of success (like in other industries) is just not adequate. For the

aforementioned reasons, the measurement success in the health care industry has to be based on

quality of services provided, cost of that service and financial performance.

New Contributions

To date, no study has examined if strategic groups can be identified by the technology

choices made by the nursing home management. The purpose of this study is to determine if

strategic groups can be classified using Porter's generic strategies based on how nursing home

management allocate their resources and how they select their technology. The study will also

determine if these groups are succinct and if they differ with respect to quality, financial

performance and cost. What and how much of technology a nursing home administrator chooses

to provide is an important strategy choice that can have significant repercussions. The study is

different from prior studies because it uses the industry's technology as the scope variables to

determine the strategic groups and will define those groups using Porter's generic strategies.

Hypothesis

Hypothesis #1: Nursing homes that have a strategy can be categorized into one of four

strategic groups: Differentiator, cost leadership, focus or best cost.

Hypothesis #2: Nursing homes in the differentiator strategic group will provide higher

quality care than other strategic groups.

Hypothesis #3: Nursing homes in focus strategic group will provide higher quality in the

area that they are focused on than other strategic groups.







Control Variables

Organizational characteristics that will be analyzed are size, for-profit status, chain

affiliation, payer mix, market competition and case mix as described in hypothesis #2.

Analysis

Similar to health care expenditure measures, the cost per patient per day tends to be

skewed to the right with some facilities having significantly higher cost than the mean. Similar to

prior studies dealing with health care cost, the study used the generalized linear model (GLM)

with gamma distribution and a log link (Fleishman et al. 2006, Pagano et al. 2005). The study

then determined the predicted value using the smearing method. The study used county level

fixed effects. The study used White's heteroskedacity-consistent covariance matrix (1980) to correct

for any heteroskedascity as found in STATA (2001). It used county level fixed effects. Dependent

variables outliers of 5 standard deviations and above were dropped from the dataset.

FlnCost = V + &giGroupl + &g2Groupl + .....+&gGroupN + &cControl + &feFE + s,

FlnCost represents one of the seven quality dependent variables.V represents the constant

value, &gn represents the effect of each strategic group on the quality variable. &gc represents the

effect of the control variables on the quality variable. &fe represents the effect of the county level

fixed effects. E represents the random error.

Hypothesis #5: Differentiation, Cost Leadership, Focus and Best Cost Strategic Groups will
have better Financial Performance than Nursing Homes that Lack a Strategy

Dependent Variables

There were two variables used to examine financial performance. The first of the two

financial performance variables used, the operating margin, is calculated using the operating

revenue minus the operating expense then divided by the operating revenue. Operating revenue

and expense were direct measures of doing business and therefore the operating margins may be

more sensitive to nursing home strategies. The second financial performance variable included

total profit margin. Profit margin not only includes the operating revenue and operating expense,
71









private-pay utilization, and Veteran's Administration utilization, average length of stay, average

patient age and case mix. For resource commitment, Marlin et al. 1999 used percentage of

nursing costs, percentage of ancillary costs, occupancy rate, semiprivate room rate, number of

beds, registered nurses (RNs) per resident, and staff per resident. The study also used cluster

analysis and found that the group with the highest private pay utilization combined with high

Medicare utilization generally performed better along financial and quality indicators.

Scope Using Technology

Prior studies of strategic groups in nursing homes have used the range of market segments

as the measure of scope commitment (Zinn et al. 1994, Marlin et al. 1999). Zinn et al. 1994 and

Marlin et al. 1999 argued that to compete effectively, nursing homes need to be responsive to

differences in payer demand characteristics because the payer mix of a nursing home is the major

basis of segmentation in the industry (Zinn et al. 1994, Marlin et al. 1999). However, using payer

demand characteristics may not be the best way of determining the strategic groups because the

payer mix of a nursing home may be more the result of good strategy. For example, many

facilities may have a strategy to acquire Medicare and private pay residents but not be successful.

Whether a facility has a high proportion Medicare and private pay beds may be independent of

the nursing homes' attempt to acquire these types of residents.

Structure Conduct Performance theory is a model used to link elements of the market

structure to business conduct and performance in industrial economics. Structure refers to market

structure defined mainly by the concentration of market share in the market. Conduct refers to

the behavior of firms whether competitive or collusive (pricing and production, production, goals

of firms, promotion) (Britton et al. 1992). Performance is mainly defined by the consequences of

market power. Based on the Structure Conduct Performance paradigm, payer demand

characteristics would be considered performance measures rather than conduct measures. In the









3 METHODOLOGY ................................. ......... ....................... 43

D ata ......... ....... ... ............. ... ............................................................... 4 3
P o p u latio n ........... ..... ... ............. ... ...........................................................4 4
Overview of M ethodology............................................. .. .... .. ..................... .....45
Hypothesis #1: Nursing homes that have a strategy can be categorized into one of four
strategic groups: Differentiator, cost leadership, focus or best cost. ................................45
S c o p e ..........................................................................4 5
Nursing Technology .................................... ..... .. ...... ............... 45
R ehab T technology ................ .. ............................................................. ..... 52
R source C om m itm ent .......................... ............................................... ............... 54
Hypothesis #1 Analysis ................................... ..... .. ...... ............... 56
Factor A analysis .................................................................................... .................... 57
Factor M atrix ..................................................................................... ..................... 58
N um ber-of-F actors P problem ......................................... .............................................59
Rotation of the Factor Structure ........................................................ ............. 59
Internal Consistency of M measures ............................................................................60
C lu ste r A n a ly sis ......................................................................................................... 6 0
Strategic G groups D defined ........................................................................ .................. 63
D eterm ine the total variance value........................................................................ 63
C lu ster v alu e ............. ................................................................. ............. 63
C lu ster rank score ............. ..................................................... ........ ........ 64
A analysis of V ariance .......... ............................... .. ..... .. .. .......... .. .... ..... .... .......... .... 65
Hypothesis #2: Nursing Homes in the Differentiator Strategic Group will have the
H highest Q quality. ..................................................................66
Dependent Variables: ............................................66
Independent V ariables ..................................... ......... ...... .... .. ........ .... 67
Control Variables.......... ........ ............. .. ... .... ..................67
A n aly sis ............................................... ..... .. ..... ........ ............... 6 8
Hypothesis #3: Nursing Homes in the Focus Strategic Group will provide Higher
Quality in the area that they are focused on compared to other Strategic Groups..............69
D dependent V ariables ......................... ......... .. .. ........... ... ...... 69
Independent Variables ............. ................................... .... ......... 69
Control Variables.......... ........ ............. .. ... .... ..................70
A n aly sis ......................................... .. .. .. ............ ....... .. ............... 7 0
Hypothesis #4: Nursing homes in the Cost Leadership Strategic Group will have the
low est costs. ..................................................................70
D dependent V ariables ......................... ......... .. .. ........... ... ...... 70
Independent V ariables .................................................. ........ ..............70
C o n tro l V ariab le s....................................................................................................... 7 1
A n aly sis ..................... .... ..... .. .. ............... ........ ...... ... .. .............. 7 1
Hypothesis #5: Differentiation, Cost Leadership, Focus and Best Cost Strategic Groups
will have better Financial Performance than Nursing Homes that Lack a Strategy ...........71
D dependent V ariables ......................... ......... .. .. ........... ... ...... 71
Independent V ariables ............. ...... ............................................... ........ 72
Control Variables.......... ........ ............. .. ... .... ..................72
A n a ly sis ................................................................................7 2


6







the relative locations of the points to each other; however, the actual coordinates of the points,

that is, the factor loadings would of course change (Kachigan 1991).

Rotational strategies: There are various rotational strategies that have been proposed. The

goal of all of these strategies is to obtain a clear pattern of loadings, that is, factors that are

somehow clearly marked by high loadings for some variables and low loadings for others.

Fabrigar et al. 1999 describe two commonly used types of rotational strategies: orthogonal

and oblique. Although orthogonal is conceptually simpler, factors have to be uncorrelated.

Oblique rotation on the other hand does not have this limitation. Since the resource factors have

the potential of being correlated to the technology that they are used for (for example, therapy

resources will be correlated with rehab technology), this study used the oblique rotation.

Internal Consistency of Measures

It was necessary to determine whether the variables of the generated factors have internal

consistency. Cronbach alpha tested the correlation between the variables of the construct. The

higher the alpha value, the closer the variables are correlated (Cronbach 1951). This value should

be above 0.7 (Allen 2002). If the alpha was lower than 0.7, the factor was dropped because the

variables did not correlate well together. To determine how much impact the individual variables

had on the overall alpha, multiple tests will provide a score of the resulting alpha. Any variable

whose exclusion causes the alpha score to rise above 0.7 were be dropped. Similarly, any

variable whose inclusion causes the alpha score to drop below 0.7 were dropped.

Cluster Analysis

Cluster analysis is an exploratory data analysis tool which aims at sorting different objects

into groups in a way that the degree of association between two objects is maximal if they belong

to the same group and minimal otherwise (Kachigan 1991). Computationally, cluster analysis

could be thought of as analysis of variance (ANOVA) "in reverse." The program starts with k

random clusters, and then move objects between those clusters with the goal to 1) minimize







Incontinence reducing technologies: All processes of care provided by the nursing staff

intended for the patient engage in a program that is directed towards accomplishing continence.

Toiletingprogram: This is a continuous facility level variable that represents the

proportion of residents in the nursing home that have a toilet program. Toileting programs

include such items like toileting schedules.

Bladder retraining: This is a continuous facility level variable that represents the

proportion of residents in the nursing home that have a bladder retraining program. Bladder

training programs include such items like toileting schedules.

Nutrition technology: All processes of care provided by the nursing staff intended for the

patient engage in a program that is directed towards maintaining a healthy weight.

Weight management program: This is a continuous facility level variable that represents

the proportion of residents in the nursing home that have a weight management program.

Resident is receiving a program of which the documented purpose and goal are to facilitate

weight gain or loss (e.g., double portions; high calorie supplements; reduced calories; 10 grams

fat).

Therapeutic diet: This is a continuous facility level variable that measures the proportion

of residents that are on a therapy diet. A diet ordered to manage problematic health conditions.

Examples include calorie-specific, low-salt, low-fat lactose, no added sugar, and supplements

during meals.

Plate guard: This is a continuous facility level variable that represents the proportion of

residents in the nursing home having a plateguard. This includes any type of specialized, altered,

or adaptive equipment to facilitate the resident's involvement in self performance of eating.

Rehab Technology

Rehab technology is utilized to improve the resident's health status. It is provided by

licensed therapy staff. There are different types of therapist that have different functions. These

52







relate to the derived factor. Factor loadings are used to reveal the extent that the variables

contribute to the meaning of the variable. Factor loading are the correlation coefficients between

the variables and the factors. A factor loading score of greater than 0.4 was the inclusion

criterion to be considered a part of that factor (Kachigan 1991).

Number-of-Factors Problem

One concern when using factor analysis is how many factors are retained. Each factor

accounts for a certain amount of variance in the data. Typically, the first factor will extract the

most variance; the second factor will extract the second most and so on. Once the variance of

each successive factor extracts is known, it is possible to determine how many factors can be

retained. There are two primary methods that can be used to determine the amount of factors

retained:

1) Eigenvalue is a measure of the variation and represents the equivalent number of

variables that the factor represents. The Kaiser criterion is a commonly used standard that

includes only factors with eigenvalues greater than 1 (Kachigan 1991). The rationale for using

such a threshold is that unless a factor extracts at least as much variation as the equivalent of one

original variable, it should be dropped. This criterion was proposed by Kaiser (1960), and is

probably the one most widely used standard in determining how many factors should be used.

2) Scree plot is another method that can be used to identify the number of factors. It is also

possible to identify how many factors to retain by plotting the incremental variance accounted

for by each successive factor. The idea is that the tail of the curve represents mostly random error

variance and therefore the factor solution is to include all factors that exist just prior to the

leveling off of the curve (Kachigan 1991).

Rotation of the Factor Structure

We could plot the factor loadings shown above in a scatter plot. In that plot, each variable

is represented as a point. In this plot we could rotate the axes in any direction without changing









Table A-36. Odds of nursing home of being in the highest total margin tier
Odds Ratio Std. Err. z P>z
Cost leader 1.22 0.09 2.58 0.01
Rehab focus 1.27 0.11 2.74 0.01
Specialty care focus 0.98 0.09 -0.25 0.80
Differentiator 1.31 0.14 2.6 0.01
Ownership 1.64 0.12 6.62 0.00
Percent Medicaid 1.06 0.03 1.98 0.05
Acuindex 1.00 0.00 1.29 0.20
Total beds 0.92 1.25 -0.06 0.95
Market competition 1.04 0.07 0.58 0.56
Chain Affiliation 10.53 15.18 1.63 0.10
Ifips_2 18.39 18.84 2.84 0.00
Note: 57 observations completely determined. Standard errors questionable









In most industries, success is most commonly defined by financial metrics; however,

Venkatraman et al. 1986 have suggested that other measures such as market share, new product

introduction and product quality can be equally important.

One measurement that has recently gained popularity in business is the balanced

scorecard (BSC) approach. A major characteristic of a BSC is that it combines long-range

strategic financial goals with day-to day operations. Kaplan and Norton (2001) stated that the

customer's perspective and internal operations are important areas that should be considered

along with the financial perspective.

Taking the customer's perspective, the ability of a nursing home to provide quality care is

crucial if a facility is going to survive over the long term. Poor quality can result in higher costs

(Weech-Maldonado et al. 2003) as well as lower revenue streams due to lower occupancy rates.

Poor quality can also result in moratorium on new admissions and even a facility's closure.

Quality in nursing homes has also generated concern for policymakers and the public ever since

the release of the Institute of Medicine (IOM) report in 1986. It is therefore very important that

quality indicators be used in conjunction with financial variables when measuring performance.

There are three key reasons why measures other than financial performance are needed in

the nursing home industry. First, the government programs like Medicare and Medicaid were

established to pay for the care of the program's beneficiaries. Although enough funds must be

allocated to keep companies viable, the purpose of the program is not to give organizations

above normal profits. Second, not all health care organizations are for profit. For example, nearly

one third of nursing homes in the US are either government or not for profit facilities (Harrington

et al. 2001) and it is important to consider what health care organizations do. The presence of

not-for-profit in the industry is another reason why performance objectives other than profit










http://www.cms.hhs.gov/NursingHomeQualityInits/downloads/MDS20rai 1202ch3.pdf Accessed March
30th, 2008

Centers for Medicare and Medicaid : State Operations Manual Appendix PP Guidance to Surveyors for Long
Term Care Facilities Available at:
http://www.cms.hhs.gov/manuals/Downloads/soml07ap_pp_guidelinesltcf.pdf Accessed March 30th, 2008

Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16(3), 297-334

Day, G., Strategic Market Planning, West Publishing Co., St. Paul, 1984

Dess, G., Davis, P., (1984), Porter's (1980) Generic Strategies as Determinants of Strategic Group Membership
and Organizational Performance. Academy of Management Journal; Sep84, Vol. 27 Issue 3, p467-488, 22p, 5
charts

Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., & Strahan, E. J. (1999). Evaluating the use of exploratory
factor analysis in psychological research. Psychological Methods, 4, 272-299.

Feng Z, Grabowski DC, Intrator O, Mor V. The effect of state medicaid case-mix payment on nursing home
resident acuity.Health Serv Res. 2006 Aug;41 (4 Pt 1):1337.

Feng Z, Grabowski DC, Intrator O, Zinn J, Mor V.Medicaid payment rates, case-mix reimbursement, and
nursing home staffing--1996-2004. Med Care. 2008 Jan;46(1):33-40.

Fiegenbaum, A; Thomas, H., INDUSTRY AND STRATEGIC GROUP DYNAMICS: COMPETITIVE
STRATEGY IN THE INSURANCE INDUSTRY, 1970-84. Journal of Management Studies, Jan93, Vol. 30
Issue 1, p69-105, 37p

Fleishman JA, Cohen JW, Manning WG, Kosinski M. Using the SF-12 health status measure to improve
predictions of medical expenditures. Med Care. 2006 May;44(5 Suppl):I54-63.

Ford EW, Duncan JD, Ginter PM., The Structure of State Health Agencies: A Strategic Analysis. MCR&R
60:1 (March 2003)

Fries BE, Schneider DP, Foley WJ, et al. Refining a case-mix measure for nursing homes: Resource Utilization
Groups (RUG-III). Med Care.1994;32:668-685.

Gapenski, L., Understanding Healthcare Financial Management, Third Edition, 2000 Health Administration
Press

GAO Report to Congressional Requesters (July 2001), Nursing Home Quality: Prevalence of Serious
Problems, While Declining Reinforces Importance of Enhanced Oversight,
http://www.gao.gov/new.items/d03561 .pdf

Grabowski DC, Castle NG. Nursing homes with persistent high and low quality. Med Care Res Rev. 2004
Mar;61(1):89-115.

Grabowski DC. A longitudinal study of Medicaid payment, private-pay price and nursing home quality. Int J
Health Care Finance Econ. 2004;4:5-26.







Cluster Analysis: The 10 factor variables and the 9 individual variables were standardized

using z scores that were based on the county mean. The dataset were then trimmed for outliers

using the 5 standard deviations for each variable. This resulted in the reduction of the number of

nursing homes used in the cluster analysis from 11601 to 9726. The newly generated variables

were used in the Ward's Minimum Variance Cluster Analysis and the k-means cluster analysis.

Ward's Minimum Variance Cluster Analysis identified 5 clusters as the best number of clusters.

This was based on a local peak in the pseudo F statistic, a small value of the pseudo t2 statistic

followed by a larger pseudo t2 statistic for the next cluster solution level, and that adding an

additional cluster increased the overall fit by less than 5 percent (Table A-6 and A-7 Appendix A).

No other cluster arrangement met the three criteria. The data was then analyzed using the k-means

with an a priori of 5 clusters.

Clusters: Table 4-3 represents the results of the cluster analysis. Using the ranking protocol

described in the methodology (Table A-8 Appendix), the different clusters were defined using the

rank score ranging from 0 to 5. 0 represented the lowest and 5 represented the highest value for

both the resource variables and the technology variables. For each cluster, the cumulative rank

score was calculated by summing the values for the scope variables and summing the resource

variables. These cumulative rank scores were plotted on a graph to generate model as indicated in

Figure 4-1. The scope score was represented by the horizontal axis and the resource score was

represented by the vertical axis. Table 4-4 provided some descriptive structural, market and case

mix variables of each strategic group.

The ANOVA test identified that the clusters were significantly different from each other for

each variable. In addition, the Tukey's test compared each pair of clusters for every variable and

found that there were significant differences between each pair 79% of the time. Finally, the

MANOVA test criteria (Wilks' Lambda) found that the entire clusters were significantly different

from each other (p <.000). Descriptives of all variables is located in table A-9 in the appendix.

77









ACKNOWLEDGMENTS

There are a number of people who provided me with the opportunity to achieve this goal.

To begin with, I would like to say thanks my mother who not only supported and encouraged me

to learn but also encouraged me to finish what I started. I would like to thank my father who

inspired me to think outside of the box. I would like to thank Dr. Weech-Maldonado for his

guidance, generosity and tremendous patience. And finally I would to thank my wife, whose

unconditional love, support and understanding made it all possible.









facility in the first place. The desired outcome for long term care residents is to maintain their

health status for as long as possible.

The second major function of a nursing home is to provide short term care (also known as

post acute or sub acute) for those residents that require skilled services after a qualified hospital

stay. To qualify for a post acute stay, residents must have either a skilled nursing care need (e.g.

IV antibiotics) or a rehab care need based on the recommendations of the hospital therapists.

The goal of post acute care is to improve the resident's health status so that the resident can be

discharged home or to the most appropriate level of care. The discharge options are quite broad

and can range from patients going home by themselves to becoming long term residents in a

nursing home. Discharge from a skilled nursing rehabilitation facility (also known as a post acute

or subacute facility) can occur when the patient reaches their highest level of function or the stay

is no longer covered by the payer. The desired outcome for a rehabilitation resident is a better

health status.

Long Term Care Resident Reimbursement

Reimbursement for long term care is primarily provided by Medicaid or by the residents

themselves (private pay). A small proportion of residents are covered by other payers like the

Department of Veteran Affairs (VA) and private insurance. Although Medicare does not pay for

a long term care resident's stay in a nursing home, Medicare does cover some ancillary services

like physical or occupational therapy.

Medicaid

Medicaid is the major payer of nursing home long-term care services. Over 50% of nursing

homes revenues and 70% of nursing home beds are covered by Medicaid (Rhoades et al. 2000).

Swan et al. 2000 describe five rate setting methodologies that state Medicaid programs use to







but also includes the non patient revenues like charitable donations and non patient expenses like

capital costs. Although profit margin is not as sensitive as operating margins to changes in the

nursing homes' strategies, it is still important because it does represent the overall picture of the

nursing homes financial performance. Financial performance is important because it strongly

influences the nursing homes behavior. If strategies that produces higher quality result in better

financial performance, then customers' needs and providers' goals will be aligned. However, if a

strategy provides higher quality of care but does not provide the facility with better financial

performance, alternative strategies by the nursing homes will likely be considered.

Independent Variables

The independent variables of interest included a dichotomous variable (0 or 1) for each

strategic group generated by the cluster analysis. The lack of strategy group was used as the

reference group since it was expected that nursing homes that had a strategy would have better

financial performance than those that did not have a strategy.

Control Variables

Organizational characteristics that were utilized as controls included size, for-profit status,

chain affiliation, Medicaid and case mix as described in Hypothesis #3. Unlike the other

hypothesis, hypothesis #5 only used Medicaid as a control variable. Some nursing homes are

trying to acquire the higher reimbursing Medicare as part of their strategy. By using the percent

Medicare variable as a control variable, it would not be possible to determine the benefit of

having a strategy directed towards acquiring rehab patients.

Analysis

Although the operating margin and the total margin were both continuous variables,

neither of them is normally distributed and therefore Ordinary Least Squares regression is not

appropriate. By the nature of the mathematical equation that is used to determine them, their

range is from negative infinity to positive one. This is because both equations use the revenue of












Table A-6. Ward's minimum variance cluster analysis
The CLUSTER Procedure
Ward's Minimum Variance Cluster Analysis
Eigenvalues of the Covariance Matrix
Eigenvalue Difference Proportion Cumulative


6.24393174
5.61295630
4.08166266
3.49904251
1.82518078
1.42486178
1.15088786
1.00157383
0.54994229
0.42184551
0.40626272
0.33925671
0.23580760
0.14780646
0.12732224
0.10034939
0.08968920
0.06519883
0.03024354
0.02011316
0.00687351


0.63097545
1.53129364
0.58262014
1.67386173
0.40031900
0.27397392
0.14931403
0.45163154
0.12809678
0.01558279
0.06700601
0.10344911
0.08800114
0.02048422
0.02697285
0.01066020
0.02449037
0.03495529
0.01013037
0.01323966


0.2280
0.2050
0.1491
0.1278
0.0667
0.0520
0.0420
0.0366
0.0201
0.0154
0.0148
0.0124
0.0086
0.0054
0.0047
0.0037
0.0033
0.0024
0.0011
0.0007
0.0003


Root-Mean-Square Total-Sample Standard Deviation
Root-Mean-Square Distance Between Observations


0.2280
0.4330
0.5821
0.7099
0.7766
0.8286
0.8706
0.9072
0.9273
0.9427
0.9575
0.9699
0.9785
0.9839
0.9886
0.9923
0.9955
0.9979
0.9990
0.9997
1.0000
= 1.141862
= 7.400109


Table A-7. Cluster History
NCL ------Clusters Joined------


CL22
CL12
CL15
CL17
CL9
CL22
CL6
CL4
CL3
CL2


FREQ SPRSQ RSQ ERSQ


CL29
CL13
CL31
CL11
CL15
CL23
CL7
CL10
CL8
CL5


0.0250
0.0258
0.0286
0.0330
0.0466
0.0550
0.0887
0.0942
0.1260
0.1472


.645
.619
.591
.558
.511
.456
.367
.273
.147
.000


.605
.583
.558
.530
.497
.458
.400
.320
.188
.000


CCC PSF PST2 e
2.70 13.1 4.3
2.42 13.4 5.4
2.13 13.8 5.2
1.79 14.3 3.8
0.78 14.4 9.1
-.12 14.7 8.4
-1.7 13.7 13.6
-2.3 13.5 11.7
-2.2 12.6 13.1
0.00 12.









Table A-30. OLS regressions of root transformed restraint with specialty care as ref. group
Coef. Std. Err. t P>t
Cost leader 0.005 0.004 1.22 0.22
Rehab focus -0.006 0.005 -1.32 0.19
Differentiator -0.002 0.005 -0.41 0.68
Lack of strategy -0.001 0.004 -0.15 0.88
Ownership 0.020 0.004 5.50 0.00
Percent Medicaid 0.000 0.000 4.06 0.00
Percent Medicare 0.000 0.000 1.16 0.25
Market competition 0.053 0.070 0.76 0.45
Occupancy -0.010 0.012 -0.84 0.40
Total beds 0.000 0.000 7.90 0.00
Chain -0.007 0.003 -2.24 0.03
Acuity index 0.011 0.002 6.07 0.00
ADL index 0.016 0.004 3.80 0.00
Tube feed care -0.049 0.041 -1.20 0.23
Respiratory care -0.012 0.022 -0.54 0.59
Suctioning care -0.138 0.101 -1.37 0.17
IV therapy care 0.028 0.054 0.51 0.61
Tracheostomy care 0.060 0.110 0.54 0.59
Need injection -0.055 0.018 -3.12 0.00
cons 0.002 0.023 0.08 0.93


Table A-31. OLS regressions of root transformed LTC skin
Coef. Std. Err.
Cost leader -0.005 0.002
Rehab focus 0.008 0.002
Differentiator -0.004 0.003
Lack of strategy -0.008 0.002
Ownership 0.006 0.002
Percent Medicaid 0.000 0.000
Percent Medicare 0.001 0.000
Market competition -0.001 0.029
Occupancy -0.035 0.006
Total beds 0.000 0.000
Chain -0.002 0.002
Acuity index 0.008 0.001
ADL index 0.007 0.002
Tube feed care 0.109 0.024
Respiratory care 0.031 0.011
Suctioning care 0.045 0.060
IV therapy care 0.150 0.032
Tracheostomy care -0.031 0.065
Need injection 0.046 0.010
cons 0.155 0.012


ulcer with specialty
t
-2.50
3.34
-1.47
-3.97
3.32
-0.71
6.94
-0.04
-5.55
3.31
-1.57
8.42
3.11
4.60
2.74
0.74
4.70
-0.47
4.74
13.38


care as ref. group
P>t
0.01
0.00
0.14
0.00
0.00
0.48
0.00
0.97
0.00
0.00
0.12
0.00
0.00
0.00
0.01
0.46
0.00
0.64
0.00
0.00









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

AN EXAMINATION OF STRATEGIC GROUP MEMBERSHIP AND TECHNOLOGY IN
THE NURSING HOME INDUSTRY

By

Alexandre Laberge

May 2009

Chair: Robert Weech-Maldonado PhD
Major: Health Services Research, Management and Policy

Purpose: The purpose of the study was to examine if strategic groups in the nursing home

industry can be determined by examining how facilities apply their technology and how facilities

commit their resources. The study also determined if these groups can be defined using Porter's

generic strategies and whether the different groups have varying strategic-performance

relationships.

Methodology: The study began by performing a factor analysis using scope (technology)

and resource commitment variables from the Minimum Data Set (MDS) and the Online Survey

Certification and Reporting (OSCAR) system. The newly defined variables were used in a

cluster analysis to identify the different clusters. The clusters were classified using Porter's

generic strategy model. Once the strategic groups were identified, the study analyzed the

strategic-performance relationships using negative binomial, ordinary least squares regression

(OLS), generalized linear model (GLM) with gamma distribution and ordered logit analytical

tools. Quality dependent variables included quality of care deficiencies, long term care (LTC)

pressure sore prevalence and post acute pressure ulcer incidence, LTC activities of daily living

(ADL) decline, post acute walking improvement, LTC bowel and bladder decline and LTC

restraint. Financial measures included cost per resident day, operating margin and total margin.









Human capital provides the physical care and possesses the knowledge of the facility.

Knowledge is one of the most critical resources that a firm possesses (Hitt et al. 1991).

The nursing home requires adequate resources if they are going to be able to provide their

technology effectively. The correlation between staffing levels and technology provided is

expected to be high. The amount of nursing staff and the skill of the nursing staff will influence

the quality outcomes (Harrington et al. 2000, Weech-Maldonado et al. 2003) and will likely

influence the quality of the technology that the facility provides.

Nursing resource: The total nursing staff levels of a nursing home is measured by adding

all the registered nurses (RN) per resident, the licensed practical nurses (LPN) per resident and

the certified nursing assistant (CNA) per resident. Although the duties for each of these staffing

types are well defined and are intended to be separate, there remains is a certain degree of

overlap. As a result some facilities may chose to substitute higher skilled staff for lower skilled

staff (E.g. CNA for RN or LPN for RN). Using lower skilled staff allows facilities to keep costs

down. However, nursing homes must have an appropriate amount skilled staff in order to provide

quality of care. Facilities who substitute staffing may lose their savings to the cost of poor

outcome. For example, preventing a pressure ulcer from occurring through good care practices is

going to cost less then treating a newly acquired ulcer.

Rehab resource: The total rehab staffing level of a nursing home is measured by adding

all the physical therapists (PT), physical therapy assistant (PTA), occupational therapist (OT),

certified occupational therapy assistant (COTA) per resident and speech pathologist per resident.

The duties for each of these staffing types for each discipline are less well defined than nursing

because therapy assistants can provide almost all the treatments that a therapist can provide. As a

result facilities may chose to substitute more skilled staff for lesser skilled staff.










Harrington, C., Zimmerman, D., Karon, S.L. Robinson, J. & Beutel, P.(2000). Nursing home staffing and its
relationship to deficiencies. Journal of Gerontology: Social Sciences, 55B, S278-S287

Harrington, C., Kovner, C., Mezey, M., Kayser-Jones, J., Burger, S., Mohler, M., Burke, S. & Zimmerman, D.
(2000) Experts recommend minimum nurse staffing standards for nursing homes in the United States. The
Gerontologist: 40, 5-16

Harrington C, Woolhandler S, Mullan J, Carrillo H, Himmelstein DU. Does investor ownership of nursing
homes compromise the quality of care? Am J Public Health. 2001 Sep;91(9): 1452-5.

Hill, C.W. Differentiation Versus Low Cost or Differentiation and Low Cost: A Contingency Framework.
Academy of Management Review 13 (1988): 401-12

Hitt M., Bierman L.,Shimizu K., Kochhar L. Direct and Moderating Effects of Human Capital on Strategy and
Performance in Professional Service Firms: A Resource-Based Perspective Michael A. Hitt, Leonard Bierman,
Katsuhiko Shimizu and Rahul Kochhar The Academy of Management Journal, Vol. 44, No. 1 (Feb., 2001), pp.
13-28

Hofer, C. Schendel, D., Strategy Formulation: Analytical Concepts, West Publishin, St. Paul, 1978.

Harrigan, C An Application of Clustering for Strategic Group Analysis. Strategic Management Journal; Jan-
Mar85, Vol. 6 Issue 1, p55-73, 19p

Hulin, C., Roznowski, M., (1985) Organizational Technologies: Effect on organizational characteristics and
individuals responses." Research in Organizational behavior, vol7, p.39-85

Hunt, M.S. "Competition in the Major Home Appliance Industry, 1960-1970." Unpublished Doctoral
dissertation, Harvard, 1972.

Instititute of Medicine, 1986. Improving the quality of care of nursing homes. Washington DC: National
Academy Press

Johnson CE, Dobalian A, Burkhard J, Hedgecock DK, Harman J. Predicting lawsuits against nursing homes in
the United States, 1997-2001. Health Serv Res. 2004 Dec;39(6 Pt 1):1713-31.

Kachigan S.K., Multivariate Statistical analysis: A conceptual introduction, Radius, New York, 1991

Kaiser, H. F. (1960). The application of electronic computers to factor analysis. Educational and Psychological
Measurements, 20, 141-51

Kaplan, and Norton. 2001. Transforming the balanced scorecard from performance measurement to strategic
management: Part I. Accounting Horizons (March): 87-104

Katz, R., Cases and Concepts in Corporate Strategy, Prentice-Hall, Englewood Cliffs, N.J. 1970

Ketchen DJ, Shook CJ. The Application of Cluster Analysis In Strategic Management Research: An Analysis
and Critique. Strategic Management Journal 17 (1996): 441-58

Kling, J A.; Smith, K A. Identifying Strategic Groups in the U.S. Airline Industry: An Application of the Porter
Model. Transportation Journal, Winter95, Vol. 35 Issue 2, p26-34, 9p











Table A-2. Continued
Post acute no skin treatment 0.01 -0.08 -0.04 -0.79
Post acute other skin tx 0.03 0.19 -0.01 0.32
Post acute ointment -0.03 0.16 0.02 0.04
Post acute pressure relief bed 0.01 0.04 0.03 0.84
Post acute pressure relief chair 0.10 0.14 -0.04 0.65
Post acute dressing application -0.04 0.07 0.01 0.03
Post acute surgical wound care 0.01 -0.06 0.00 0.12
Post acute ulcer care 0.00 0.07 -0.05 0.01
Post acute skin nutrition program 0.06 0.27 -0.03 0.09
Post acute turning repositioning 0.07 0.07 -0.06 0.26
Post acute OT minutes -0.14 -0.13 0.14 0.06
Post acute PT minutes -0.10 -0.19 0.14 0.08
Post acute other restorative care 0.23 0.04 0.11 -0.02
Post acute restorative communication 0.49 -0.03 0.14 0.01
Post acute restorative amputation 0.31 -0.02 -0.05 0.03
Post acute restorative eating 0.65 -0.03 0.29 0.03
Post acute restorative dressing 0.86 -0.01 0.26 0.00
Post acute restorative transfers 0.90 -0.02 0.19 0.01
Post acute restorative bed mobility 0.88 -0.03 0.15 0.01
S Post acute restorative splinting 0.30 -0.04 0.01 0.03
Post acute restorative Passive ROM 0.58 -0.02 -0.04 0.00
Post acute restorative Active ROM 0.65 0.01 0.01 0.01
Post acute restorative ambulation 0.79 0.02 0.10 0.01
Values over 0.4 are highlighted and represent the initial variables used to define the factor


-0.31
0.31
0.66
0.01
-0.06
0.77
0.42
0.71
0.52
0.39
0.05
0.10
-0.03
0.02
0.04
0.05
-0.03
-0.03
0.00
0.09
0.03
-0.07
-0.08


-0.03 -0.02
-0.06 0.00
-0.02 0.12
0.05 0.03
-0.09 0.06
0.06 0.07
0.15 -0.07
-0.03 0.15
-0.13 0.00
0.01 0.02
0.69 -0.10
0.67 -0.08
-0.02 0.18
-0.01 -0.12
0.01 0.12
0.00 -0.01
-0.01 -0.01
-0.03 0.02
-0.01 0.00
-0.03 0.52
-0.10 0.54
-0.08 0.46
-0.08 0.15


-0.09 -0.02
0.20 0.05
0.03 0.02
-0.04 -0.01
0.13 0.06
-0.07 -0.03
0.08 -0.12
-0.13 0.05
0.17 0.14
0.35 0.02
0.06 -0.10
0.11 -0.12
0.07 -0.01
-0.07 0.00
-0.01 -0.01
0.02 0.01
0.04 0.01
0.05 -0.01
0.03 0.00
0.05 0.01
0.02 0.02
0.05 0.01
0.10 -0.01


-0.08
0.15
0.00
-0.06
-0.02
0.09
0.27
-0.13
-0.16
-0.18
-0.02
0.02
-0.19
-0.01
0.00
0.02
0.04
0.02
0.00
0.01
-0.03
-0.03
0.05







therapy beds, flotation, water, or bubble mattress or pad placed on the bed. Include pressure

relieving, pressure reducing, and pressure redistributing devices.

Turning /Repositioning: This is a continuous facility level variable that represents the

proportion of residents in the nursing home that are placed on a turning-repositioning schedule.

Turning-repositioning schedules are an important intervention to prevent a new pressure sore

from occurring or to prevent an already acquired pressure sore from worsening. It includes a

continuous, consistent program for changing the resident's position and realigning the body.

"Program" is defined as "a specific approach that is organized, planned, documented, monitored,

and evaluated."

Ulcer care: This is a continuous facility level variable that represents the proportion of

residents in the nursing home that are receiving ulcer care treatment. Residents who get ulcer

care already have ulcers.

Surgical care: This is a continuous facility level variable that represents the proportion of

residents in the nursing home that are receiving surgical care treatment. Residents who get

surgical care already have ulcers.

Ointment: This is a continuous facility level variable that represents the proportion of

residents in the nursing home that are receiving ointment care. It includes ointments or

medications used to treat a skin condition (e.g., cortisone, antifungal preparations,

chemotherapeutic agents, etc.).

Dressing care: This is a continuous facility level variable that represents the proportion of

residents in the nursing home that are receiving dressing care. Residents who get dressing care

already have ulcers.

N\ki nutrition program care: This is a continuous facility level variable that represents the

proportion of residents in the nursing home that are on a skin nutrition program. Dietary

measures received by the resident for the purpose of preventing or treating specific skin

47










BIOGRAPHICAL SKETCH

Alexandre Laberge completed his honors bachelors of physical therapy at the University of

Western Ontario in London, Ontario in 1995. He worked as a physical therapist in a variety of settings

including hospitals, skilled nursing facilities, assisted living facilities and in the home health settings.

He completed his Masters of Business Administration in information technology at Goldey Beacom

College in 2002. He completed his Ph.D. in health services research, management and policy in May in

2009. He worked in skilled nursing facilities for over 6 years as a therapist as well as a rehab manager.

In the 4.5 years that he worked as a research assistant, he worked closely with Dr Weech-Maldonado

on a variety of projects that examined the relationship between cost, quality and financial performance

in nursing homes.







K-means cluster analysis was specifically designed for large datasets (Rudolph et al. 1991).

Since this study began with over 11500 observations, K means cluster analysis was the most

appropriate. K-means cluster analysis requires the researcher to establish the number of clusters

apriori. This presents a challenge because it is not known if the number of clusters selected a

priori is the most well defined clusters. Using theory to determine the number of clusters is

considered to be a valid method to identify the k-means apriori (Ketchen et al. 1996). However,

Ketchen et al. (1994) also recommended using a quantitative method in conjunction with the

inductive method to increase the robustness of the number of clusters selected. One such method

is the Ward's minimum variance method which has been used by numerous strategic group

studies (Marlin et al. 1999, Ford et al. 1998, Zinn et al. 1994, Short et al. 2002).

One problem with the Ward's minimum variance method is that it functions poorly with

very large datasets and it is very susceptible to outliers. FASTCLUS was used as technique to

reduce the variables into a smaller number of clusters that would be more manageable for the

Ward's minimum variance method (SAS 1990). FASTCLUS uses k-means cluster analysis to

reduce the number of observations to a level that the Wards minimum variance can manage. In

this study, the initial FASTCLUS reduced the dataset to 100 cluster/observations. The study also

removed outlier observations that were greater than five standard deviations.

Ward's minimum variance method was used for determining the number of clusters to be

used. The study used the following decision criteria to identify the optimal cluster solutions

(SAS 1990).

1. a local peak in the cubic clustering criterion

2. a local peak in the pseudo F statistic,

3. a small value of the pseudo t2 statistic and a larger pseudo t2 statistic for the next cluster
solution level,

4. an additional cluster increasing the overall fit by less than 5 percent, and

5. The clusters obtained explaining at least 65 percent of the overall variance.

62









Table A-24. GLM of PA walking improve with gamma distribution with rehab focus as ref group
Coef. Std. Err. z P>z
Cost leader -0.019 0.018 -1.09 0.28
Differentiator -0.050 0.022 -2.32 0.02
Specialty care focus -0.014 0.019 -0.73 0.47
Lack of strategy 0.010 0.018 0.57 0.57
Ownership -0.087 0.014 -6.04 0.00
Percent Medicaid -0.004 0.000 -10.18 0.00
Percent Medicare -0.002 0.001 -2.81 0.01
Market competition -0.425 0.248 -1.71 0.09
Occupancy -0.160 0.043 -3.73 0.00
Total beds 0.000 0.000 -0.53 0.60
Chain -0.032 0.012 -2.74 0.01
Acuity index -0.028 0.007 -4.07 0.00
ADL index -0.110 0.017 -6.49 0.00
Tube feed care -0.666 0.143 -4.66 0.00
Respiratory care 0.139 0.081 1.71 0.09
Suctioning care 0.444 0.419 1.06 0.29
IV therapy care 0.240 0.213 1.13 0.26
Tracheostomy care -0.238 0.449 -0.53 0.60
Need injection 0.038 0.068 0.56 0.58
Ifips_2 0.218 0.283 0.77 0.44
Ifips_3 0.209 0.188 1.11 0.27


Table A-25. OLS regressions of root
Coef.
Cost leader -0.030
Differentiator 0.015
Specialty care focus -0.012
Lack of strategy -0.061
Ownership 0.034
Percent Medicaid 0.001
Percent Medicare 0.001
Market competition -0.031
Occupancy 0.014
Total beds 0.000
Chain -0.021
Acuity index 0.034
ADL index 0.027
Tube feed care 0.568
Respiratory care 0.088
Suctioning care 0.535
IV therapy care 0.270
Tracheotomy care -0.160
Need injection 0.183
cons -2.528


transformed PA skin
Std. Err.
0.016
0.021
0.017
0.016
0.014
0.000
0.001
0.308
0.046
0.000
0.012
0.007
0.019
0.156
0.082
0.413
0.194
0.431
0.069
0.097


ulcers with rehab
t
-1.91
0.74
-0.69
-3.80
2.38
3.39
0.70
-0.10
0.31
5.08
-1.76
4.88
1.42
3.64
1.07
1.30
1.40
-0.37
2.63
-26.06


focus as ref.


group
P>t
0.06
0.46
0.49
0.00
0.02
0.00
0.48
0.92
0.75
0.00
0.08
0.00
0.16
0.00
0.28
0.20
0.16
0.71
0.01
0.00







and tacit knowledge to improve the quality of patient care through their involvement in

establishing processes and supervising other physical therapy staff

OTFTE /Total Residents: Occupational Therapist (OT) per resident is a continuous

variable represents the ratio total number of OT FTEs divided by the total number of residents in

the facility. OT evaluates the patient and oversees the plan of care. They supervise the Certified

Occupational Therapy Assistant

COTA FTE /Total Residents: Certified Occupational Therapy Assistant (COTA) per

resident is a continuous variable represents the ratio total number of COTA FTEs divided by the

total number of residents in the facility. COTAs provide the treatment to the patients under the

supervision of the occupational therapist.

Percent OT: Percent of the occupational therapy staff that are OTs is a continuous staffing

mix variable. Higher OT staffing mixes are advantageous since OTs can use their professional

skills and tacit knowledge to improve the quality of patient care through their involvement in

establishing processes and supervising other occupational therapy staff

Wages: Nursing homes average wages is a continuous variable derived by dividing the

total salary expense by the total number of working hours reported in the Medicare costs reports.

Wages are considered resources measures because nursing homes utilize the wages to control

their cost.

Hypothesis #1 Analysis

Since there was a total of 68 variables of interest (technology and resource variables), it

was necessary to reduce the number of variables into composite variables by using factor

analysis. The internal consistency of the newly created composite variables was tested using

Cronbach's (1951) alpha methodology. Some variables that were not included in the composite

scores were still included in the cluster analysis as individual variables.







variability within clusters and 2) maximize variability between clusters. In other words, the

similarity rules apply maximally to the members of one cluster and minimally to members

belonging to the rest of the clusters. Cluster analysis can also be described as a set of techniques

for partitioning a set of objects into relatively homogenous subsets based on inner objects

similarities. The process begins by measuring a set number of objects (nursing homes) on each of

the variables of interest. A measure of difference between each pair of nursing homes is

obtained. Porter's competitive strategy typology was used to define the strategic groups of

nursing homes. Similar to Marlin et al. 1999 and Zinn et al. 1994, this study used the scope and

resource commitment measures in the cluster analysis.

Prior studies argue that due to the large interstate variation in legislation, they could only

use one state for their analysis (Zinn et al. 1994, Marlin et al. 1999). Although it is true that

states have different governing laws, different reimbursement systems and different survey

methodologies, limiting the analysis to one state limits the generalizability of the results. Rather

than limiting to one state, this study standardized all variables using z-score based on the mean

of the nursing home market (county). Regulation and reimbursement may vary not only by state

but also by county. Similarly income and demographics can vary by county. All these factors

may influence the nursing home's strategy on how to use their resources and how much

technology to provide. Therefore it is necessary to compare nursing homes to others in their

county. Z-score or normal score is a dimensionless quantity derived by subtracting the

population mean from an individual raw score and then dividing the difference by the population

standard deviation. In this study, the county mean and the county standard deviation were used.

This study used an inductive methodology to determine the strategic group clusters. K-

means is a common method used to determine clusters using inductive reasoning (Ketchen 1996,

Harrigan 1981). A significant number of studies have used k-means to identify the strategic

groups (Marlin et al. 1999, Ford et al. 1998, Zinn et al. 1994, Short et al. 2002, Dess et al. 1984).

61









to be small, managers decide not so much how to be unique, but rather which group of

competitors their strategies should be similar to (Marlin et al. 2002). The strategic group model

is one approach that can help managers and researchers identify the best strategy.

Purpose

The purpose of the study is to examine if strategic group structure of the nursing home

industry can be determined by how facilities focus their technology and how facilities commit

their resources. The study will also determine if these groups can be defined using Porter's

generic strategies and will examine whether the different groups have different strategic-

performance relationships.

Research Questions

What strategic groups exist in the nursing home industry?

How does nursing home strategic group membership affect quality of care?

How does nursing home strategic group membership affect costs?

How does nursing home strategic group membership affect financial performance?









Table A-18. Negative binomial of quality of care deficiencies with rehab focus as reference group
IRR Std. Err. z P>z
Cost leader 1.067 0.032 2.14 0.03
Differentiator 1.007 0.038 0.18 0.86
Specialty care focus 1.008 0.033 0.25 0.80
Lack of strategy 1.132 0.035 4.03 0.00
Ownership 1.099 0.029 3.59 0.00
Percent Medicaid 1.005 0.001 7.69 0.00
Percent Medicare 1.005 0.001 3.75 0.00
Market competition 0.773 0.367 -0.54 0.59
Occupancy 0.830 0.065 -2.38 0.02
Total beds 1.003 0.000 14.74 0.00
Chain 1.024 0.021 1.18 0.24
Acuity index 1.040 0.013 3.20 0.00
ADL index 0.952 0.029 -1.63 0.10
Tube feed care 1.009 0.237 0.04 0.97
Respiratory care 0.972 0.157 -0.18 0.86
Suctioning care 0.811 0.454 -0.37 0.71
IV therapy care 3.826 1.304 3.94 0.00
Tracheostomy care 1.142 0.745 0.20 0.84
Need injection 1.189 0.141 1.46 0.14
Ifips_2 1.206 0.470 0.48 0.63
Ifips_3 1.025 0.198 0.13 0.90


Table A-19. OLS regressions of root
Coef.
Cost leader -0.006
Differentiator -0.012
Specialty care focus -0.003
Lack of strategy -0.007
Ownership 0.006
Percent Medicaid 0.000
Percent Medicare 0.001
Market competition 0.077
Occupancy 0.014
Total beds 0.000
Chain 0.004
Acuity index 0.009
ADL index 0.021
Tube feed care -0.061
Respiratory care -0.040
Suctioning care 0.097
IV therapy care 0.004
Tracheostomy care -0.245
Need injection 0.004
cons 0.253


transformed bladder
Std. Err.
0.003
0.004
0.003
0.003
0.002
0.000
0.000
0.064
0.008
0.000
0.002
0.001
0.003
0.027
0.014
0.067
0.040
0.074
0.012
0.017


decline with rehab
t
-1.97
-3.27
-0.86
-2.32
2.46
-5.35
3.84
1.21
1.70
3.66
1.97
6.66
6.68
-2.24
-2.99
1.44
0.09
-3.32
0.33
14.72


focus as ref.


group
P>t
0.05
0.00
0.39
0.02
0.01
0.00
0.00
0.23
0.09
0.00
0.05
0.00
0.00
0.03
0.00
0.15
0.93
0.00
0.74
0.00









pay nursing homes: retrospective, prospective class, prospective facility-specific, adjusted, and

combination.

Retrospective payments freely adjust rates to current costs. In contrast, prospective

payment rates are not fully adjusted to cost during the costs year. Rates are usually set based on

prior year costs. Prospective class rates are the same for all nursing homes in the state while

prospective facility-specific are set based on the costs for each facility. Adjusted reimbursement

is similar to retrospective reimbursement however adjusted systems allow interim rates to

increase during a year but not to fully reflect costs. Finally, there are also combination

reimbursement methods which are a mixture of retrospective and prospective payment systems

(Swan et al. 2000). As of 1997, only one state used retrospective reimbursement, 25 used

prospective, 21 used adjusted and 3 used combinations (Feng et al. 2006).

Since the late 1990s, most state Medicaid programs have used a prospective per diem

system to reimburse nursing homes for all the cost that is incurred in providing the care to the

resident in the facility (Feng et al. 2008). With this reimbursement methodology, a nursing home

that provides a lot of nursing services will not necessarily be compensated more than a facility

that provides fewer nursing services.

An increasing number of states have modified their reimbursement methodology by

making case mix adjustments to their per diem rate. This allows nursing homes to collect higher

rates for residents that have higher acuity and that are likely to cost the facility more to care for

them. The number of states that use case mix reimbursement increased from 19 in 1991 to 35 in

2004 (Feng et al. 2006).

The best-known and most widely used case mix methodology is the Resource Utilization

Groups system (RUGs), currently in its third version (Fries et al. 1994). This system classifies








All the above quality variables with skewness under 4 were analyzed using OLS with

county level fixed effects. The quality variables that had skewness over 4 were initially analyzed

by using the generalized linear model (glm) with gamma distribution and a log link.

Unfortunately, the glm was not able to find a solution when using post acute pressure ulcer

incidence quality variable. Therefore the PA pressure ulcer incidence was analyzed using OLS

with fixed effects.

Hypothesis #2

Hypothesis #2 was partially supported. Table 4-5 (full regression results are located in tables

A-9 to A-16 in the appendix) illustrates that the differentiator strategic group had either the same

or better long term care quality than all other strategic groups in all measures except long term

care pressure ulcers. Differentiators had lower 4 point ADL decline than cost leaders and those

nursing homes that lacked a strategy. Differentiators also had fewer pressures sores than nursing

homes that had a rehab focus or had a specialty care focus. However, with respect to the post

acute measures, differentiators performed poorly. The differentiator strategic group generally had

worse outcomes with less walking improvement than all other groups. Differentiators had higher

incidence of post acute pressure sore than lack of strategy and cost leaders.

Table 4-5. Strategic group quality of care with Differentiator / nursing care focus cluster as the
reference group
Dependent Variables Cost Rehab focus Specialty care Lack of Strategy

Quality of care deficiencies 1.06 (0.09)* 0.99 (0.86) 1.00 (0.97) 1.12 (0.00)*
Long Term care measures

Bladder Decline 0.006 (0.04)* 0.012 (0.00)* 0.009 (0.01)* 0.005 (0.12)
Bowel Decline 0.008 (0.02)* 0.009 (0.03)* 0.012 (0.00)* 0.005 (0.16)
ADL 4 pt decline 0.008 (0.00)* 0.003 (0.39) 0.005 (0.10)* 0.004 (0.13)
Restraint use 0.011 (0.01) 0.004 (0.46) 0.001 (0.19) 0.001 (0.19)
Pressure Ulcers -0.001 (0.66) 0.012 (0.00)* 0.004 (0.14) -0.004 (0.09)
Post acute measures
Ambulation improvement 0.031 (0.10)y 0.050 (0.02)y 0.037 (0.08)y 0.060 (O.OO)y
Pressure Ulcers incidence -0.045 (0.02)y -0.015 (0.46) -0.027(0.20) -0.076 (O.OO)y
Quality of care results are incidence rate ratio
Square root transformation LTC bladder decline, bowel decline pressure ulcer prevalence, LTC 4 point ADL decline, restraints, Post acute pressure
sore incidence are OLS regression coefficients. Post acute walking improvement was glm with gamma distribution. Represents significant values
in the desired direction. y Represents significant values in the negative direction.









nursing home industry, measures of this type of performance include percent of beds filled by

Medicare patients and percent of beds filled by private pay residents. Both Medicare and private

pay clientele pay higher reimbursement rates than other payers like Medicaid. Market shares of

Medicare and/or private pay residents are the result of a strategy but do not give any information

about what strategy was used by the nursing home to accomplish that performance.

Unlike the aforementioned performance measures, technology focus and resource

deployment decisions are made by the business level management and are under the direct

control of management. Decisions to increase or decrease the amount of nursing technologies

provided (services like restorative care or skin ulcer prevention) can be made independent of the

environment that the facility is located in. However the resulting quality of care caused by

providing these services can influence the performance of the facility as measured by the ability

of the facility to draw higher reimbursing Medicare or private pay residents.

In the nursing home industry, technology includes all the processes of care that are

provided to the residents by the facility's healthcare staff. Nursing homes use their technology to

transform their inputs into outputs. The nursing home industry is different from other industries

because of the way technology is applied and the way the product is measured.

Resource Commitment

Resource commitment refers to the commitment of resources to functional areas that are

needed to gain and maintain competitive advantage in targeted market segments. For nursing

home administrators, resource commitment decisions should be reflected by labor, price and

capacity decisions (Zinn et al. 1994). Nursing homes have some control over their labor but they

have limited control over their products' price. Government reimbursement accounts for over

62% of their revenue (MedPAC 2007). After the implementation of the Balanced Budget Act of

1997, Medicare dictated the price that they are going to pay for the nursing home stay. Similarly










Laberge A, Weech-Maldonado R, Johnson CE, Jia H, Dewald L.Outsourcing veterans for long-term care:
comparison of community and state veterans' nursing homes. J Health Hum Serv Adm. 2008 Spring;30(4):441-
67.

Leask, G.; Parker, D. Strategic groups, competitive groups and performance within the U.K. pharmaceutical
industry: Improving our understanding of the competitive process. Strategic Management Journal, Jul2007,
Vol. 28 Issue 7, p723-745

Marlin D, Sun M, Huonker JW. Strategic groups and performance in the nursing home industry: a
reexamination.Med Care Res Rev. 1999 Jun;56(2):156-76

Marlin D, Huonker JW, Sun M. An examination of the relationship between strategic group membership and
hospital performance. Health Care Manage Rev. 2002 Fall;27(4):18-29

Mukamel DB, Glance LG, Li Y, Weimer DL, Spector WD, Zinn JS, Mosqueda L. Does risk adjustment of the
CMS quality measures for nursing homes matter? Med Care. 2008 May;46(5):532-41.

Mehra, Ajay. RESOURCE AND MARKET BASED DETERMINANTS OF PERFORMANCE IN THE U.S.
BANKING INDUSTRY. Strategic Management Journal, Apr96, Vol. 17 Issue 4, p307-322

Murray, A.L. "A Contingency view of porter's generic strategies." Academy of Management Review 13
(1988): 390-400

National Institute of Health, http://www.nlm.nih.gov/medlineplus/nursinghomes.html, accessed March 20th,
2008

O'Neill C, Harrington C, Kitchener M, Saliba D. Quality of care in nursing homes: an analysis of relationships
among profit, quality, and ownership. Med Care. 2003 Dec;41(12): 1318-30.

Pagano E, Bo S, Petrinco M, Rosato R, Merletti F, Gregori D. Factors affecting hospitalization costs in Type 2
diabetic patients. J Diabetes Complications. 2009 Jan-Feb;23(1):1-6. Epub 2008 Apr 16.

Pfeffer, J. 1994. Competitive advantage through people. Boston: Harvard Business School Press

Porter, M.E., "Competitive Strategy, New York: Free press, 1980

Rhoades J, Sommers J., Expenses and Sources of Payment for Nursing Home Residents 1996. Rockville, MD:
Agency for Healthcare Research and Quality; 2000. MEPS Research Findings No 13. AHRQ Pub. No 01-0010

SAS manual 1990, SAS Technical Report A-108

Scanlon, W. 1980. "A theory of the nursing home market." Inquiry 17 (1):25-41.

Schnelle JF. Determining the relationship between staffing and quality. Gerontologist. Feb 2004;44(1): 10-12

Shen YC, Eggleston K, Lau J, Schmid CH.Hospital ownership and financial performance: what explains the
different findings in the empirical literature? Inquiry. 2007 Spring;44(1):41-68

Short JC, Palmer TB, Ketchen DJ Jr. Resource-based and strategic group influences on hospital performance.
Health Care Manage Rev. 2002 Fall;27(4):7-17.







Percent Medicare beds: Facility level variable that represents the percent of the total

number of resident beds reimbursed by Medicare. Medicare only reimburses for resident who are

at the facility for short term stay. This variable is used because it is necessary to account for the

proportion of the patients that are receiving rehabilitation at the facility. It is generated by

dividing number of Medicare residents by the total number of beds for that facility.

Resource Commitment

Resource commitment refers to the commitment of resources to the different technology

areas that are needed to gain and maintain competitive advantage in targeted market segments

(Zinn et al. 1994). The primary resource is staffing levels (the nurses and the therapist) since it is

the health care staff that provides the technology. Therefore, organizations manipulate their

staffing levels to control costs and ensure quality in an attempt to perform well financially.

This study divided the resource commitment into two separate subtypes because they are

mutually exclusive of each other and do not overlap.

Resource commitment nursing: Resource commitment nursing subtype includes all the

licensed and non licensed nursing staff. They are responsible for all the nursing technologies

RNFTE / Total Residents: Registered Nurse (RN) per resident is a continuous variable

representing the ratio of the total number of RN FTEs divided by the total number of residents in

the facility. RNs are involved in establishing processes and supervising other nursing staff.

LPNFTE /Total Residents: Licensed practical nurse (LPN) per resident is a continuous

variable representing the ratio of the total number of LPN FTEs divided by the total number of

residents in the facility. Both measures are from the OSCAR dataset. LPNs' are able to provide

some skilled nursing processes of care like dispensing medication. They have to be under the

supervision of a nurse.

CNA FTE / Total Residents: Certified Nursing Assistant (CNA) per resident is a

continuous variable representing the ratio of the total number of CNA FTEs divided by the total

54







Financial Performance

The study found that rehab focus, nursing care focus group and cost leaders had better

financial performance than the nursing homes that lacked a strategy and the specialty care focus

group. The poor financial performance of the specialty care focus group could be related to higher

costs of having more RNs without the benefit of generating higher revenue streams. Unlike the

rehab group where the more therapy results in more revenue, the specialty care focus groups may

have limited ability to increase revenue. Medicare and some Medicaid programs can (depending

on the state) provide more payment for nursing care; however the reimbursement is provided

based on the utilization of services and not based on the staffing levels of the facility. In addition,

these groups provided a care to patients with highly specific and costly conditions. MedPAC 2007

has expressed concern that the current Medicare reimbursement system does not cover the costs of

care adequately to patients with needs over than therapy. Rehab focus nursing homes would be

able to generate the greatest amount of revenues since they had the most Medicare rehab patients.

This would explain why even though the rehab focus group had some of the highest costs, they

still managed to be one of the best financial performers. These results are providing some support

with prior research that found that nursing homes with higher Medicare residents and private pay

residents had the best financial performance (Zinn et al. 1994, Marlin et al. 1999). However, in

this study there were other groups that also performed well. Differentiators managed to have good

financial performance without a high degree of rehab. One explanation is that by 2004, most states

through Medicaid used case mix reimbursement that would compensate facilities that provided

certain types of nursing technology (e.g. restorative care). Nursing homes that provide services

that can generate revenue and lower costs via better quality can increase their odds of achieving

better financial performance.

The hypothesis of this study stated that nursing homes that lacked a strategy would not

perform as well as nursing homes that had a strategy. The results suggest that this is true with one

94










Table A-4. Eigenvalues of scope (technology) factor analysis


1
2
3
4
5
6
7



3.50

3.00

2.50

2.00

1.50

1.00

0.50

0.00


Eigenvalues of the Correlation Matrix:
Eigenvalue Difference Proportion
3.16 1.77 0.45
1.39 0.32 0.20
1.08 0.43 0.15
0.65 0.30 0.09
0.35 0.10 0.05
0.24 0.11 0.03
0.13 0.02


Cumulative
0.45
0.65
0.80
0.90
0.95
0.98
1


0 1 2 3 4 5 6 7 8


Figure A-2. Scree plot of resource staffing intensity from factor analysis


Table A-5. Factor analysis of resource staffing intensity with three factors
Total Total Total
therapist therapy nursing
assistants
Total PT per Resident 0.94987 0.13329 0.09479


Total OT per Resident
Total PTA per Resident
Total OTA per Resident
RN per Resident
LPN per Resident
CNA per Resident
Values over 0.4 are higl


0.93244
-0.02059
0.39775
0.0587
0.15186


0.10457
0.92192
0.78854
0.03395
0.56287


0.13311
0.05157
0.21303
0.79055
0.61385


0.13473 0.15559 0.86025
lighted and represent the initial variables used to define the factor


-*- Seriesi


7







position to another in bed. It is a routine service and does not require the skill of the therapist.

However the goals of the program need to be generated by the therapist.

Restorative dressing: The restorative dressing variable is facility level and is continuous. It

represents the average number of days that a resident receives restorative dressing while

remaining in the facility. It is generated by dividing the resident level restorative variable that

represents the number of days of dressing provided in the 7 days before the assessment date by

the total number of residents in the facility. Restorative dressing have the nursing staff help dress

the residents in order to improve the resident's ability to dress by themselves. It is a routine

service and does not require the skill of the therapist. However the goals of the program need to

be generated by the therapist.

Restorative eating: The restorative eating variable is facility level and is continuous. It

represents the average number of days that a resident receives restorative eating while remaining

in the facility. It is generated by dividing the resident level restorative variable that represents the

number of days of eating provided in the 7 days before the assessment date by the total number

of residents in the facility. Restorative eating have the nursing staff help the residents with eating

in order to improve the resident's ability to eat by themselves. It is a routine service and does not

require the skill of the therapist. However the goals of the program need to be generated by the

therapist.

Restorative splinting: The restorative splinting variable is facility level and is continuous.

It represents the average number of days that a resident receives restorative splinting while

remaining in the facility. It is generated by dividing the resident level restorative variable that

represents the number of days of splinting provided in the 7 days before the assessment date by

the total number of residents in the facility. Restorative splinting have the nursing staff assist

with the donning and doffing of splints in order to improve the resident's ability to splint







The composite (or index) and the appropriate individual variables were transformed into

cluster variables using z-scores. Wards minimum variance method was used to identify the

number of clusters and k-means cluster analysis to identify the different strategic groups. The

analysis of variance and covariance (ANOVA) statistical test with Tukey's comparison test was

used for each variable of interest to determine if the strategic groups were significantly different

from each other. The Multivariate analysis of variance and covariance (MANOVA) test with

Wilk's Lambda was used to determine if the overall clusters were significantly different from

each other. Each variable was given a weighted rank score for each cluster. The weighted rank

score was then aggregated by type of commitment (scope or resource). Using the aggregated

weighted rank score, the clusters were classified based on Porter's generic strategies as defined

in the technology, resources and strategic group model presented in Figure #2.

Factor Analysis

Factor analysis is a variable reducing technique. It is a family of procedures designed to

remove the redundancy from a set of highly correlated variables and representing the variables in

a smaller set of derived variables or 'factors' (Kachigan 1991). Factor analysis can be seen as the

grouping of similar variables (Kachigan 1991). These groupings of variables (or factors) can be

used to describe a particular construct. For example, in the nursing home industry, two of the

treatments used in an attempt to decrease pressure sores include placing a pressure relieving

device in a chair and placing a pressure relieving device in a bed. It would be expected that these

two processes of care are going to be highly correlated since both processes are similar in nature.

The study used exploratory factor analysis to identify scope commitment and resource

commitment variables that are correlated to each other. These variables were combined into new

index variables that were a measure of what was considered a technology or resource factor. In

the above example, if the factor analysis determined that the two pressure relieving device









The center of the circle depicts the centroid of the cluster. To be in the cluster, a nursing

home is closer to the centroid of its cluster than the centroid of any other cluster in the industry.

All nursing homes within the groups are located within the outer boundaries of each cluster. A

nursing home is not able to be in two separate clusters.

The upper scope boundary is a vertical line that passes through the centroid of the cluster

that has the highest scope (technology) value. The lower resource boundary is a horizontal line

that passes through the centroid that has the lowest resources. There are nursing homes that fall

beyond both these boundaries; however the centroids of the clusters cannot fall beyond these

boundaries.

Differentiator

For a nursing home to differentiate itself from its competitors, it must provide superior

processes of care. By taking on the strategy of providing more service than the competitors in

their local markets, these nursing homes are more likely to attract residents to their facility

resulting in higher occupancy rates and choose residents that can provide them with higher

margins. A differentiating nursing home would provide a high volume of service relative to the

average nursing home in the market. In Figure #2, the differentiating strategic group would be

expected to be in the upper right quadrant of the graph. This is because the differentiating group

is defined as the highest provider of service (technology). If a high volume of service is

provided, it would be expected that facilities in the differentiation strategic groups would also

have to commit a greater amount of resources. The centroid of the cluster demarks the upper

scope boundary.

Cost Leadership

Some nursing homes will be focused on keeping their costs low. Since staffing is the

significant cost driver in the nursing home industry, keeping lower staffing levels is one obvious


































2009 Alexandre Laberge










Post acute LTC skin LTC Pressure Post acute Therapy Toilet Weight
restorative care restorative devices skin care minutes program program
Table A-3. Continued
Post acute no skin treatment 0.01 -0.08 -0.04 -0.79 -0.31 -0.03 -0.09 -0.02
Post acute other skin tx 0.03 0.19 -0.01 0.32 0.31 -0.06 0.20 0.05
Post acute ointment -0.03 0.16 0.02 0.04 0.66 -0.02 0.03 0.02
Post acute pressure relief bed 0.01 0.04 0.03 0.84 0.01 0.05 -0.04 -0.01
Post acute pressure relief chair 0.10 0.14 -0.04 0.65 -0.06 -0.09 0.13 0.06
Post acute dressing application -0.04 0.07 0.01 0.03 0.77 0.06 -0.07 -0.03
Post acute surgical wound care 0.01 -0.06 0.00 0.12 0.42 0.15 0.08 -0.12
Post acute ulcer care 0.00 0.07 -0.05 0.01 0.71 -0.03 -0.13 0.05
Post acute skin nutrition program 0.06 0.27 -0.03 0.09 0.52 -0.13 0.17 0.14
Post acute turning repositioning 0.07 0.07 -0.06 0.26 0.39 0.01 0.35 0.02
Post acute OT minutes -0.14 -0.13 0.14 0.06 0.05 0.69 0.06 -0.10
Post acute PT minutes -0.10 -0.19 0.14 0.08 0.10 0.67 0.11 -0.12
Post acute other restorative care 0.23 0.04 0.11 -0.02 -0.03 -0.02 0.07 -0.01
Post acute restorative communication 0.49 -0.03 0.14 0.01 0.02 -0.01 -0.07 0.00
Post acute restorative amputation 0.31 -0.02 -0.05 0.03 0.04 0.01 -0.01 -0.01
Post acute restorative eating 0.65 -0.03 0.29 0.03 0.05 0.00 0.02 0.01
Post acute restorative dressing 0.86 -0.01 0.26 0.00 -0.03 -0.01 0.04 0.01
S Post acute restorative transfers 0.90 -0.02 0.19 0.01 -0.03 -0.03 0.05 -0.01
Post acute restorative bed mobility 0.88 -0.03 0.15 0.01 0.00 -0.01 0.03 0.00
Post acute restorative splinting 0.30 -0.04 0.01 0.03 0.09 -0.03 0.05 0.01
Post acute restorative Passive ROM 0.58 -0.02 -0.04 0.00 0.03 -0.10 0.02 0.02
Post acute restorative Active ROM 0.65 0.01 0.01 0.01 -0.07 -0.08 0.05 0.01
Post acute restorative ambulation 0.79 0.02 0.10 0.01 -0.08 -0.08 0.10 -0.01
Values over 0.4 are highlighted and represent the initial variables used to define the fact









LIST OF FIGURES


Figure page

2-1 T he nursing hom e process ........................................................................ ...................30

2-2 Strategic Group Model Based on Scope and Resource commitment.............................36

4-1 Strategic group model based on scope and resource commitment in the nursing
hom e indu story ....................................................... .................. 8 1

A-i Scree plot of scope (technology) factor analysis .................................... ............... 101

A-2 Scree plot of resource staffing intensity from factor analysis............... ... ...............106









Hypothesis #4: Nursing homes in the cost leadership strategic group will have lower

costs than other strategic groups.

Hypothesis #5: Nursing homes that lack a strategy will have lower financial performance

than nursing homes that are in a strategic group.









Movement between groups is limited by mobility barriers, which are associated with the

cost of moving from one strategic group to another (Porter 1980). Conceptually, mobility

barriers behave similar to barriers to entry in Porter's five forces analysis framework (Barney

1991). This framework describes entry of new organizations to the market place as threats to the

profitability of the incumbent organizations. By establishing barriers to entry, an organization

can block out new competitors from entering the market, thereby maintaining their above normal

margins (Caves and Porter 1977). The existence of mobility barriers can allow two groups within

the same industry to exist without directly competing with each other (Leask 2007). In the

nursing home industry, facilities that are a part of the strategic groups concerned with efficiency

and keeping operating cost as low as possible are different from those nursing homes that

attempt to differentiate themselves by providing better care (Marlin 1999). Although these

strategies are very different, it is possible that nursing homes in either one of these strategic

group types will have good financial performance.

For an organization to move from one strategic group to another, it would be necessary for

them to change their business strategy and put themselves at risk of not recouping the resources

they invest (Porter 1980). These costs offer protection to group members by discouraging entry

of rivals into the group (Porter 1980). An example would be an organization from a cost

leadership strategic group trying to move to a differentiating strategic group. This organization

would have to make some considerable investment and they would also have to change the

organizational culture from being cost focus to providing more and better services.

The cost of movement between strategic groups can vary. For a facility to move from cost

leadership to focus may require significantly less resources than moving from cost leadership to

differentiator. Therefore the higher the cost to change strategic groups, the higher the mobility









Table A-12. OLS regressions of root transformed bowel decline with differentiators as ref. group
Coef. Std. Err. t P>t
Cost leaders 0.008 0.004 2.31 0.02
Rehab focus 0.009 0.004 2.22 0.03
Specialty care focus 0.012 0.004 2.92 0.00
Lack of strategy 0.005 0.004 1.41 0.16
Ownership 0.005 0.003 1.83 0.07
Percent Medicaid 0.000 0.000 -3.76 0.00
Percent Medicare 0.001 0.000 3.82 0.00
Market competition 0.040 0.050 0.79 0.43
Occupancy 0.013 0.009 1.49 0.14
Total beds 0.000 0.000 2.94 0.00
Chain 0.004 0.002 1.79 0.07
Acuity index 0.011 0.001 7.26 0.00
ADL index 0.025 0.003 7.43 0.00
Tube feed care -0.039 0.029 -1.34 0.18
Respiratory care -0.040 0.015 -2.74 0.01
Suctioning care 0.014 0.077 0.18 0.86
IV therapy care 0.016 0.044 0.36 0.72
Tracheostomy care -0.101 0.085 -1.19 0.23
Need injection 0.020 0.013 1.55 0.12
cons 0.176 0.017 10.11 0.00


Table A-13. OLS regressions


Cost leaders
Rehab focus
Specialty care focus
Lack of strategy
Ownership
Percent Medicaid
Percent Medicare
Market competition
Occupancy
Total beds
Chain
Acuity index
ADL index
Tube feed care
Respiratory care
Suctioning care
IV therapy care
Tracheostomy care
Need injection
cons


of root transformed
Coef.
0.008
0.003
0.005
0.004
0.001
0.000
0.000
0.083
0.006
0.000
-0.001
0.002
-0.003
-0.028
-0.005
-0.054
-0.013
-0.072
-0.008
0.374


ADL decline with
Std. Err.
0.003
0.003
0.003
0.003
0.002
0.000
0.000
0.053
0.007
0.000
0.002
0.001
0.003
0.026
0.013
0.057
0.033
0.066
0.011
0.016


differentiators as ref
t
2.94
0.86
1.67
1.50
0.50
0.10
2.97
1.57
0.84
-0.71
-0.72
1.46
-1.01
-1.10
-0.42
-0.95
-0.39
-1.10
-0.69
23.87


group
P>t
0.00
0.39
0.10
0.13
0.62
0.92
0.00
0.12
0.40
0.48
0.47
0.14
0.31
0.27
0.67
0.34
0.70
0.27
0.49
0.00







conditions e.g., wheat-free diet to prevent allergic dermatitis, high calorie diet with added

supplements to prevent skin breakdown, high protein supplements for wound healing. Vitamins

and minerals, such as Vitamin C and Zinc, which are used to mange a potential or active skin

problem, should be coded here.

Other skin care: This is a continuous facility level variable that represents the proportion

of residents in the nursing home that are receiving other skin care. Residents who get other skin

care already have ulcers.

Restorative technology: All processes of care provided by the nursing staff that involves

having the patient engage in an activity that is directed towards accomplishing an ADL goal that

will result in improved health status.

Restorative ambulation: The restorative ambulation variable is facility level and is

continuous. It represents the average number of days that a resident receives restorative

ambulation while remaining in the facility. It is generated by dividing the resident level

restorative variable that represents the number of days of ambulation provided in the 7 days

before the assessment date by the total number of residents in the facility. Restorative ambulation

has the nursing staff walk the residents in order to improve the distance that the resident can

ambulate. It is a routine service and does not require the skill of the therapist. However the goals

of the program need to be generated by the therapist.

Restorative Passive Range of Motion (PROM): The restorative PROM variable is facility

level and is continuous. It represents the average number of days that a resident receives

restorative PROM while remaining in the facility. It is generated by dividing the resident level

restorative variable that represents the number of days of PROM provided in the 7 days before

the assessment date by the total number of residents in the facility. Restorative PROM has the

nursing staff range the residents in order to improve the PROM of the resident. It is a routine







CHAPTER 3
METHODOLOGY

Data

This study combined the On-line Survey Certification of Automated Records (OSCAR),

the Minimum Data Set (MDS), the Area Resource File (ARF) and the Medicare Cost Reports.

The OSCAR dataset is a facility level dataset that is routinely collected through the Medicare

and Medicaid certification process conducted by state licensure and certification agencies. This

data is updated annually as part of the recertification survey. OSCAR provides a snapshot about

the characteristics and care processes occurring within the nursing homes at the time of the

survey. It includes variables like staffing levels, ownership status, number of residents, facility

acuity level as well as quality outcome measures like state survey care deficiencies and pressure

sore prevalence.

The Omnibus Budget Reconciliation Act of 1987 (OBRA) was the first major legislative

step to improve the quality of care in nursing homes. It established the Resident Assessment

Instrument (RAI) which was implemented in 1991. The RAI is the basis of the Minimum Data

Set (MDS) which is a dataset that records information about nursing home residents. The MDS

includes demographic information, health status as well as amount and types of services that a

resident receives while they reside at the nursing home. Each long term care resident is assessed

when first admitted to a nursing home, then each quarter thereafter. The post acute resident is

assessed on the 5th day, the 14th day, the 30th day, 60th day and 90th day (for as long they remain

in the facility). The CMS MDS is the most comprehensive dataset available for examining

quality outcomes in the long-term care setting. OSCAR and Quality Indicators (QI) generated

from the MDS are used by the CMS as part of its Nursing Home Compare website.

The MDS was acquired using a reuse agreement from the CMS. Data is currently located

behind a firewall in a fully secured server of the College of Public Health and Health Professions













Table A-3. Factor analysis with 10 factors of scope (technology) variables
Rotated Factor Pattern
Post acute LTC skin LTC Pressure Post acute Therapy Toilet Weight
restorative care restorative devices skin care minutes program program
Long term care plateguard -0.01 0.13 0.07 0.02 -0.04 0.01 0.02 0.00
Long term care Weight program -0.01 -0.01 0.04 0.03 0.02 -0.03 0.02 0.92
Long term care Therapy Diet 0.01 0.17 0.07 -0.03 0.03 0.05 -0.05 -0.02
Long term care Bladder training -0.03 0.07 0.08 -0.11 0.02 0.05 0.30 -0.05
Long term care toilet program 0.07 -0.05 0.02 0.20 -0.06 -0.02 0.79 0.05
Long term care no skin treatment 0.00 0.09 0.00 -0.71 -0.24 -0.05 -0.11 -0.01
Long term care other skin tx -0.03 0.63 0.14 0.24 0.05 -0.05 0.08 0.00
Long term care ointment -0.06 0.70 0.18 0.02 0.13 0.02 -0.02 -0.02
Long term care pressure relief bed -0.05 0.57 0.18 0.59 -0.19 0.04 -0.11 -0.05
Long term care pressure relief chair 0.00 0.61 0.12 0.51 -0.20 -0.05 0.00 -0.02
Long term care dressing application -0.06 0.77 0.11 -0.07 0.16 0.13 -0.18 -0.01
Long term care surgical wound care 0.06 0.40 -0.09 -0.07 0.08 0.35 -0.07 0.01
Long term care ulcer care -0.03 0.72 0.07 -0.06 0.15 0.11 -0.20 0.00
Long term care skin nutrition program 0.02 0.60 0.08 0.04 0.21 -0.09 0.12 0.09
S Long term care turning repositioning -0.02 0.69 0.14 0.18 0.03 0.04 0.19 -0.03
Long term care OT minutes 0.03 0.34 -0.11 -0.05 -0.11 0.75 -0.09 0.10
Long term care PT minutes 0.03 0.38 -0.10 -0.05 -0.10 0.74 -0.09 0.09
Long term care other restorative care -0.07 0.15 0.47 0.01 -0.04 0.02 0.08 -0.02
Long term care restorative communication 0.17 0.01 0.39 0.04 0.04 -0.01 -0.07 0.01
Long term care restorative amputation -0.04 -0.03 0.38 0.05 -0.03 -0.02 0.01 0.00
Long term care restorative eating 0.17 0.12 0.72 0.00 0.01 0.06 -0.01 0.03
Long term care restorative dressing 0.30 0.11 0.74 -0.02 0.01 0.02 0.03 0.03
Long term care restorative transfers 0.31 0.14 0.73 0.00 0.00 0.01 0.01 0.00
Long term care restorative bed mobility 0.39 0.06 0.64 0.00 0.03 0.00 0.01 0.03
Long term care restorative splinting 0.02 -0.03 0.18 0.05 0.23 0.14 -0.03 0.00
Long term care restorative Passive ROM 0.12 0.31 0.42 0.03 -0.03 -0.03 0.03 0.00
Long term care restorative Active ROM 0.10 0.28 0.47 0.01 -0.09 0.01 0.05 -0.01
Long term care restorative ambulation 0.05 0.33 0.62 0.02 -0.07 0.03 0.07 -0.02
Post acute plateguard 0.03 0.02 -0.01 0.06 -0.03 -0.14 0.12 0.10
Post acute Weight program -0.02 0.01 0.03 0.03 0.05 -0.03 0.01 0.93
Post acute Therapy Diet -0.01 -0.04 -0.04 0.04 0.10 -0.15 0.02 0.05
Post acute Bladder training -0.01 -0.04 0.06 -0.05 0.14 0.05 0.39 -0.07
Post acute toilet program 0.11 -0.06 -0.05 0.18 -0.03 -0.07 0.80 0.09









APPENDIX
BACKGROUND TABLES

Table A-1. Eigenvalues of scope (technology) factor analysis


Eigenvalue
7.03
4.89
3.76
2.68
2.29
2.21
1.80
1.78
1.54
1.51
1.46
1.37
1.34
1.28
1.21
1.17
1.12
1.06
1.04
1.01
0.95
0.85
0.81
0.77
0.70
0.69
0.63
0.63
0.62
0.60
0.54
0.50
0.47
0.43
0.43
0.41
0.39
0.38
0.36
0.33
0.32
0.30
0.25
0.24
0.21
0.20
0.19
0.18
0.18
0.17


Difference
2.134
1.138
1.074
0.391
0.080
0.416
0.017
0.241
0.031
0.043
0.098
0.024
0.067
0.061
0.048
0.047
0.057
0.022
0.027
0.064
0.101
0.037
0.044
0.066
0.014
0.053
0.009
0.006
0.022
0.060
0.033
0.032
0.038
0.008
0.021
0.021
0.006
0.024
0.026
0.009
0.018
0.050
0.017
0.023
0.011
0.011
0.008
0.005
0.005
0.008


Proportion
0.126
0.087
0.067
0.048
0.041
0.040
0.032
0.032
0.028
0.027
0.026
0.024
0.024
0.023
0.022
0.021
0.020
0.019
0.019
0.018
0.017
0.015
0.015
0.014
0.013
0.012
0.011
0.011
0.011
0.011
0.010
0.009
0.008
0.008
0.008
0.007
0.007
0.007
0.006
0.006
0.006
0.005
0.005
0.004
0.004
0.004
0.003
0.003
0.003
0.003


Cumulative
0.126
0.213
0.280
0.328
0.369
0.408
0.440
0.472
0.500
0.527
0.553
0.577
0.601
0.624
0.646
0.666
0.686
0.705
0.724
0.742
0.759
0.774
0.789
0.802
0.815
0.827
0.839
0.850
0.861
0.871
0.881
0.890
0.898
0.906
0.914
0.921
0.928
0.935
0.941
0.947
0.953
0.958
0.963
0.967
0.971
0.974
0.978
0.981
0.984
0.987







number of residents in the facility. Both measures are from the OSCAR dataset. CNAs' are

mostly involved with the direct care of patients.

Percent Licensed: Percent nursing staff that are licensed is a continuous staffing mix

variables that measures the proportion of nursing staff that have a license. Higher percent

licensed will give the nursing home a higher capacity to have knowledge and intangible

resources.

Percent of nursing staff that are RNs: Percent of nursing staff that are RNs is a continuous

staffing mix variable that measures the proportion of the nursing staff that are RNs. Higher RN

staffing mixes are advantageous since RNs can use their professional skills and tacit knowledge

to improve the quality of patient care through their involvement in establishing processes and

supervising other nursing staff (Weech-Maldonado et al. 2003).

Resource commitment rehab: Resource commitment rehab includes all therapists and

therapy assistants. Only licensed rehab staff is included because the use of rehab aides for

therapy treatments is minimal due to state and federal practice guidelines.

PTFTE/ Total Residents: Physical Therapist (PT) per resident is a continuous variable

represents the ratio total number of PT FTEs divided by the total number of residents in the

facility. PTs evaluate the patient and oversee the plan of care. They supervise the physical

therapy assistants

PTA FTE/ Total Residents: Physical Therapy Assistants (PTA) per resident is a continuous

variable represents the ratio total number of PTA FTEs divided by the total number of residents

in the facility. PTAs provide the treatment to the patients under the supervision of the physical

therapist.

Percent PT: Percent of the physical therapy staff that are PTs is a continuous staffing mix

variable. Higher PT staffing mixes are advantageous since PTs can use their professional skills









Table A-22. OLS regressions of log transformed restraint with rehab focus as ref. group
Coef. Std. Err. t P>t
Cost leader 0.011 0.004 2.48 0.01
Differentiator 0.004 0.006 0.74 0.46
Specialty care focus 0.006 0.005 1.32 0.19
Lack of strategy 0.006 0.005 1.24 0.22
Ownership 0.020 0.004 5.50 0.00
Percent Medicaid 0.000 0.000 4.06 0.00
Percent Medicare 0.000 0.000 1.16 0.25
Market competition 0.053 0.070 0.76 0.45
Occupancy -0.010 0.012 -0.84 0.40
Total beds 0.000 0.000 7.90 0.00
Chain -0.007 0.003 -2.24 0.03
Acuity index 0.011 0.002 6.07 0.00
ADL index 0.016 0.004 3.80 0.00
Tube feed care -0.049 0.041 -1.20 0.23
Respiratory care -0.012 0.022 -0.54 0.59
Suctioning care -0.138 0.101 -1.37 0.17
IV therapy care 0.028 0.054 0.51 0.61
Tracheostomy care 0.060 0.110 0.54 0.59
Need injection -0.055 0.018 -3.12 0.00
cons -0.004 0.023 -0.19 0.85


Table A-23. OLS regressions of root
Coef.
Cost leader -0.013
Differentiator -0.012
Specialty care focus -0.008
Lack of strategy -0.016
Ownership 0.006
Percent Medicaid 0.000
Percent Medicare 0.001
Market competition -0.001
Occupancy -0.035
Total beds 0.000
Chain -0.002
Acuity index 0.008
ADL index 0.007
Tube feed care 0.109
Respiratory care 0.031
Suctioning care 0.045
IV therapy care 0.150
Tracheostomy care -0.031
Need injection 0.046
cons 0.163


transformed skin ulcer with
Std. Err.
0.002
0.003
0.002
0.002
0.002
0.000
0.000
0.029
0.006
0.000
0.002
0.001
0.002
0.024
0.011
0.060
0.032
0.065
0.010
0.012


rehab focus
t
-5.69
-4.19
-3.34
-6.89
3.32
-0.71
6.94
-0.04
-5.55
3.31
-1.57
8.42
3.11
4.60
2.74
0.74
4.70
-0.47
4.74
13.90


as ref group
P>t
0.00
0.00
0.00
0.00
0.00
0.48
0.00
0.97
0.00
0.00
0.12
0.00
0.00
0.00
0.01
0.46
0.00
0.64
0.00
0.00









A-32 GLM with gamma distribution PA walking improve with specialty care as ref group..121

A-33 OLS regressions of root transformed PA skin ulcers with specialty care as ref. group ..121

A-34 GLM regressions of nursing home cost with gamma distribution............................122

A-35 Odds of nursing home of being in the highest operating margin tier ............................122

A-36 Odds of nursing home of being in the highest total margin tier ..............................123

A-37 Correlations of independent variables ........ .......... ...... .... ............... 124













Scree plot


8.00
7.00
6.00
5.00
S-*- Series1
4.00 3
3.00
2.00
1.00 -
0.00
0 10 20 30 40 50 60


Figure A-1. Scree plot of scope (technology) factor analysis









Nursing homes also face increased competition from growing industries like assisted living

facilities and home health agencies that tend to draw the less dependent (and more profitable)

private pay residents. This ultimately leaves nursing homes with the residents who have higher

medical acuity, greater functional dependence and have higher care costs (Weech-Maldonado et

al. 2003).

With the limited ability to control their revenue and with the increased competition created

by substitutes, nursing homes find themselves in a precarious position. For nursing homes to

thrive (or even just survive), it is necessary for them to identify strategies that allow them to

achieve competitive advantage.

The Importance of Focusing on Strategy

Government payers and interest groups are interested in persuading nursing homes to

provide high quality of care. One way to encourage facilities to provide better quality of care is

to demonstrate that better quality results in better financial performance. Some recent studies

have determined a positive relationship between financial performance and quality of care

(Weech-Maldonado et al. 2003, 2008).

Even though nursing homes may have the incentive to provide better quality, it is also

necessary for the nursing home to determine the best strategies to be able to provide high quality

care. Without incentives and a strategy, a nursing home will have some difficulty achieving these

desired outcomes.

Managers are interested in achieving above normal profits. Strategies are plans related to

the mission and the vision of an organization that provide leadership with the means to achieve

high performance. Like any industry, nursing homes attempt to gain competitive advantage by

selecting and implementing an effective strategy that the competitors are not able to reproduce

(Barney 1991). Because the number of truly distinct strategies available in any industry is likely







Some of the case mix variables may have greater influence on the quality outcomes than others.

The rehab focus groups may be 'penalized' by taking on residents that are more prone to pressure

ulcers and whose risk factors are not captured by the risk adjustments of the quality measures or

control variables used in the analysis.

In the case of the walking improvement, the rehab focus group provided nearly 25% more

minutes then any other group and yet did not attain better outcomes than the specialty care focus,

cost leaders or lack of strategy groups. There are a few reasons that may explain why the rehab

focus group does not have better walking improvement outcomes. First, rehab focus nursing

homes may be downgrading the progress of their residents. Rehab focus strategic groups are

accustomed to providing care to rehab patients and may be more rigorous in the application of

walking improvement scales. A second reason is that there may be an upper therapy minute

threshold where adding more minutes of therapy does not translate into better outcomes. The

rehab group provided significantly more minutes of therapy when compared to the specialty care

focus, the cost leader groups and the lack of strategy group yet did not have better therapy

outcomes. The fact that differentiators provide considerably less therapy and have worst outcomes

when compared to all other groups supports the notion that there is at least a minimal threshold

level of minutes of therapy that needs to be achieved in order for basic outcomes to occur.

Another possibility is that since revenue streams are enhanced by providing higher intensity of

therapy for a longer period of time, these facilities may be more conservative with their 14 day

assessments in an effort to increase the length of stay. By having a lower 14 day assessment value,

the facility could be able to justify keeping the resident at the facility for a longer period time.

Over the past 7 years, MedPAC have issued statements of concern that the current post acute

reimbursement schematic promotes therapy overutilization. More research is needed to determine

if the lack of post acute quality outcomes difference between the rehab group and the cost leader,







population means are equal across groups; however, this method uses a multivariate approach. It

was used in Zinn et al. 1994 study that examined strategic groups.

Hypothesis #2: Nursing Homes in the Differentiator Strategic Group will have the Highest
Quality.

Dependent Variables:

There were a total of 8 dependent variables that were quality measures; five of which

relate to long term care quality, two of which measure post acute quality and one measure of the

overall quality of the facility. The long term care (LTC) and post acute (PA) measures were

quality indicator (QI) variables; they included LTC bladder decline in continence, LTC bowel

decline in continence, LTC 4 point decline in ADL function, the LTC prevalence of the pressure

sores, LTC restraint prevalence, PA walking improvement, and PA pressure sore incidence. The

bladder and bowel decline variables are measures of the proportion of residents who had become

incontinent. The ADL decline in function is a risk adjusted ratio of the proportion of the

residents who experienced a 4 point ADL decline (out of 16 points) relative to the total number

of residents in the nursing home. The restraint prevalence is the proportion of residents that are

restrained at the facility. The pressure sores prevalence is a risk adjusted ratio of the proportion

of residents who have a pressure sores relative to the total number of residents in the nursing

home. The walking improvement variable measures the change in ambulation status between the

5 day assessment and the 14 day assessment. Pressure sores incidence variable is a measure of

the amount of post acute residents who developed a pressure sore while remaining in the nursing

home relative to the total number of post acute residents in the nursing home. The quality

indicators were continuous variables calculated using the MDS based on the methodology as

described by Abt Associates 2004. These quality measures are included on the Nursing Home

Compare website and are currently used by social workers and consumers to differentiate

between good and bad facilities. The eighth quality variable was the total number of quality of

care deficiencies. This is an overall measure of the facility's quality. Quality of care deficiencies
66









(2003) study found that there was a positive relationship between quality of care and financial

performance in nursing homes.

A nursing home can also use cost leadership in which the nursing home attempts to have

relative lower costs. One would expect that nursing homes that follow the cost leadership

paradigm would be more efficient at using their staff

The third strategy is a focus strategy, in which the firm concentrates on a particular group

of customers, geographic markets, or product line segments. In the case of the nursing home

industry, some facilities may choose to focus on rehabilitation because of the higher revenues per

bed that Medicare pays.

Porter (1980) considered firms that attempted to apply both a differentiating strategy and a

cost leader strategy at the same time as 'stuck in the middle' or 'muddlers'. Stuck in the middle

or muddlers are considered those organization who have no coherent strategy. These

organizations did not satisfy the demands of either cost leaders or differentiators (Marlin et al.

2002). In contrast, Hill et al. 1988 and Wright et al. 1987 have posited that some nursing homes

can be successful at differentiating themselves from their competitors while controlling their

costs at the same time. Marlin et al. 2002 coined this type of nursing home as 'best costs' and

found that the best cost strategy existed in the hospital industry. Porter has acknowledged that on

rare occasions, firms can be successful with both differentiating strategy and a cost leader

strategy (Murray et al.1988).

Marlin et al. (2002) used strategic groups as defined by Porter's theory of competitive

strategy in the hospital industry. Marlin et al. 2002 used an objective classification procedure to

classify four strategic groups: differentiation, low cost, best costs and the muddlers. The

variables for the cluster analysis were selected by a panel of experts. Those variables selected to









Results: The study identified four strategic groups (differentiator, cost leader, rehab focus,

and specialty care focus) and a lack of strategy group. Differentiators, rehab focus and specialty

care focus groups had less quality of care deficiencies than the lack of strategy group and cost

leaders. Differentiators generally had better long term care quality outcomes compared to other

strategic groups and the lack of strategy group. Cost leaders, differentiators and lack of strategy

groups had lower cost than the rehab focus and specialty care focus groups. Cost leaders,

differentiators and rehab focus group had better financial performance than the specialty care

focus group and lack of strategy group.

Conclusions: The study indicates that using technology and resources commitments is an

effective way of identifying strategies groups. The study also found that some strategic groups

have better financial performance than other strategic groups or than nursing homes that have no

strategy at all. Nursing homes that provide a high level of technology had better quality, without

increased cost or sacrificing financial performance. The study also suggests that nursing homes

that service residents with high cost conditions may not be adequately reimbursed.







The study identified 5 of the strategic groups listed in the hypothesis (nursing care focus

(differentiator), cost leader, rehab focus, specialty care focus and lack of strategy). The study did

not identify the best cost group and did not identify a well defined differentiator group. Instead,

the study identified three focus groups; rehab focus, specialty care focus and a nursing care group.

The rehab group from this study appeared similar to the high Medicare facilities in the prior

nursing home strategic group studies. This group not only had a high proportion of Medicare beds,

but also provided a high level of therapy minutes and skin care. The specialty care focus group

had a high RN intensity and had a high concentration of residents with specific care needs.

Residents with specialty care needs like tracheotomy care or suctioning require the nursing home

that has a capacity to deliver a high level of skilled services. These conditions are also only

minimally covered by Medicare in the post acute setting and have long been considered

underserved by MedPAC (MedPAC 2003, 2007). Beginning in 2004, nursing homes could get

more reimbursement for residents with these extended needs. This change in reimbursement only

started in 2004 and it may be possible that the facilities had not been able to fully take advantage

of it.

One unexpected finding was that the defined differentiator strategic group provided the least

therapy services. The fact that this group is defined as a differentiator when it provides the lowest

amount of therapy brings up the question of the appropriateness of using differentiator as this

group's definition. The group clearly provided the most nursing care technology and therefore was

reclassified as the nursing care focus strategic group. Whether this group is considered a

differentiator or a nursing care focus strategic group, this group still represents nursing homes that

are trying to separate themselves from their competition by providing a high level of technology

and therefore they can still be defined by using Porter's generic strategies.

Out of all the clusters identified, the cost leader proved to be the best defined. The cost

leader group had the lowest staffing levels and lower staffing mixes. Lower staffing mixes would

89









Table 4-3. Rank scores scope and resource variables
Variables Cluster 1 Cluster 2 Cluster 3 Cluster 4 Cluster 5
Differentiator Cost Rehab Specialty Lack of
TLqA d F FQSt-t


ea er ocus care ocu y


Average Wages 1.985
RN per Resident 1.70*
LPN per Resident 0.993,4
CNA per Resident 4.362,3,4
Therapist per Resident 1.00*
Therapy assistant per resident 1.19*
Percent Licensed PT 2.542,3,5
Percent Licensed OT 2.612'3'5
Percent RN 1.15*
Percent Licensed Nurse 0.41*

Total Resource 17.91*

LTC Skin Process 3.23*
Acute Skin Process 4.682,4,5
LTC Restorative Process 5.00*
Acute Restorative Process 5.00*
ReliefDevices 4.112A45
Toilet Program 4.96*
Weight Program 5.003,5
Therapy Minutes 0.00*
Percent Medicare 0.103,4

Total Scope 32.1*
*Significantly different from all clusters,
Superscript is significantly different than identified groups


0.813,4
0.00*
1.073,4
3.59*
0.00*
2.48*
0.00*
0.00*
0.00*
0.071,3,4

8.06*

0.60*
1.56*
0.101,3,4
0.141,3,4
1.672 4A5
1.49*
3.733,5
1.601 3,5
0.153,4

11.03*


4.042,5
2.72*
5.00*
5.00*
5.00*
5.00*
1.58*
1.94*
1.42*
1.76*

33.49*

5.00*
5.002,4,5
0.89*
0.00*
5.001,4,5
0.111,2
0.00,2,4
5.00*
5.00*

26.10*


5.001,2,5
5.00*
0.00*
0.00*
1.75*
1.62*
2.442,3,5
2.602,3,5
5.00
5.00*

28.32*

2.07*
3.03*
0.34*
0.33*
1.5124,5
0.091,2
2.533
1.681,3,5
0.74*

12.32


0.001,2
0.40*
0.863,4
4.052,3,4
2.23*
0.00*
5.00*
5.00*
0.25*
0.001,3,4

17.78*

0.00*
0.00*
0.00,3,4
0.141,3,4
0.00*
0.001 2
1.031,2
1.13*
0.003,4

2.29


The descriptives values were not adjusted for the county level fixed effects and therefore were not

used in the determination of the clusters.

The plotted clusters in Figure 4-2 were then categorized using Porter's generic strategies.

Based on the definition described in the methodology, Cluster #1 was initially classified as

Differentiator because it's scope score was the highest (technology cumulative rank score).

However this group was ranked the lowest in all the categories that was related to rehab (i.e.

physical therapy minutes and percent Medicare beds). Therefore this cluster may be more

appropriately classified as a nursing care focus group rather than a differentiator. Although these

nursing homes provided a high degree of technology, they did not use the highest amount of

resources. They tended to have median level wages, median level total RN staffing, and below

median level of percent licensed therapist. They had low staffing mix as a result of their high


78


=13.8"**
:744***
:121***
=367***
:1110***
:1492***
=1180***
:906***
=1538***
=833***

=186***

:186***
:58.35***
=688***
=1987***
=39.8***
:19.7***
=8.62***
=386***
=918***

=59***









CHAPTER 2
CONCEPTUAL FRAMEWORK

This study will begin by defining strategic group theory, Porters' generic strategy theory,

technology and Cool and Schendel's scope and resource commitment theory. The study will then

explain how these theories are integrated to generate a model that uses Porter's generic strategies

to define strategic groups in the nursing home. Finally, the study will define how the different

strategic groups can have different strategic-performance relationships as it pertains to quality,

cost and financial performance.

Strategic Groups

The concept of strategic groups was introduced by Hunt (1972) in his thesis that

examined the appliance industry. He discovered that there was less competitive rivalry than what

industry concentration ratios suggested. He attributed this to the existence of subgroups within

the industry that effectively reduce the number of competitors in each market. Caves and Porter

(1977) expanded the theory by defining strategic groups as a set of firms that face similar threats

and opportunities that are different from the threats and opportunities faced by other firms in the

industry. Strategic group theory stipulates that within an industry, there could be groupings of

organizations that have very different strategies and yet may still have good financial

performance. This theory was a shift in thinking from the industrial organizational theory where

it was generally believed that there was only one right way of doing things (Leask 2007).

Analyzing strategic groups gives insight to different competitors' approaches to the marketplace

(Harrigan 1985). The number of strategic groups within an industry corresponds to the number

of unique strategies within that industry (Marlin et al. 1999). Strategic groups are persistent

strategic characteristics of an industry, which are protected by mobility barriers (Porter 1980).









there is an MD order, that the goals are generated by the physical therapist and the program

requires some oversight to ensure that the service is provided. The responsibility of the

organizational features of a restorative program is placed on the licensed nurse.

Failure to provide these processes of care can result in a decline in the resident's health

status. One example of nursing technology is pressure sore prevention. If pressure sore

preventative processes are not provided, a resident will be more likely experience a decline in

their health status because they can develop a pressure sore. Having a patient on a turning

schedule every 2 hours or providing them with a pressure relieving device reduces the likelihood

of a pressure sore will occur. Turning schedules relieve pressure and allow blood to flow to the

skin. Pressure relieving devices distribute the pressure over a greater surface area and lower the

pressure on the high-risk breakdown areas like the coccyx (CMS 2008). A nursing home that

puts emphasis on nursing technology may be trying to get better nursing outcomes. More

processes of care and better outcomes are indicators of better quality (Schnell et al. 2004). Most

of nursing home technology is provided by the nurses. A facility focused on better quality may

be trying to differentiate themselves from their competitors. By differentiating themselves from

their competitors, a facility may get a greater proportion of private pay residents and/or higher

profit margin Medicare residents. In addition, better outcomes can also lead to lower costs

(Weech-Maldonado et al. 2003).

Rehabilitation technology: Rehabilitation technology can be provided to either post acute

or long term care residents. These processes of care consist primarily of therapy services and are

geared to improve the resident's health status. Rehabilitation technology is not considered

routine because of its dynamic nature. Therapy services vary based on the needs of the residents.

Some residents may have difficulty with transferring from the chair to the toilet while others







are better than others with respect to ensuring quality of care. This study has found some evidence

that nursing homes with strategies directed toward providing care have better quality of care than

those nursing homes focused on cost savings. The study also found some evidence that nursing

homes that provide a high level of nursing care technology to their long term care residents have

better quality of care outcomes for that population. The study also identified a group of nursing

homes that provide care to patient with needs other than rehabilitation that may not be receiving

adequate funding. The specialty care group provides service to a population that is very

vulnerable. Future policies should be directed to provide incentives that increase the amount of the

important processes of care provided and not just address structural factors like staffing.

Another factor is that is of interest to government agencies like CMS is that facilities that

provided a significantly higher amount of therapy did not have better post acute outcomes with

respect to walking improvement. At the same time, the strategic group that provided significantly

less therapy than the other groups had worse walking improvement. From these results, it appears

that there is a minimum amount of therapy required for adequate improvement, but there is limit

beyond which more therapy does not lead to any more significant benefit. With the high costs of

post acute care which is almost solely paid for by the government, understanding how much

therapy is sufficient is be very valuable. Further research is necessary to determine the minimum

amount of therapy that is necessary.

Similar to managers, policymakers are interested to know that a strategy of providing a

high degree of service does not necessarily costs more. The fact that some facilities are able to

provide a high level of service and yet keep their cost down can guide policymakers to provide

incentives that focus on care provided as well as staffing. Further research is needed examining

the relationship between processes of care and the outcomes that they would be expected to

influence.







not have any Medicare residents or even Medicare beds available and therefore they would not

have to file a Medicare cost report. Therefore, this study is limited to nursing homes that are non

government, free standing and have Medicare certified beds. The total number of facilities in the

study is approximately 11601 facilities in the year 2004.

Overview of Methodology

To identify the strategic groups and the strategy-performance relationships of the

strategic groups, the study performed a factor analysis on the scope and resource commitment

variables. The resulting variables were used in a cluster analysis to identify the different clusters.

Once clusters were identified, ordinary least squares regression (OLS), negative binomial and

ordered logit was used to determine if the different groups had different financial and quality

strategic-performance relationships.

Hypothesis #1: Nursing homes that have a strategy can be categorized into one of four
strategic groups: Differentiator, cost leadership, focus or best cost.

Scope

Similar to Zinn et al. (1994) and Marlin et al. (1999), this study used the scope

commitment and resource commitment (also known as resource deployment) dimensions to

identify the strategic groups. The scope of this study included technologies used by nursing

homes to take care of the residents. The technology of a nursing home can be classified into two

main categories: nursing technology and rehab technology.

Nursing Technology

Nursing technology includes many different types of processes. The overall care of the residents

can be divided into several different component technologies that are specifically designed to

address the varied needs of the nursing home resident. Each of these technologies has a specific

purpose and if not applied properly will have an undesired outcome. Some of the essential

technologies include skin care technology, restorative technology, incontinence reducing

technologies and nutritional technologies.








choose to substitute a licensed staff member with a non licensed staff in an attempt to keep costs

low. By generating an index that included all three nursing variables, the ability of the cluster

analysis to identify nursing homes with different types of staffing intensity strategies would be

compromised. It was therefore deemed necessary to include the RN, LPN and CNA variable in the

cluster analysis as individual variables.

Table 4-1. Internal consistency of the composite variables
Cluster variable Alpha Variables used
Long term care skin Long term care ointment, dressing application process, surgical wound care, ulcer
processes of care 0.82 care, skin nutrition and turning repositioning program
Post acute skin processes
of care 0.72 Post acute ulcer care, dressing application process,
Long term care restorative Long term care restorative dressing, restorative transfers, restorative bed mobility,
processes 0.86 restorative PROM, Restorative AROM, restorative ambulation and restorative eating

Post acute restorative Post acute restorative dressing, restorative transfers, restorative bed mobility,
processes 0.91 restorative PROM, Restorative AROM, restorative ambulation and restorative eating
Long term care pressure relieving bed and chair, Post acute pressure relieving bed
Relief Devices 0.81 and chair
Therapy minutes 0.93 Long term care and Post acute physical and occupational therapy
Toileting program 0.88 Long term care and Post acute toileting program
Weight management prg 0.88 Long term care and post acute weight management program
Therapist 0.91 Total physical and occupational therapist per resident
Therapy assistants 0.77 Total physical and occupational therapy assistants per resident

Table 4-2 lists the nursing variables and all the individual variables utilized in the cluster analysis.

Other variables included the percent of nurses that were RNs, percent of nurses that were

Table 4-2. Individual variables used in cluster analysis that were not included in a factor variable.
Individual Variables Definition

RN Registered Nurse Per Resident
LPN Licensed Practical Nurse Per Resident
CNA Certified Nurse Assistant Per Resident
RN Percent Percent of Nurses are Registered nurse
License Nurse Percent Percent of Nurses are Licensed
PT Percent Percent of Physical Therapist are licensed
OT Percent Percent of Occupational Therapist are licensed
Wages Facility direct care average wage
Percent Medicare Percent residents who are post acute
licensed, percent of physical therapist, and the percent occupational therapist. As described in the

methodology, wages and percent Medicare beds were also used as individual variables.









AN EXAMINATION OF STRATEGIC GROUP MEMBERSHIP AND TECHNOLOGY IN
THE NURSING HOME INDUSTRY























By

ALEXANDRE LABERGE


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2009







building operated by the IT department at the University of Florida. It was originally acquired by

University of South Florida.

The Area Resource File (ARF) contains county level data on supply and demand factors,

such as composition of health care workforce, and socioeconomic and demographic

characteristics of the population.

The Medicare Cost Report data is a public access relational dataset that can be acquired

from the CMS via the CMS website. All CMS certified nursing homes that have Medicare beds

are required to submit the Medicare cost report on an annual basis. MedPAC has repeatedly been

using this dataset to report the financial information of nursing homes to congress over the past 7

years (MedPAC 2007). This dataset provides financial data like total patient revenue, total

patient costs, and total revenue.

Population

This study included all nursing homes in the United States that had Center for Medicaid

and Medicare certification and that were included in the OSCAR, MDS and Medicare cost report

datasets in 2004. There were 16000+ nursing homes in the United States in 2004. However, a

total of 4500 facilities were omitted because they were either hospital based (2000 facilities) or

government based (700 facilities), did not have Medicare beds (1200 facilities) or were the only

nursing home in a county and would not have any market competition (500 facilities). Hospital

based facilities were omitted because they may behave differently from free standing facilities

(Weech-Maldonado et al. 2003, Weech-Maldonado et al. 2004). For example, hospitals may

place patients in their own subacute beds that they cannot place in other nursing homes due to

the level of acuity of the patient or because the patient lacks insurance. Government facilities

may behave differently than non-government facilities because of their different financing

structure. They are less influenced by market forces because of funding that they may receive

directly from public sources. Finally, some facilities that may be included in the OSCAR may

44









Table A-26. Negative binomial of quality of care deficiencies with specialty care as reference group
IRR Std. Err. z P>z
Cost leader 1.058 0.030 2.00 0.05
Rehab focus 0.992 0.033 -0.25 0.80
Differentiator 0.999 0.037 -0.04 0.97
Lack of strategy 1.123 0.032 4.02 0.00
Ownership 1.099 0.029 3.59 0.00
Percent Medicaid 1.005 0.001 7.69 0.00
Percent Medicare 1.005 0.001 3.75 0.00
Market competition 0.773 0.367 -0.54 0.59
Occupancy 0.830 0.065 -2.38 0.02
Total beds 1.003 0.000 14.74 0.00
Chain 1.024 0.021 1.18 0.24
Acuity index 1.040 0.013 3.20 0.00
ADL index 0.952 0.029 -1.63 0.10
Tube feed care 1.009 0.237 0.04 0.97
Respiratory care 0.972 0.157 -0.18 0.86
Suctioning care 0.811 0.454 -0.37 0.71
IV therapy care 3.826 1.304 3.94 0.00
Tracheostomy care 1.142 0.745 0.20 0.84
Need injection 1.189 0.141 1.46 0.14
Ifips_2 1.206 0.470 0.48 0.63

Table A-27. OLS regressions of root transformed bladder decline with specialty care focus as ref
group
Coef. Std. Err. t P>t
Cost leader -0.003 0.003 -1.13 0.26
Rehab focus 0.003 0.003 0.86 0.39
Differentiator -0.009 0.004 -2.68 0.01
Lack of strategy -0.004 0.003 -1.55 0.12
Ownership 0.006 0.002 2.46 0.01
Percent Medicaid 0.000 0.000 -5.35 0.00
Percent Medicare 0.001 0.000 3.84 0.00
Market competition 0.077 0.064 1.21 0.23
Occupancy 0.014 0.008 1.70 0.09
Total beds 0.000 0.000 3.66 0.00
Chain 0.004 0.002 1.97 0.05
Acuity index 0.009 0.001 6.66 0.00
ADL index 0.021 0.003 6.68 0.00
Tube feed care -0.061 0.027 -2.24 0.03
Respiratory care -0.040 0.014 -2.99 0.00
Suctioning care 0.097 0.067 1.44 0.15
IV therapy care 0.004 0.040 0.09 0.93
Tracheostomy care -0.245 0.074 -3.32 0.00
Need injection 0.004 0.012 0.33 0.74
cons 0.251 0.017 14.71 0.00



































This document is dedicated to the four women in my life









Table A-32. GLM with gamma distribution PA walking improve with specialty care as ref. group
Coef. Std. Err. Z P>z
Cost leader -0.005 0.016 -0.33 0.74
Rehab focus 0.014 0.019 0.73 0.47
Differentiator -0.037 0.021 -1.76 0.08
Lack of strategy 0.024 0.016 1.45 0.15
Ownership -0.087 0.014 -6.04 0.00
Percent Medicaid -0.004 0.000 -10.18 0.00
Percent Medicare -0.002 0.001 -2.81 0.01
Market competition -0.425 0.248 -1.71 0.09
Occupancy -0.160 0.043 -3.73 0.00
Total beds 0.000 0.000 -0.53 0.60
Chain -0.032 0.012 -2.74 0.01
Acuity index -0.028 0.007 -4.07 0.00
ADL index -0.110 0.017 -6.49 0.00
Tube feed care -0.666 0.143 -4.66 0.00
Respiratory care 0.139 0.081 1.71 0.09
Suctioning care 0.444 0.419 1.06 0.29
IV therapy care 0.240 0.213 1.13 0.26
Tracheostomy care -0.238 0.449 -0.53 0.60
Need injection 0.038 0.068 0.56 0.58
Ifips_2 0.218 0.283 0.77 0.44
cons -0.35 0.188 1.11 0.27


Table A-33. OLS regressions of root
Coef.
Cost leader -0.018
Rehab focus 0.012
Differentiator 0.027
Lack of strategy -0.049
Ownership 0.034
Percent Medicaid 0.001
Percent Medicare 0.001
Market competition -0.031
Occupancy 0.014
Total beds 0.000
Chain -0.021
Acuity index 0.034
ADL index 0.027
Tube feed care 0.568
Respiratory care 0.088
Suctioning care 0.535
IV therapy care 0.270
Tracheostomy care -0.160
Need injection 0.183
cons -2.540


transformed PA skin
Std. Err.
0.016
0.017
0.021
0.016
0.014
0.000
0.001
0.308
0.046
0.000
0.012
0.007
0.019
0.156
0.082
0.413
0.194
0.431
0.069
0.096


ulcers with specialty
T
-1.15
0.69
1.28
-3.00
2.38
3.39
0.70
-0.10
0.31
5.08
-1.76
4.88
1.42
3.64
1.07
1.30
1.40
-0.37
2.63
-26.36


care as ref. group
P>t
0.25
0.49
0.20
0.00
0.02
0.00
0.48
0.92
0.75
0.00
0.08
0.00
0.16
0.00
0.28
0.20
0.16
0.71
0.01
0.00









TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ............................................... ..................................... .....................4

LIST O F TA BLE S ........................................................................................................... ..... 8

LIST OF FIGURES ............ ...................................... .....................11

A B S T R A C T ......... ....................... ............................................................ 12

CHAPTER

1 INTRODUCTION ............... .......................................................... 14

What Are Nursing Homes and What Are Their Roles in the Health Care System? .............15
Long Term Care Resident Reimbursement ............................................................16
M e d icaid ............................................................................... 16
P riv a te P a y ................................................................................................................. 1 8
M medicare ................................................................................................... ............. 18
Post Acute Care Reim bursem ent.......................................................................... 19
N using H om e E nvironm ent............................................................................ .............. 20
The Importance of Focusing on Strategy...................................................... ..................21
P u rp o se ..............................................................................................2 2
Research Questions ........................................................ .. ......... ............... 22

2 CONCEPTUAL FRAMEWORK............................................................ ...............23

S trateg ic G rou p s ........................................................................... 2 3
N ursing H om e T technology ........................................................................... ....................26
N using H om e R sources ............................................................................. .................... 28
Scope and R source C om m itm ent............................................................................ .....31
Scope U sing T technology .......................................................................... ..............32
R source C om m itm ent ....................................................................... .... ...................33
P orter's G generic Strategies ....................................................................................... .......... 34
Strategic Group Model Using Scope (Technology) Commitment, Resources
Commitment and Porter's Generic Strategy Theory.........................................................36
Differentiator ...................................................... ................... ............... 37
C ost L leadership ............. ................................................................. ..............37
B e st C o st .................................................................................................................... 3 8
F o c u s ....................................................................... 3 8
L ack of Strategy ................................................................... 39
M easu rin g P perform an ce .................................................................................................... 3 9
N ew C o n trib u tio n s ............................................................................................................ 4 1
H y p o th e sis ..........................................................................4 1







the dependent variables were transformed by using the log function. If the initial log

transformation did not result in proper skewness and kurtosis, other transformations were

attempted. Other possible transformations included the cubed root, the root, the square and the

cube. All of them would be attempted until a transformed variable was identified that had a

skewness and kurtosis closest to 0 and 3 respectively. If the skewness was not able to be resolved

and the variable seemed to have a gamma distribution, the study used the generalized linear

model with gamma distribution and a log link. After the OLS analysis, it was necessary to

determine if the regressions pass the OLS linearity assumptions. The linearity assumption was

tested through observation of the residual-predicted values plots and the Hosmer-Lemeshow test.

The quality of care deficiencies is a count variable and was analyzed using the negative

binomial equation. The study used White's heteroskedacity-consistent covariance matrix (1980) to

correct for a heteroskedascity. It used county level fixed effects. Dependent variable outliers of 5

standard deviations and above were dropped from the dataset.

Qual = V + &giGroupl + &g2Groupl + .....+&gnGroupN + &cControl + &feFE + E

Qual represents one of the seven quality dependent variables.V represents the constant

value, while &gn represents the effect of each strategic group on the quality variable. &gc

represents the effect of the control variables on the quality variable. &fe represents the effect of

the county level fixed effects. E represents the random error.

Hypothesis #3: Nursing Homes in the Focus Strategic Group will provide Higher Quality
in the area that they are focused on compared to other Strategic Groups.

Dependent Variables

Hypothesis #3 used the same quality of care dependent variables as Hypothesis #2.

Independent Variables

As in hypothesis #2, the independent variables of interest included a dichotomous

variable (0 or 1) for each strategic group generated by the cluster analysis. However, unlike







Skin care technology is designed to prevent pressure ulcers from occurring and to treat

pressure ulcers that already exist. Failure to apply preventative skin care technology properly can

result in pressure ulcers. Failure to treat already existing pressure ulcers can result in pressure

ulcer worsening. Restorative nursing technologies engage the resident in functional activities.

Residents functional abilities can decline in the nursing home setting and restorative technology

can be used to help maintain their functional levels. Incontinence reducing technologies assist

residents to maintain or regain their continence through toileting programs or bladder retraining

programs. Nutritional technologies help residence maintain a healthy weight. In the nursing

home setting, residents are at risk of either gaining weight or losing weight depending on their

medical condition. Weight management programs and therapeutic diets assist with proper

nutrition while the use of plate guards and other devices ensures that residents adequately feed

themselves.

Skin care technology: All processes of care that involve either preventative skin care or

treatment skin care are considered skin care technology.

Pressure relieving seat: This is a continuous facility level variable that represents the

proportion of residents in the nursing home that are given a pressure relieving seat. Pressure

relieving seats are an important intervention when trying to prevent a new pressure sore from

occurring or to prevent an already acquired pressure sore from worsening. Includes gel, air (e.g.,

Roho), or other cushioning placed on a chair or wheelchair. Include pressure relieving, pressure

reducing, and pressure redistributing devices.

Pressure relieving bed: This is a continuous facility level variable that represents the

proportion of residents in the nursing home that are given a pressure relieving bed. Pressure

relieving beds are an important intervention to prevent new pressure sores from occurring or to

prevent an already acquired pressure sore from worsening. Includes air fluidized, low air loss











Table A-2. Factor analysis with 10 factors of scope (technology) variables
Rotated Factor Pattern
Factor Factor2 Factor3 Factor4 Factor5 Factor6 Factor? Factor 8 Factor9 FactorlO
Long term care plateguard -0.01 0.13 0.07 0.02 -0.04 0.01 0.05 0.02 0.00 0.76
Long term care Weight program -0.01 -0.01 0.04 0.03 0.02 -0.03 0.03 0.02 0.92 0.03
Long term care Therapy Diet 0.01 0.17 0.07 -0.03 0.03 0.05 0.13 -0.05 -0.02 0.70
Long term care Bladder training -0.03 0.07 0.08 -0.11 0.02 0.05 0.12 0.30 -0.05 -0.23
Long term care toilet program 0.07 -0.05 0.02 0.20 -0.06 -0.02 -0.06 0.79 0.05 0.22
Long term care no skin treatment 0.00 0.09 0.00 -0.71 -0.24 -0.05 0.00 -0.11 -0.01 -0.06
Long term care other skin tx -0.03 0.63 0.14 0.24 0.05 -0.05 0.01 0.08 0.00 0.22
Long term care ointment -0.06 0.70 0.18 0.02 0.13 0.02 0.06 -0.02 -0.02 0.16
Long term care pressure relief bed -0.05 0.57 0.18 0.59 -0.19 0.04 0.02 -0.11 -0.05 0.04
Long term care pressure relief chair 0.00 0.61 0.12 0.51 -0.20 -0.05 0.06 0.00 -0.02 0.08
Long term care dressing application -0.06 0.77 0.11 -0.07 0.16 0.13 0.05 -0.18 -0.01 0.06
Long term care surgical wound care 0.06 0.40 -0.09 -0.07 0.08 0.35 0.06 -0.07 0.01 0.01
Long term care ulcer care -0.03 0.72 0.07 -0.06 0.15 0.11 0.03 -0.20 0.00 0.01
Long term care skin nutrition program 0.02 0.60 0.08 0.04 0.21 -0.09 -0.08 0.12 0.09 -0.06
Long term care turning repositioning -0.02 0.69 0.14 0.18 0.03 0.04 0.00 0.19 -0.03 -0.03
Long term care OT minutes 0.03 0.34 -0.11 -0.05 -0.11 0.75 0.05 -0.09 0.10 -0.02
S Long term care PT minutes 0.03 0.38 -0.10 -0.05 -0.10 0.74 0.04 -0.09 0.09 0.01
I Long term care other restorative care -0.07 0.15 0.47 0.01 -0.04 0.02 0.21 0.08 -0.02 -0.17
Long term care restorative communication 0.17 0.01 0.39 0.04 0.04 -0.01 -0.13 -0.07 0.01 0.02
Long term care restorative amputation -0.04 -0.03 0.38 0.05 -0.03 -0.02 0.05 0.01 0.00 -0.19
Long term care restorative eating 0.17 0.12 0.72 0.00 0.01 0.06 0.05 -0.01 0.03 0.08
Long term care restorative dressing 0.30 0.11 0.74 -0.02 0.01 0.02 0.03 0.03 0.03 0.13
Long term care restorative transfers 0.31 0.14 0.73 0.00 0.00 0.01 0.06 0.01 0.00 0.09
Long term care restorative bed mobility 0.39 0.06 0.64 0.00 0.03 0.00 -0.03 0.01 0.03 -0.01
Long term care restorative splinting 0.02 -0.03 0.18 0.05 0.23 0.14 0.61 -0.03 0.00 -0.03
Long term care restorative Passive ROM 0.12 0.31 0.42 0.03 -0.03 -0.03 0.59 0.03 0.00 0.17
Long term care restorative Active ROM 0.10 0.28 0.47 0.01 -0.09 0.01 0.51 0.05 -0.01 0.17
Long term care restorative ambulation 0.05 0.33 0.62 0.02 -0.07 0.03 0.35 0.07 -0.02 0.22
Post acute plateguard 0.03 0.02 -0.01 0.06 -0.03 -0.14 0.12 0.12 0.10 0.24
Post acute Weight program -0.02 0.01 0.03 0.03 0.05 -0.03 0.03 0.01 0.93 0.02
Post acute Therapy Diet -0.01 -0.04 -0.04 0.04 0.10 -0.15 0.29 0.02 0.05 0.10
Post acute Bladder training -0.01 -0.04 0.06 -0.05 0.14 0.05 0.15 0.39 -0.07 -0.20
Post acute toilet program 0.11 -0.06 -0.05 0.18 -0.03 -0.07 -0.03 0.80 0.09 0.12







service and does not require the skill of the therapist. However the goals of the program need to

be generated by the therapist.

Restorative Active Range ofMotion (AROM): The restorative AROM variable is facility

level and is continuous. It represents the average number of days that a resident receives

restorative AROM while remaining in the facility. It is generated by dividing the resident level

restorative variable that represents the number of days of AROM provided in the 7 days before

the assessment date by the total number of residents in the facility. Restorative AROM has the

nursing staff range the residents in order to improve the AROM of the resident. It is a routine

service and does not require the skill of the therapist. However the goals of the program need to

be generated by the therapist.

Restorative transfers: The restorative transfer variable is facility level and is continuous. It

represents the average number of days that a resident receives restorative transfers while

remaining in the facility. It is generated by dividing the resident level restorative variable that

represents the number of days of transfers provided in the 7 days before the assessment date by

the total number of residents in the facility. Restorative transfers have the nursing staff transfer

the residents in order to improve the resident's ability to transfer from one position to another. It

is a routine service and does not require the skill of the therapist. However the goals of the

program need to be generated by the therapist.

Restorative bed mobility: The restorative bed mobility variable is facility level and is

continuous. It represents the average number of days that a resident receives restorative bed

mobility while remaining in the facility. It is generated by dividing the resident level restorative

variable that represents the number of days of bed mobility provided in the 7 days before the

assessment date by the total number of residents in the facility. Restorative bed mobility has the

nursing staff transfer the residents in order to improve the resident's ability to transfer from one









Table A-28. OLS regressions of root transformed bowel
Coef. Std. Err.
Cost leader -0.003 0.003
Rehab focus -0.002 0.003
Differentiator -0.012 0.004
Lack of strategy -0.006 0.003
Ownership 0.005 0.003
Percent Medicaid 0.000 0.000
Percent Medicare 0.001 0.000
Market competition 0.040 0.050
Occupancy 0.013 0.009
Total beds 0.000 0.000
Chain 0.004 0.002
Acuity index 0.011 0.001
ADL index 0.025 0.003
Tube feed care -0.039 0.029
Respiratory care -0.040 0.015
Suctioning care 0.014 0.077
IV therapy care 0.016 0.044
Tracheostomy care -0.101 0.085
Need injection 0.020 0.013
cons 0.188 0.017


decline with specialty
t
-1.10
-0.64
-2.92
-2.18
1.83
-3.76
3.82
0.79
1.49
2.94
1.79
7.26
7.43
-1.34
-2.74
0.18
0.36
-1.19
1.55
10.99


care as ref. group
P>t
0.27
0.52
0.00
0.03
0.07
0.00
0.00
0.43
0.14
0.00
0.07
0.00
0.00
0.18
0.01
0.86
0.72
0.23
0.12
0.00


Table A-29. OLS regressions of root transformed ADL decline with
Coef. Std. Err.
Cost leader 0.003 0.002
Rehab focus -0.002 0.003
Differentiator -0.005 0.003
Lack of strategy -0.001 0.002
Ownership 0.001 0.002
Percent Medicaid 0.000 0.000
Percent Medicare 0.000 0.000
Market competition 0.083 0.053
Occupancy 0.006 0.007
Total beds 0.000 0.000
Chain -0.001 0.002
Acuity index 0.002 0.001
ADL index -0.003 0.003
Tube feed care -0.028 0.026
Respiratory care -0.005 0.013
Suctioning care -0.054 0.057
IV therapy care -0.013 0.033
Tracheostomy care -0.072 0.066
Need injection -0.008 0.011
cons 0.379 0.016


specialty
t
1.39
-0.86
-1.67
-0.28
0.50
0.10
2.97
1.57
0.84
-0.71
-0.72
1.46
-1.01
-1.10
-0.42
-0.95
-0.39
-1.10
-0.69
24.46


care as ref. group
P>t
0.16
0.39
0.10
0.78
0.62
0.92
0.00
0.12
0.40
0.48
0.47
0.14
0.31
0.27
0.67
0.34
0.70
0.27
0.49
0.00










LIST OF REFERENCES


Agresti A, Finlay B, Statistical methods for social sciences, third edition, New Jersey 1997

Allen, M.J., & Yen, W. M. (2002). Introduction to Measurement Theory. Long Grove, IL: Waveland Press.

American Physical Therapy Association History of Therapy caps Available at:
http://www.apta.org/AM/Template.cfm?Section=Media&TEMPLATE=/CM/ContentDisplay.cfm&CONTENT
ID=47095 Accessed March 30th, 2008

Amit, R., & Schoemaker, P. f. H. 1993. Strategic assets and organizational rent. Strategic Management
Journal, 14: 33-4B

Ansoff, H. Igor, Corporate Strategy, McGraw-Hill, New York, 1965

Abt Associates (2004), National Nursing Home Quality Measures User's Manual, Available at:
http://www.cms.hhs.gov/NursingHomeQualityInits/downloads/NHQIQMUsersManual.pdf Accessed March
30th, 2008

Barney J B. 1991. Firm resources and sustained competitive advantage. Journal ofManagement, 17: 99-129.

Britton, L. C.; Clark, T. A. R.; Ball, D. F..Modify or Extend? The Application of the Structure Conduct
Performance Approach to Service Industries. By: Service Industries Journal, Jan92, Vol. 12 Issue 1, p34-43,
10p;

Center for Medicare and Medicaid 2007. National Health Expenditures by type of service and source of funds,
CY 1960-2006. Available at:
http://www.cms.hhs.gov/NationalHealthExpendData/02-NationalHealthAccountsHistorical.asp#TopOfPage
Accessed March 30th, 2008

Center for Disease Control. 2007. NCHS faststats Nursing home care. Available at: http://www.cdc.gov/ Cohen
nchs/fastats/nursingh.htm Accessed March 30th, 2007

Center for Medicare and Medicaid 2007 Medicare Long Term Care Available at:
http://www.medicare.gov/longtermcare/static/home.asp Accessed March 30th, 2008

Cool K, Schendel, D STRATEGIC GROUP FORMATION AND PERFORMANCE: THE CASE OF THE
U.S. PHARMACEUTICAL INDUSTRY, 1963-1982. Management Science; Sep87, Vol. 33 Issue 9, p1102-
1124, 23p

Castle NG. Strategic groups and outcomes in nursing facilities. Health Care Manage Rev. 2003 Jul-
Sep;28(3):217-27.

Cool, K., Schendel, D., (1988) Performance differences among strategic group members. Strategic
Management Journal, 9 (3), pp 207-223

Caves, R., Porter, M., (1977), From entry barriers to mobility barriers: conjectural decisions and contrived
deterrence to new competition. Quarterly Journal of Economics p.241-261

Centers for Medicare and Medicaid: RAI MDS Users Manual Available at:







processes of care include physical therapy for mobility issues and occupational therapy for ADL

issues other than mobility.

Long term care rehabilitation: All processes of care provided by the therapy staff that

involves rehabilitation services to the long term care residents.

Mean PT minutes long term care per resident: Facility level variable that represents the

average number of minutes of physical therapy per resident provided in the 7 days prior to the

quarterly assessment date. The numerator is the number of minutes of PT and the denominator is

the number of residents in the nursing home. This is a MDS resident level variable that is raised

to the facility level.

Mean OTminutes long term care per resident: Facility level variable that represents the

average number of minutes of occupational therapy per resident provided in the 7 days prior to

the quarterly assessment date. The numerator is the number of minutes of OT and the

denominator is the number of residents in the nursing home. This is a MDS resident level

variable that is raised to the facility level.

Post acute rehabilitation: All processes of care provided by the therapy staff that involves

rehabilitation services to the long term care residents.

Mean PT minutes post acute per resident: Facility level variable that represents the

average number of minutes of Physical Therapy the 7 days prior to the 14 day assessment date.

The numerator is the number of minutes of PT and the denominator is the number of residents in

the nursing home. This is a MDS resident level variable that is raised to the facility level.

Mean OTminutes post acute per resident: Facility level variable that represents the

average number of minutes of Occupational Therapy provided in the 7 days prior to the 14 day

assessment date. The numerator is the number of minutes of PT and the denominator is the

number of residents in the nursing home. This is a MDS resident level variable that is raised to

the facility level.







pressure sores and also provided the most skin care. Another concern about the post acute pressure

ulcer incidence variable was that it did not have a normal distribution and an adequate

transformation was not achieved. When the GLM analysis was attempted, the analysis could not

converge. Therefore the results of this measure may not be reliable.

Another limitation of the study was the magnitude of the impact of some of the dependent

variables. 4 point ADL decline and post acute improvement in walking had very small regression

coefficients relative to the constant. Although these variables presented statistical significance, the

small size of the effect may limit the relevance.

The study used cross-sectional data and therefore caution should be taken when trying to

infer a causal relationship. Finally, the datasets included in the study are administrative data.

Although the nursing homes can be held accountable for the accuracy of the data, some

inconsistencies can occur depending on the individual who is entering the data. For example, the

MDS is entered by at least one nurse in each of the 16000 nursing homes in the United States.









Table A-9. Descriptives of dependent variables, independent variables and control variables
Cluster 1 Cluster 2 Cluster 3 Cluster 4 Cluster 5
Differentiator Cost Leader Rehab focus Specialty Lack of
N=990 N=2752 N=1816 care strategy
N=1634 N=2534
Avg Wages 12.61 12.52 12.99 12.76 12.45
RN per Resident 0.06 0.04 0.07 0.08 0.05
LPN per Resident 0.14 0.14 0.17 0.14 0.14
CNA per Resident 0.44 0.42 0.47 0.31 0.44
Therapist per Res. 0.02 0.01 0.04 0.02 0.02
Therapist asst per Res. 0.01 0.02 0.03 0.01 0.01
Percent Licensed PT 0.62 0.39 0.54 0.6 0.84
Percent Licensed OT 0.65 0.41 0.6 0.64 0.86
Percent RN 0.09 0.07 0.1 0.17 0.08
LTC Skin Process 0.83 0.72 0.87 0.78 0.69
Acute Skin Process 0.91 0.83 0.91 0.86 0.81
Relief device 1.95 1.8 1.98 1.79 1.72
LTC Restorative Process 4.28 1.8 2.03 1.88 1.67
Acute Restorative Process 3.69 0.47 0.39 0.58 0.44
Therapy minutes 324.7 357.8 439.6 357.3 349.6
Toilet Program 0.68 0.59 0.58 0.57 0.57
Weight Program 0.34 0.32 0.28 0.31 0.29
Percent Medicare 10.7 10.82 22.2 11.88 10.47
Percent Other 23.39 21.7 26.56 24.68 22.65
Percent Medicaid 65.91 67.48 51.24 63.45 66.88
Acuindex 10.39 10.2 10.45 10.25 10.17
ADL index 10.13 10.02 10.23 10.03 9.98
Percent tube feed 0.07 0.06 0.06 0.06 0.06
Percent tracheotomy 0.01 0.01 0.01 0.01 0
Percent requiring suctioning 0.01 0.01 0.01 0.01 0.01
Percent requiring respiratory care 0.11 0.1 0.12 0.11 0.1
Percent requiring IV 0.01 0.01 0.02 0.01 0.01
Percent requiring injection 0.15 0.15 0.17 0.16 0.15
Quality care deficiency 1.782 1.928 1.749 1.777 1.929
Bladder decline 0.186 0.193 0.209 0.192 0.187
Bowel decline 0.169 0.176 0.188 0.176 0.168
ADL 4 point decline 0.165 0.172 0.17 0.169 0.17
Restraint 0.076 0.079 0.075 0.076 0.077
Long term care pressure ulcer 0.08 0.076 0.093 0.081 0.075
Post acute walking improvement 0.232 0.238 0.246 0.24 0.246
Post acute pressure ulcer 0.187 0.173 0.177 0.174 0.169
Ownership 0.71 0.76 0.8 0.77 0.76
Market Competition 0.13 0.14 0.14 0.16 0.14
Chain 0.52 0.6 0.66 0.63 0.58
Total beds 114 130 116 116 111







of the barrier. This study partially supports that theory. Certainly a nursing home would have to

incur significant cost to enter the rehab focus group thereby protecting this group with a relatively

high mobility barrier. However, the mobility barrier to the differentiator may not be as high since

these groups staffing levels, wages and cost are considerably low.

Since it appears that nursing homes that provide higher levels of technology can have low

costs and be profitable, managers with low performing strategy can improve their performance

providing more care with the staff that they have.

The study found that business level decisions of technology and resource utilization can be

used to identify strategies that lead to greater financial performance and lower cost. The study

identified cost leaders, rehab focus and differentiators groups as having significantly better

financial performance than the RN focus strategy and nursing homes that had no strategy. The

rehab focus group had the highest proportion of Medicare beds and other payer beds. Although

this supports prior studies that found that those facilities that had higher Medicare beds and

private pay beds perform well financially (Marlin et al. 1999 and Zinn et al. 1994), it found that

this strategy was not significantly better than the differentiation or the cost leadership. Therefore,

managers may be able to choose from one of these three alternatives and still have their facility

perform well financially.

One of the primary functions of managers is to control costs. With differentiators able to

provide a high degree of care at the same costs as the cost leaders raises the question of whether

there are cost savings in keeping staffing levels and staffing mixes low. Further research is needed

to know how the differentiators were able to provide all the services yet have similar costs.

Policy Implications

State and federal governments have a considerable interest in nursing homes since they

provide over 60% of the funding. In return for their investment, they expect quality care for the

residents. Therefore, it is in the government's best interest to know if there are some strategies that

96







hopefully help managers address the problems directly. In this study the results of these measures

were mixed.

When compared to other groups, the nursing care focus strategic group generally provided

better care to their long term care residents. The nursing care focus, however, also had poor post

acute outcomes. The fact that this group provided the least amount of minutes of therapy, had a

lower proportion of post acute patients, and had the poorest post acute quality provides further

support to classify this cluster as a nursing technology focus group rather than an overall

differentiator group. These results suggest that there may be a relationship between a strategic

group focus and the quality of care provided.

The other two focus strategic groups did not fair as well. The rehab focus strategic group

had a high level of rehab staffing and provided high amount of rehab technology. The specialty

care focus had a high level of RN staffing and provided care to residents with specific care needs.

It would be expected that the rehab focus group would have the better post acute care

outcomes than other groups. However, rehab focus strategic group had a higher incidence of

pressure sores and did not have better walking improvement than the specialty nurse, cost leaders

or nursing homes that lacked strategy. This was surprising since the rehab focus did not have

better outcomes even though they provided significantly more therapy. There are a number of

possible explanations.

The rehab focus strategic group had higher acuity than other strategic groups. Although the

quality measures used are risk adjusted and the case mix of the facility was controlled for, there

may be factors that may not be captured. Mukamel et al. 2008 has determined that the current

quality measures including pressure ulcers would benefit with a greater degree of risk adjustment.

The rehab focus group has a high proportion of Medicare patients and may have the expertise to

handle patients with more co morbidities that may not be captured by the risk adjustments of the

outcome variables. Similarly, different strategic groups had different types of acuity measures.

91









progress. The difference in reimbursement rate between the RUG levels is significant (MDS

2007).

Higher rehab RUG reimbursement levels are associated with higher therapy utilization

(and higher reimbursement rates) but are not necessarily associated with higher acuity (MedPAC

2007). For example, a post acute patient with a knee replacement that only requires supervision

with his ADLs may be classified at an ultra high RUG level because the patient can tolerate high

volume of therapy, not because the patient has higher acuity. In contrast, a post acute patient that

is dependent and has low tolerance to activity will only classify at a high or medium RUG level.

Based on acuity scores, the second patient would be considered more acute but the facility would

be compensated less. Therefore, nursing homes that have a higher proportion of residents in the

highest RUG level will have higher revenue but they will not necessarily have higher comparable

costs. MedPAC (2007) found that nursing homes increased the proportion of their residents

receiving the two highest paying RUG levels (ultra and very) by 50% between the years 2000

and 2005.

Nursing Home Environment

Similar to most health care organizations, nursing homes exist in a turbulent, complex and

uncertain environment. Since the Balanced Budget Act 1997 (BBA), the nursing home industry

has been dealing with a number of cost control mechanisms imposed by federal and state

government payers. This problem affected the whole industry since government payers like

Medicare, Medicaid and the VA account for over 62% of the nursing home industries' revenues

(MedPAC 2003). Public payers often dictate price of a resident's per diem stay to the nursing

home. By removing price from the equation, nursing homes must find other ways to compete for

residents that provide the nursing home with the highest revenue and that have the lowest cost.







indicate a cost leader because it indicates that the nursing home is substituting CNAs for RNs. In

addition, on observation of the structural features of this group, the cost leaders also tended to be

larger and have lower acuity. It is clear that this group's strategy was to keep cost under control by

providing a low amount of service.

Finally, the lack of strategy group was identified because it used a median amount of

resources with almost no technology. In addition, the nursing homes that lacked strategy appeared

not to have any specialty services.

By using technology and resource deployment, this study found that strategic groups not

only exist, but can be classified using Porter's generic strategies. The next question is whether

strategic-performance relationships exist between the different strategic groups and quality, costs

and financial performance.

Quality

This study identified that the nursing care focus (differentiator), specialty care focus and

rehab focus strategic groups had lower quality of care deficiencies than the nursing homes that

were in the cost leader strategic group or the nursing homes that lacked strategy. This finding

provides some evidence that focusing on providing quality either by providing technology or

having better staffing mixes can result in some better outcomes. However, only quality of care

deficiencies had consistent results through these three strategic groups. Quality of care

deficiencies is a general measure of the facilities' quality and can relate to anything from staff that

is inadequately trained to poor care practices. The Nursing Home Compare website lists and

describes all of facilities deficiencies from the most recent state survey. It is a good general

measure of quality and has been used in past research (Harrington et al. 2000, 2001). One

limitation of deficiencies is that it does not highlight specific care outcomes. Other measures like

the QIs that were derived by Abt are specifically designed to identify the specific quality issues

like pressure ulcer prevalence and ADL decline. Knowing the specific quality areas would









homes have been faced with considerable challenges that include increased government

regulation (GAO 2001, Walshe et al. 2002), increased litigation (Johnson et al. 2003) and

increased competition (Weech-Maldonado et al. 2003).

In order to survive, nursing home managers need to identify effective strategies so that

they can continue to provide a very valuable service to a population with few options. Some of

those strategies include providing high quality care, keeping costs as low as possible and

focusing on a specific market segment of the industry. These different types of strategy do not

only relate to better financial performance, but can also influence quality.

To fully understand these different types of strategies, it is necessary to have a clear grasp

of what nursing homes are, to have insight on what their role is in the health care industry and to

comprehend how they function in an uncertain environment.

What Are Nursing Homes and What Are Their Roles in the Health Care System?

A nursing home is a place for people who do not need to be in a hospital but cannot be

cared for at home. Most nursing homes have nursing aides and skilled nurses on hand 24 hours a

day (National Institutes of Health 2008). The stay in a nursing home can be permanent or can be

temporary.

Nursing homes have two major functions. The first function is to provide long term care to

those individuals who require more care than can be provided by their families or can be

provided by any other type of placement. The desired outcome for the long term care resident is

that during their stay, their health status remains at least at the same level as it was on admission.

It is understood, however, that the longer the resident remains in a facility, the more likely they

will experience a decline in their health status. The decline occurs as a result of the normal

progression of the pre existing medical conditions that typically precipitated the admission to the




Full Text

PAGE 1

AN EXAMINATION OF STRA TEGIC GROUP MEMBERSHIP AND TECHNOLOGY IN THE NURSING HOME INDUSTRY By ALEXANDRE LABERGE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009 1

PAGE 2

2009 Alexandre Laberge 2

PAGE 3

This document is dedicated to the four women in my life 3

PAGE 4

ACKNOWLEDGMENTS There are a number of people who provided me with the opportunity to achieve this goal. To begin with, I would like to say thanks my mother who not only supported and encouraged me to learn but also encouraged me to finish what I started. I would like to thank my father who inspired me to think outside of the box. I would like to thank Dr. Weech-Maldonado for his guidance, generosity and tremendous patience. And finally I would to thank my wife, whose unconditional love, support and unde rstanding made it all possible. 4

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF TABLES ...........................................................................................................................8 LIST OF FIGURES .......................................................................................................................11 ABSTRACT ...................................................................................................................................12 CHAPTER 1 INTRODUCTION................................................................................................................. .14 What Are Nursing Homes and What Are Their Roles in the Health Care System? ...............15 Long Term Care Resident Reimbursement .....................................................................16 Medicaid ..........................................................................................................................16 Private Pay .......................................................................................................................18 Medicare ..........................................................................................................................18 Post Acute Care Reimbursement ............................................................................................19 Nursing Home Environment ...................................................................................................20 The Importance of Focusing on Strategy ................................................................................21 Purpose ...................................................................................................................................22 Research Questions .................................................................................................................22 2 CONCEPTUAL FRAMEWORK...........................................................................................23 Strategic Groups .....................................................................................................................23 Nursing Home Technology .....................................................................................................26 Nursing Home Resources .......................................................................................................28 Scope and Resource Commitment ..........................................................................................31 Scope Using Technology .................................................................................................32 Resource Commitment ....................................................................................................33 Porters Generic Strategies .....................................................................................................34 Strategic Group Model Using Scope (T echnology) Commitment, Resources Commitment and Porters Generic Strategy Theory ...........................................................36 Differentiator ...................................................................................................................37 Cost Leadership ...............................................................................................................37 Best Cost ..........................................................................................................................38 Focus ................................................................................................................................38 Lack of Strategy ..............................................................................................................39 Measuring Performance ..........................................................................................................39 New Contributions ..................................................................................................................41 Hypothesis ..............................................................................................................................41 5

PAGE 6

3 METHODOLOGY.................................................................................................................4 3 Data .........................................................................................................................................43 Population ...............................................................................................................................44 Overview of Methodology ......................................................................................................45 Hypothesis #1: Nursing homes that have a strategy can be categorized into one of four strategic groups: Differentiator, cost leadership, focus or best cost. ..................................45 Scope ...............................................................................................................................45 Nursing Technology ........................................................................................................45 Rehab Technology ...........................................................................................................52 Resource Commitment ....................................................................................................54 Hypothesis #1 Analysis ..........................................................................................................56 Factor Analysis ................................................................................................................57 Factor Matrix ...................................................................................................................58 Number-of-Factors Problem ............................................................................................59 Rotation of the Factor Structure ......................................................................................59 Internal Consistency of Measures ...................................................................................60 Cluster Analysis ...............................................................................................................60 Strategic Groups Defined ................................................................................................63 Determine the total variance value ...........................................................................63 Cluster value .............................................................................................................63 Cluster rank score .....................................................................................................64 Analysis of Variance .......................................................................................................65 Hypothesis #2: Nursing Home s in the Differentiator Strategic Group will have the Highest Quality. ..................................................................................................................66 Dependent Variables: ......................................................................................................66 Independent Variables .....................................................................................................67 Control Variables .............................................................................................................67 Analysis ...........................................................................................................................68 Hypothesis #3: Nursing Homes in the Focus Strategic Group will provide Higher Quality in the area that they are focuse d on compared to other Strategic Groups. .............69 Dependent Variables .......................................................................................................69 Independent Variables .....................................................................................................69 Control Variables .............................................................................................................70 Analysis ...........................................................................................................................70 Hypothesis #4: Nursing homes in the Cost Leadership Strategic Group will have the lowest costs. ........................................................................................................................70 Dependent Variables .......................................................................................................70 Independent Variables .....................................................................................................70 Control Variables .............................................................................................................71 Analysis ...........................................................................................................................71 Hypothesis #5: Differentiation, Co st Leadership, Focus and Best Cost Strategic Groups will have better Financial Performance than Nursing Homes that Lack a Strategy ...........71 Dependent Variables .......................................................................................................71 Independent Variables .....................................................................................................72 Control Variables .............................................................................................................72 Analysis ...........................................................................................................................72 6

PAGE 7

4 RESULTS...................................................................................................................... .........74 Hypothesis #1 .........................................................................................................................74 Hypothesis #2 .........................................................................................................................82 Hypothesis #3 .........................................................................................................................83 Hypothesis #4 .........................................................................................................................84 Hypothesis #5 .........................................................................................................................85 5 DISCUSSION................................................................................................................... ......88 Quality ....................................................................................................................................90 Cost .........................................................................................................................................93 Financial Performance ............................................................................................................94 Managerial Implications .........................................................................................................95 Policy Implications .................................................................................................................96 Conclusion ..............................................................................................................................98 Limitations ..............................................................................................................................98 APPENDIX: BACKGROUND TABLES................................................................................100 LIST OF REFERENCES .............................................................................................................125 BIOGRAPHICAL SKETCH .......................................................................................................130 7

PAGE 8

LIST OF TABLES Table page 4-1 Internal consistency of the composite variables ................................................................76 4-2 Individual variables used in cluster analysis that we re not included in a factor variable. ..............................................................................................................................76 4-3 Rank scores scope and resource variables .........................................................................78 4-4 Rank scores of struct ure and market factors ......................................................................79 4-5 Strategic group quality of care with Differentiator / nursi ng care focus cluster as the reference group ..................................................................................................................82 4-6 Strategic group quality of care with re hab focus cluster as the reference group...............83 4-7 Strategic group quality of care with Speci alty care focus clus ter as the reference group ..................................................................................................................................84 4-8 Strategic groups and costs with cost l eader as reference group regression results ............85 4-9 Strategic groups and predicted costs ..................................................................................85 4-10 Odds Ratio of Nursing Home Strategy ..............................................................................86 4-11 Probability of nursing homes being in the highest operating margin tier ..........................86 4-12 Probability of nursing homes bei ng in the highest total margin tier ..................................87 A-1 Eigenvalues of scope (technology) factor analysis ..........................................................100 A-2 Factor analysis with 10 factor s of scope (technology) variables .....................................102 A-3 Factor analysis with 10 factor s of scope (technology) variables .....................................104 A-4 Eigenvalues of scope (technology) factor analysis ..........................................................106 A-5 Factor analysis of resource staffing intensity with three factors ......................................106 A-6 Ward's minimum variance cluster analysis ......................................................................107 A-7 Cluster History .................................................................................................................107 A-8 Determination of rank scores table ..................................................................................108 A-9 Descriptives of dependent variables, independent variables and control variables .........109 8

PAGE 9

A-10 Negative binomial of quality of care defi ciencies with differentiators as reference group ................................................................................................................................110 A-11 OLS regressions of root transformed bladder decline with differentiators as ref. group ................................................................................................................................110 A-12 OLS regressions of root tr ansformed bowel decline with differentiators as ref. group ...111 A-13 OLS regressions of root transformed ADL decline with differentiators as ref. group ....111 A-14 OLS regressions of log transformed rest raints with differentiators as ref. group ............112 A-15 OLS regressions of root transformed LTC ulcers with di fferentiators as ref. group .......112 A-16 GLM with gamma distribution with PA walking improve with differentiators as ref. group ................................................................................................................................113 A-17 OLS regressions of root tr ansformed PA skin ulcers with differentiators as ref. group ..113 A-18 Negative binomial of quality of care defi ciencies with rehab focus as reference group .114 A-19 OLS regressions of root transformed bladder decline with rehab focus as ref. group .....114 A-20 OLS regressions of root transformed bowel decline with rehab focus as ref. group .......115 A-21 OLS regressions of root transformed ADL decline with rehab focus as ref. group ........115 A-22 OLS regressions of log transformed restraint with rehab focus as ref. group .................116 A-23 OLS regressions of root transformed skin ulcer with rehab focus as ref. group .............116 A-24 GLM of PA walking improve with gamma distribution with rehab focus as ref. group .117 A-25 OLS regressions of root tr ansformed PA skin ulcers with rehab focus as ref. group ......117 A-26 Negative binomial of quality of care defi ciencies with specialty care as reference group ................................................................................................................................118 A-27 OLS regressions of root tr ansformed bladder decline with specialty care focus as ref. group ................................................................................................................................118 A-28 OLS regressions of root transformed bowel decline with specialty care as ref. group ...119 A-29 OLS regressions of root transformed ADL decline with specialty care as ref. group .....119 A-30 OLS regressions of root transformed restraint with sp ecialty care as ref. group .............120 A-31 OLS regressions of root tr ansformed LTC skin ulcer with specialty care as ref. group ..120 9

PAGE 10

A-32 GLM with gamma distribution PA walking improve with specialty care as ref. group ..121 A-33 OLS regressions of root tr ansformed PA skin ulcers with specialty care as ref. group ..121 A-34 GLM regressions of nursing hom e cost with gamma distribution ...................................122 A-35 Odds of nursing home of being in the highest operating margin tier ..............................122 A-36 Odds of nursing home of being in the highest total margin tier ......................................123 A-37 Correlations of independent variables .............................................................................124 10

PAGE 11

LIST OF FIGURES Figure page 2-1 The nursing home process .................................................................................................30 2-2 Strategic Group Model Based on Scope and Resource commitment ................................36 4-1 Strategic group model based on scope and resource commitment in the nursing home industry.....................................................................................................................81 A-1 Scree plot of scope (technology) factor analysis ............................................................101 A-2 Scree plot of resource staffi ng intensity from factor analysis .........................................106 11

PAGE 12

Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy AN EXAMINATION OF STRA TEGIC GROUP MEMBERSHIP AND TECHNOLOGY IN THE NURSING HOME INDUSTRY By Alexandre Laberge May 2009 Chair: Robert Weech-Maldonado PhD Major: Health Services Research, Management and Policy Purpose: The purpose of the study was to examine if strategic groups in the nursing home industry can be determined by examining how faci lities apply their technology and how facilities commit their resources. The study also determined if these groups can be defined using Porters generic strategies and whether the different groups have varying strategic-performance relationships. Methodology : The study began by performing a factor analysis using scope (technology) and resource commitment variables from the Minimum Data Set (MDS) and the Online Survey Certification and Reporting (OSCAR) system. The newly defined variables were used in a cluster analysis to identify the different cluste rs. The clusters were classified using Porters generic strategy model. Once the strategic group s were identified, the study analyzed the strategic-performance relationships using negativ e binomial, ordinary least squares regression (OLS), generalized linear model (GLM) with ga mma distribution and ordered logit analytical tools. Quality dependent variab les included quality of care defi ciencies, long term care (LTC) pressure sore prevalence and post acute pressure ulcer incidence, LTC ac tivities of daily living (ADL) decline, post acute walking improveme nt, LTC bowel and bladder decline and LTC restraint. Financial measures in cluded cost per resident day, oper ating margin and total margin. 12

PAGE 13

13 Results: The study identified four strategic groups (differentiator, cost leader, rehab focus, and specialty care focus) and a lack of strate gy group. Differentiators, rehab focus and specialty care focus groups had less quality of care deficien cies than the lack of strategy group and cost leaders. Differentiators generally had better l ong term care quality outcomes compared to other strategic groups and the lack of strategy group. Cost leaders, diffe rentiators and lack of strategy groups had lower cost than the rehab focus a nd specialty care focus groups. Cost leaders, differentiators and rehab focus group had better financial performance than the specialty care focus group and lack of strategy group. Conclusions: The study indicates that using technology and resour ces commitments is an effective way of identifying strategies groups. The study also found that some strategic groups have better financial performance than other st rategic groups or than nursing homes that have no strategy at all. Nursing homes that provide a hi gh level of technology had better quality, without increased cost or sacrificing fi nancial performance. The study also suggests that nursing homes that service residents with high cost conditions may not be adequately reimbursed.

PAGE 14

CHAPTER 1 INTRODUCTION Nursing homes, like other health care orga nizations face many challenges. They have a limited ability to dictate their pr ice and to control their costs. Nursing homes also face higher competition from the rapid rise in home health ag encies and assisted living facilities. To deal with these challenges, nursing homes must devise and implement effective strategies if they are to remain profitable. These strategies can va ry from focusing on providing quality service, controlling cost or identifying niche market s. Although there are ove r 16,000 nursing homes in the United States, the number of different strategi es to choose from is sm all. It is therefore possible that nursing homes can be categorized into different groups that have similar strategies. Nursing homes are an important component of the health care industry consisting of 8% of the United States health care expenditures (Cen ters for Medicare and Medicaid 2007). The nursing home industry provides care for 1.6 million of the frailest people in the United States (Center for Disease Control and Prevention 2007). A study by the U.S. Department of Health and Human Services says that people who reach the age of 65 have a 40 percent chance of entering a nursing home. Of those who enter a nursing home, 10 percent of them will remain there five years or more (Centers for Medicare and Medicaid 2007). Expenditures for nursing home care skyrockete d in the years following the implementation of the hospital prospective payment system of 19 83. The industry grew from 50 billion dollars in 1990 to 90 billion dollars in 1997 (Centers for Medicare and Medicaid 2007). However, this changed after the implementation of the Bala nced Budget Act of 1997 (BBA 1997) where the industry only grew 10% between the years 1999 and 2004. The BBA cuts in reimbursement were so severe that a number of nursing homes went bankrupt including some of the nations largest nursing home chains (GAO 2000). Since the implem entation of the BBA, the remaining nursing 14

PAGE 15

homes have been faced with considerable challenges that include increased government regulation (GAO 2001, Walshe et al. 2002), incr eased litigation (Johnson et al. 2003) and increased competition (Weech-Maldonado et al. 2003). In order to survive, nursing home managers ne ed to identify effective strategies so that they can continue to provide a very valuable service to a population with few options. Some of those strategies include providing high quality care, keeping costs as low as possible and focusing on a specific market segment of the indu stry. These different types of strategy do not only relate to better financ ial performance, but can also influence quality. To fully understand these different types of stra tegies, it is necessary to have a clear grasp of what nursing homes are, to have insight on what th eir role is in the health care industry and to comprehend how they function in an uncertain environment. What Are Nursing Homes and What Are Their Roles in the Health Care System? A nursing home is a place for people who do not need to be in a hospital but cannot be cared for at home. Most nursing homes have nursing aides and skilled nurses on hand 24 hours a day (National Institutes of Health 2008). The stay in a nursing home can be permanent or can be temporary. Nursing homes have two major functions. The firs t function is to provide long term care to those individuals who require more care than can be provided by their families or can be provided by any other type of placement. The desired outcome for the long term care resident is that during their stay, their health status remains at least at the same level as it was on admission. It is understood, however, that the longer the resident remains in a facility, the more likely they will experience a decline in their health status The decline occurs as a result of the normal progression of the pre existing medical conditions that typically precipitated the admission to the 15

PAGE 16

facility in the first place. The desired outcome for long term care residents is to maintain their health status for as long as possible. The second major function of a nursing home is to provide short term care (also known as post acute or sub acute) for those residents that require skilled se rvices after a qualified hospital stay. To qualify for a post acute st ay, residents must have either a skilled nursing care need (e.g. IV antibiotics) or a rehab care need based on th e recommendations of the hospital therapists. The goal of post acute care is to improve the residents health stat us so that the resident can be discharged home or to the most appropriate leve l of care. The discharge options are quite broad and can range from patients going home by themse lves to becoming long term residents in a nursing home. Discharge from a skilled nursing reha bilitation facility (als o known as a post acute or subacute facility) can occur wh en the patient reaches their highest level of function or the stay is no longer covered by the payer. The desired out come for a rehabilitation resident is a better health status. Long Term Care Resident Reimbursement Reimbursement for long term care is primarily provided by Medicaid or by the residents themselves (private pay). A small proportion of residents are covered by other payers like the Department of Veteran Affairs (VA) and private insurance. A lthough Medicare does not pay for a long term care residents stay in a nursing home, Medicare does cover some ancillary services like physical or occupational therapy. Medicaid Medicaid is the major payer of nursing home long-term care services. Over 50% of nursing homes revenues and 70% of nursing home beds are covered by Medicaid (Rhoades et al. 2000). Swan et al. 2000 describe five rate setting methodologies that state Medicaid programs use to 16

PAGE 17

pay nursing homes: retrospective, prospective clas s, prospective facility-s pecific, adjusted, and combination. Retrospective payments freely adjust rates to current costs. In contrast, prospective payment rates are not fully adjusted to cost duri ng the costs year. Rates are usually set based on prior year costs. Prospective cl ass rates are the same for all nursing homes in the state while prospective facility-specific are set based on the costs for each f acility. Adjusted reimbursement is similar to retrospective reimbursement however adjusted systems allow interim rates to increase during a year but not to fully reflect costs. Finally, there are also combination reimbursement methods which are a mixture of retrospective and prospective payment systems (Swan et al. 2000). As of 1997, only one state used retrospective reimbursement, 25 used prospective, 21 used adjusted and 3 us ed combinations (Feng et al. 2006). Since the late 1990s, most state Medicaid prog rams have used a prospective per diem system to reimburse nursing homes for all the cost that is incurred in providing the care to the resident in the facility (Feng et al. 2008). With this reimbursement methodology, a nursing home that provides a lot of nursing services will not ne cessarily be compensated more than a facility that provides fewer nursing services. An increasing number of states have m odified their reimbursement methodology by making case mix adjustments to their per diem ra te. This allows nursing homes to collect higher rates for residents that have higher acuity and that are likely to cost the facility more to care for them. The number of states that use case mix reimbursement increased from 19 in 1991 to 35 in 2004 (Feng et al. 2006). The best-known and most widely used case mix methodology is the Resource Utilization Groups system (RUGs), currently in its third version (Fries et al. 1994). This system classifies 17

PAGE 18

residents into homogeneous categ ories based on their estimated re source utilization. Associated with each of these categories is a case-mix index, which represents, at least relatively, the time, or cost, of caring for the average resident in the group. A higher case-mix index indicates a greater degree of complexity and, consequently, a greater need for input resources. Under this system, nursing homes with a higher case-mix in dex, on average, would be reimbursed a higher rate (Zinn et al. 2008). Private Pay Private paying residents reimburse approxima tely 18% of the beds (GAO 2000). Private pay reimbursement is commonly based on a per diem rate, which is on average 1.4 times the amount that Medicaid reimburses (Grabowski 2004). Medicaid only pays for those residents w ho do not have enough funds to pay for the nursing home stay themselves. Residents who have some assets will pay the private pay rates until their personal funds are exhausted. Once the resident has spent down their assets, they usually qualify for Medicaid. Medicare There are some processes of care provided to th e long term care residents that are paid for by Medicare (Wodchis et al. 2004). Physical, occupational and speech therapy are three services that can be provided to long term care residents by nursing homes. These services are reimbursed over and above the Medicaid per diem rate or the private pay rate. Therapy services are paid for under Medicare part B and is cons idered an outpatient service. Although the price is set by the Center for Medicaid and Medica re (CMS) however the amount of therapy services provided was not limited until 2005. In 2005, the CMS implemented a cap on therapy services that limited residents to a maximum annual benefit of 1700 dolla rs of physical and speech therapy combined and 1700 dollar cap on occupational thera py. Initially passed by congress in 1997, 18

PAGE 19

implementation of the caps was delayed by a num ber of moratoriums. Even in 2005, after the caps were implemented, nursing ho mes were able to circumvent the caps if a resident had therapy needs that were deemed medical necessary (APTA 2008). Post Acute Care Reimbursement Reimbursement for post-acute care is prim arily reimbursed for by Medicare; however it can also be covered by private insurance and th e VA. Approximately 12% of the beds in nursing homes are occupied by rehabilitation patients paid for by Medicare (GAO 2000). These patients usually have experienced a declin e in their health status and re quire rehabilitation before they can return either home or to the next appropriate level of care. Medicare reimbursement uses a prospective pa yment system. Similar to Medicaid case mix reimbursement, residents are classified usi ng the Resource Utilizati on Groups (RUG). These groups are commonly determined by the amount of th erapy and nursing services that the resident requires. For example, a facility will receive a higher daily rate if the resident gets more therapy. Each rehab RUG level has a minimum amount of therapy minutes that have to be provided within 7 days prior to a given assessment refe rence date. For example, for a patient to be classified as an ultra high reha b RUG level, they would have to receive 720 minutes of therapy during the 7 day assessment period. Patients ar e assessed a maximum of 5 times depending on the amount of time they remain at the facility for rehab. The assessments include the 5-day, 14 day, 30 day, 60 day and the 90 day assessment. The 5 day assessment identifies the RUG level for day 1 to 14. The 14 day assessment identifies the RUG level for day 15-30. Each subsequent assessment identifies the RUG level for the fo llowing period. Although the each resident is eligible for up to 100 days of rehab per medical incident, most patients are discharged from subacute services well before the 100 day lim it because they go home or are no longer making 19

PAGE 20

progress. The difference in reimbursement rate between the RUG levels is significant (MDS 2007). Higher rehab RUG reimbursement levels are a ssociated with higher therapy utilization (and higher reimbursement rates) but are not necessarily associat ed with higher acuity (MedPAC 2007). For example, a post acute patient with a kn ee replacement that only requires supervision with his ADLs may be classified at an ultra hi gh RUG level because the patient can tolerate high volume of therapy, not because the patient has higher acuity. In contrast, a post acute patient that is dependent and has low tolerance to activity will only classify at a high or medium RUG level. Based on acuity scores, the second patient would be considered more acute but the facility would be compensated less. Therefore, nursing homes th at have a higher proportion of residents in the highest RUG level will have highe r revenue but they will not nece ssarily have higher comparable costs. MedPAC (2007) found that nursing homes increased the proportion of their residents receiving the two highest paying RUG levels (ultra and very) by 50% between the years 2000 and 2005. Nursing Home Environment Similar to most health care organizations, nursing homes exist in a turbulent, complex and uncertain environment. Since the Balanced Budget Act 1997 (BBA), the nursing home industry has been dealing with a number of cost c ontrol mechanisms imposed by federal and state government payers. This problem affected th e whole industry since government payers like Medicare, Medicaid and the VA account for over 62% of the nursing home industries revenues (MedPAC 2003). Public payers ofte n dictate price of a residents per diem stay to the nursing home. By removing price from the equation, nursi ng homes must find other ways to compete for residents that provide th e nursing home with the highest revenue and that have the lowest cost. 20

PAGE 21

Nursing homes also face increased competition from growing industries like assisted living facilities and home health agenci es that tend to draw the less dependent (and more profitable) private pay residents. This ultimately leaves nursing homes with the re sidents who have higher medical acuity, greater functional dependence and have higher care costs (Weech-Maldonado et al. 2003). With the limited ability to control their reve nue and with the increased competition created by substitutes, nursing homes find themselves in a precarious position. For nursing homes to thrive (or even just survive), it is necessary fo r them to identify strategies that allow them to achieve competitive advantage. The Importance of Focusing on Strategy Government payers and interest groups ar e interested in persuading nursing homes to provide high quality of care. One way to encourage facilities to provide better quality of care is to demonstrate that better quality results in be tter financial performance. Some recent studies have determined a positive re lationship between financial pe rformance and quality of care (Weech-Maldonado et al. 2003, 2008). Even though nursing homes may have the incenti ve to provide better quality, it is also necessary for the nursing home to determine the best strategies to be able to provide high quality care. Without incentives and a strategy, a nursing home will have some difficulty achieving these desired outcomes. Managers are interested in achieving above normal profits. Stra tegies are plans related to the mission and the vision of an organization that provide leadership with the means to achieve high performance. Like any industry, nursing homes attempt to gain competitive advantage by selecting and implementing an effective strategy that the competitors ar e not able to reproduce (Barney 1991). Because the number of truly distinct strategies available in any industry is likely 21

PAGE 22

22 to be small, managers decide not so much how to be unique, but rather which group of competitors their strategies should be similar to (Marlin et al. 2002). The strategic group model is one approach that can help managers and researchers identify the best strategy. Purpose The purpose of the study is to examine if strategic group struct ure of the nursing home industry can be determined by how facilities focus their technology a nd how facilities commit their resources. The study will also determine if these groups can be defined using Porters generic strategies and will examine whether th e different groups have different strategicperformance relationships. Research Questions What strategic groups exist in the nursing home industry? How does nursing home strategic group memb ership affect quality of care? How does nursing home strategic group membership affect costs? How does nursing home strategic group member ship affect financial performance?

PAGE 23

CHAPTER 2 CONCEPTUAL FRAMEWORK This study will begin by defining strategic grou p theory, Porters generic strategy theory, technology and Cool and Schendels scope and re source commitment theory. The study will then explain how these theories are integrated to genera te a model that uses Porters generic strategies to define strategic groups in the nursing home. Finally, the study will define how the different strategic groups can have differe nt strategic-performance relations hips as it pertains to quality, cost and financial performance. Strategic Groups The concept of strategic groups was introduced by Hunt (1972) in his thesis that examined the appliance industry. He discovered that there was less competitive rivalry than what industry concentration ratio s suggested. He attributed this to the existence of subgroups within the industry that effectively reduce the number of competitors in each market. Caves and Porter (1977) expanded the theory by defini ng strategic groups as a set of fi rms that face similar threats and opportunities that are different from the threats and opportunities faced by other firms in the industry. Strategic group theory stipulates that within an industry, there could be groupings of organizations that have very different stra tegies and yet may still have good financial performance. This theory was a shift in thinki ng from the industrial organizational theory where it was generally believed that there was only one right way of doing things (Leask 2007). Analyzing strategic groups gives insight to different competitors approaches to the marketplace (Harrigan 1985). The number of strategic group s within an industry corresponds to the number of unique strategies within that industry (Marlin et al. 1999). St rategic groups are persistent strategic characteristics of an industry, which are protected by mobility barriers (Porter 1980). 23

PAGE 24

Movement between groups is limited by mobility barriers, which are associated with the cost of moving from one strategic group to another (Porter 1980). Conceptually, mobility barriers behave similar to barrier s to entry in Porters five fo rces analysis framework (Barney 1991). This framework describes entr y of new organizations to the ma rket place as threats to the profitability of the incumbent organizations. By establishing barriers to entry, an organization can block out new competitors from entering the market, thereby maintaining their above normal margins (Caves and Porter 1977). The existence of mobility barriers can a llow two groups within the same industry to exist without directly co mpeting with each other (Leask 2007). In the nursing home industry, facilities that are a part of the strategic gr oups concerned with efficiency and keeping operating cost as low as possible are different from those nursing homes that attempt to differentiate themselves by provi ding better care (Marlin 1999). Although these strategies are very different, it is possible that nursing homes in either one of these strategic group types will have good financial performance. For an organization to move from one strategic group to anot her, it would be necessary for them to change their business strategy and put th emselves at risk of not recouping the resources they invest (Porter 1980). These costs offer pr otection to group member s by discouraging entry of rivals into the group (Porter 1980). An ex ample would be an organization from a cost leadership strategic group trying to move to a differentiating strategic group. This organization would have to make some considerable investment and they would also have to change the organizational culture from being cost focu s to providing more and better services. The cost of movement between strategic groups can vary. For a facility to move from cost leadership to focus may require significantly le ss resources than moving from cost leadership to differentiator. Therefore the highe r the cost to change strategi c groups, the higher the mobility 24

PAGE 25

barriers. Higher barriers offer higher insularity to the member s of the groups, thus limiting competition and group membership (Marlin et al. 2002). The height of the mobility barrier is expected to be directly rela ted to the performance of th e strategic group (Porter 1980). Since the introduction of the strategic group concept by Hunt (1972) and the additional contributions by Caves and Porter (1978), there have been a significant number of empirical studies that have examined stra tegic groups and their impact on firm performance. Studies that have examined the beer (Tremblay 1985), banki ng (Mehra 1996), air line industry (Kling et al.1988), and the pharmaceutical industry (Figen baum et al. 1992, Leask 2007) have all found the existence of strategic groups with strategicperformance relationships. Direct care health care industries have also been examined using strate gic group theory in the hospital (Marlin et al. 2002, Ketchen et al. 2003) and the nursing hom e industry (Castle 2003, Ma rlin et al. 1999, Zinn et al. 1994). The three studies that have examined strategic groups in the nursing home industry all have found significant difference between the groups and have also found some strategyperformance relationships. Castle 2003 used a surv ey that asked the leadership of the nursing homes to self characterize themselves using Mi les and Snows categorie s (1990): prospectors, analyzers, reactors or defenders. He found that the Miles and Snows prospectors had higher levels of financial performance and higher qua lity. Zinn et al. 1994 a nd Marlin et al. 1999 attempted to determine if th ere was a relationship between strategic group membership and nursing home performance. Both studies used scope and resource commitment as defined by Cool and Shendel (1988) to determine their clusters and found the existence of succinct strategic groups in the nursing home industry. 25

PAGE 26

Nursing Home Technology Hulin and Roznowski (1985) defined technology as the physical and knowledge processes by which materials in some forms are transfor med into outputs. In the nursing home industry, technology generally refers to the care processes provided by the staff to the residents. In order to provide these processes, an organization requires sufficient resources. For example, in the nursing home industry, a facility that fails to have adequate staffi ng levels (the primary resource in the nursing home industry) is not going to be able to adequately provide the care processes (i.e. technology) that are necessary to have good patient care outcomes. The Centers for Medicare and Medicaid (CMS) statute 483.25 dictate that each resident must receive from the facility the necessary care and services to attain or maintain the highest practicable physical, mental, and psychosocial we ll-being, in accordance with the comprehensive assessment and plan of care (CMS 2008). Nursing homes use a number of different techno logies that allows them to address the multiple needs of their residents. Each of th ese technologies addresses a specific care need. There are two broad categories of nursing hom e technology: nursing technology and rehab technology Nursing technology: Nursing technology includes all of the care processes that are provided by the nursing staff. These processes of care are required so that the resident can maintain or improve their health status. Nursing provides routine care such as bathing, providing medication and repositioning the residents. Nursing technology also includes less routine services like skin ulcer care or catheter placement. Most routine services require that the nursing home establish a program in order for the proce sses to be provided properly. For example, even though a routine technology like rest orative ambulation may only requ ire one nursing assistant to walk with the resident, a restorative program must be established. The program requires that 26

PAGE 27

there is an MD order, that the goals are generated by the physical th erapist and the program requires some oversight to ensure that the service is provided. The responsibility of the organizational features of a restorativ e program is placed on the licensed nurse. Failure to provide these processes of care can result in a decline in the residents health status. One example of nursing technology is pr essure sore prevention. If pressure sore preventative processes are not provided, a resident will be more likely experience a decline in their health status because they can develop a pressure sore. Having a patient on a turning schedule every 2 hours or providing them with a pressure relieving device reduces the likelihood of a pressure sore will occur. Turning schedules relieve pressure and allow blood to flow to the skin. Pressure relieving devices distribute the pressure over a great er surface area and lower the pressure on the high-risk br eakdown areas like the coccyx (C MS 2008). A nursing home that puts emphasis on nursing technology may be tr ying to get better nursing outcomes. More processes of care and better outcomes are indicato rs of better quality (Sch nell et al. 2004). Most of nursing home technology is provided by the nu rses. A facility focused on better quality may be trying to differentiate themselves from their competitors. By differentiating themselves from their competitors, a facility may get a greater pr oportion of private pay residents and/or higher profit margin Medicare residents. In addition, be tter outcomes can also lead to lower costs (Weech-Maldonado et al. 2003). Rehabilitation technology: Rehabilitation technology can be provided to either post acute or long term care residents. These processes of care consist primar ily of therapy services and are geared to improve the residents health stat us. Rehabilitation technology is not considered routine because of its dynamic nature. Therapy serv ices vary based on the needs of the residents. Some residents may have difficulty with transfer ring from the chair to the toilet while others 27

PAGE 28

have difficulty walking. Both these issues would be identified and addressed by physical therapy through an assessment and then would be treated with a specifically program tailored to the individual residents needs. Even the treatment protocol can change over time. Initially the resident may require treatment to increase their st rength and balance so they are able to stand. Once the patient is able to stand, physical th erapy will then focus on balance and walking. Therapy assessment can only be performed by a licensed therapist while therapy treatments can only be performed by either a licensed therapist or licensed therapy assi stant. Nursing homes have an incentive to use rehab technology ove r nursing technology since rehab technology provides better revenue streams. MedPAC 2007 has warned that the current Medicare reimbursement system favors over utilizati on of therapy. Although it may be fiscally advantageous to provide rehabili tation, it only serves as a few of the processes of care that a resident receives on any given da y. The therapist actually only pl ays a small role in the overall care of residents when compared to role of the nurses. Nursing Home Resources According to the resource based view of the firm, performance differences across firms can be attributed to the variance in the firms res ources and capabilities (Hitt et al. 1990). Resources that are valuable, unique and difficult to imitate can provide the basis for firms competitive advantages (Amit & Schoemaker, 1993; Barney 1991). Firms employ both tangible resources (such as buildings and financia l resources) and intangible res ources (like knowledge and brand equity) in the development and implementation of strategies. Human capital has long been argued as a critical resource in most firms (Pfeffer, 1994). In the nursing home setting, the tangible resources like human cap ital and intangible resources like staff expertise are especially important because most of the nursing homes expe nses are the nurses and therapist salaries. In addition, a facilities technology is dependent of the amount and quality of their health care staff. 28

PAGE 29

Human capital provides the physical care and po ssesses the knowledge of the facility. Knowledge is one of the most critical resour ces that a firm possesses (Hitt et al. 1991). The nursing home requires adequate resources if they are going to be able to provide their technology effectively. The correlation between staffing levels and technology provided is expected to be high. The amount of nursing staff and the skill of the nursing staff will influence the quality outcomes (Harrington et al. 2000, We ech-Maldonado et al. 2003) and will likely influence the quality of the techno logy that the facility provides. Nursing resource : The total nursing staff levels of a nursing home is measured by adding all the registered nurses (RN) per resident, the licensed practical nurses (LPN) per resident and the certified nursing assi stant (CNA) per resident. Although th e duties for each of these staffing types are well defined and are in tended to be separate, there remains is a certain degree of overlap. As a result some facilities may chose to substitute higher skilled staff for lower skilled staff (E.g. CNA for RN or LPN for RN). Using lower skilled staff a llows facilities to keep costs down. However, nursing homes must ha ve an appropriate amount skilled staff in order to provide quality of care. Facilities who substitute staffi ng may lose their savings to the cost of poor outcome. For example, preventing a pressure ulcer from occurring throug h good care practices is going to cost less then trea ting a newly acquired ulcer. Rehab resource: The total rehab staffing level of a nursing home is measured by adding all the physical therapists (PT) physical therapy assistant (PTA ), occupational therapist (OT), certified occupational therapy assistant (COTA) pe r resident and speech pathologist per resident. The duties for each of these staffing types for each discipline are less well defined than nursing because therapy assistants can provide almost all the treatments that a therapist can provide. As a result facilities may chose to substitute mo re skilled staff for lesser skilled staff. 29

PAGE 30

Figure #1 illustrates the relationship between the technology provided by the facility and the resources needed to provide that tec hnology. A nursing homes pr imary function is to provide care to all the facilitys residents. The required inputs not onl y include th e facilities tangible and the intangible resour ces but also include the reside nts health status on admission. The technology includes all the nursing homes care pr ocesses that use the facilities tangible and intangible resources. Staffing le vels are the tangible resources that provide the physical care. The intangible resources include all the knowledge that the nurs ing home staff possesses and is measured by the amount of the facilitys st aff is licensed (Weech-Maldonado et al. 2003). Similar to inputs, the outputs also use the measure of the health status of the residents. However, the measure only occurs after the nursing home care processes have been applied. Figure 2-1. The nursing home process One factor that is unique for long term care is that because the resident is not likely to leave the facility, the desired output is the sust ained health and well being of the resident. The measure of quality is how well th e nursing home maintains (or improves) a residents health and well being during their stay in the nursing home. This is not equally true for the post acute 30

PAGE 31

resident since Medicare only pays for patients stay as long as the patients health status improves. Therefore the desired ou tput for the post acute patient is a better health status in hope that the resident can be discharged home. Scope and Resource Commitment Early writings postulated that business level strategy, the level relevant to strategic groups, consists minimally of two sets of activities: (1) business scope commitments and (2) resource commitments (Ansoff 1965; Katz 1970; Hofer and Schendel 1978; Day 1984; Cool and Schendel 1988). Scope commitments are those decisions that lead to the selection of market segments to compete and to the types of services to offer (Cool and Schendel 1988, Zinn et al. 1994). While applicable across diverse industries, the variab les chosen to operationalize these strategic dimensions should be industry specific, reflectin g the bases for competition in the industry under study (Cool and Schendel 1988). Resource commitmen t includes business level deployment of resources to those functional ar eas that are necessary to achieve and maintain competitive advantage (Cool and Schendel 1988). There have been some nursing home studies that have utilized scope and resource commitment as the parameters to identify whet her strategic groups existed in the nursing home industry. Zinn et al. 1994 used scope commitme nt measures of percent Medicare, percent Medicaid, percent independent capacity, case mix, average lengt h of stay and percent over 85. Zinn et al. 1994 used resource commitment measur es of size, private pay rate, and occupancy, RN per resident and staff per resi dent. Using cluster analysis, they found that facilities with the highest private pay residents had the best patient outcomes and the facilities with the highest percent Medicare beds had better financial performance. Marlin et al. 1999 used Medicaid utilization, Medicare u tilization, insurance/health maintena nce organization (H MO) utilization, 31

PAGE 32

private-pay utilization, and Vete rans Administration utilization, average length of stay, average patient age and case mix. For resource commitme nt, Marlin et al. 1999 used percentage of nursing costs, percentage of ancillary costs, occupancy rate, semiprivate room rate, number of beds, registered nurses (RNs) per resident, and st aff per resident. The study also used cluster analysis and found that the group with the highest private pay utilization combined with high Medicare utilization generally performed better along financial and quality indicators. Scope Using Technology Prior studies of strategic groups in nursing hom es have used the range of market segments as the measure of scope commitment (Zinn et al. 1994, Marlin et al. 1999 ). Zinn et al. 1994 and Marlin et al. 1999 argued that to compete effectively, nursing ho mes need to be responsive to differences in payer demand characteristics becau se the payer mix of a nursing home is the major basis of segmentation in the indus try (Zinn et al. 1994, Ma rlin et al. 1999). However, using payer demand characteristics may not be the best way of determining the strategic groups because the payer mix of a nursing home may be more the result of good strategy. For example, many facilities may have a strategy to acquire Medicare and private pay residents but not be successful. Whether a facility has a high pr oportion Medicare and private pay beds may be independent of the nursing homes attempt to acqui re these types of residents. Structure Conduct Performance theory is a m odel used to link elements of the market structure to business conduct and performance in i ndustrial economics. Struct ure refers to market structure defined mainly by the concentration of market share in the market. Conduct refers to the behavior of firms whether co mpetitive or collusive (pricing and production, production, goals of firms, promotion) (Britton et al. 1992). Perfor mance is mainly defined by the consequences of market power. Based on the Structure C onduct Performance paradigm, payer demand characteristics would be considered performance measures rather than conduct measures. In the 32

PAGE 33

nursing home industry, measures of this type of performance include percent of beds filled by Medicare patients and percent of beds filled by private pay residents. Both Medicare and private pay clientele pay higher reimburseme nt rates than other payers lik e Medicaid. Market shares of Medicare and/or private pay residents are the result of a stra tegy but do not give any information about what strategy was used by the nursi ng home to accomplish that performance. Unlike the aforementioned performance m easures, technology focus and resource deployment decisions are made by the business level management and are under the direct control of management. Decisions to increase or decrease the amount of nursing technologies provided (services like restorative care or skin ulcer prevention) can be made independent of the environment that the facility is located in. However the resulting quality of care caused by providing these services can influe nce the performance of the facility as measured by the ability of the facility to draw higher reimburs ing Medicare or private pay residents. In the nursing home industry, technology incl udes all the processes of care that are provided to the residents by the facilitys health care staff. Nursing homes use their technology to transform their inputs into output s. The nursing home industry is different from other industries because of the way technology is applied and the way the product is measured. Resource Commitment Resource commitment refers to the commitment of resources to functional areas that are needed to gain and maintain competitive advantage in targeted market segments. For nursing home administrators, resource commitment deci sions should be reflec ted by labor, price and capacity decisions (Zinn et al. 1994 ). Nursing homes have some c ontrol over their labor but they have limited control over their products price. Government reimbursement accounts for over 62% of their revenue (MedPAC 200 7). After the implementation of the Balanced Budget Act of 1997, Medicare dictated the price that they are going to pay for the nursing home stay. Similarly 33

PAGE 34

most states have changed their Medicaid reim bursement schematic in the direction of a prospective payment system. Only about 15% of nursing home beds are private pay where a facility can set their price. Capacity decisions can be impacted by economies of scale where the decision to be larger can allow the nursing home to benefit from the efficiencies created by more resources derived from size. Weech-Maldonado et al 2008 found that larger fa cilities have better financial performance. Industrial organization st udies have found that firms following similar strategies are likely to be of comparable scale (Porter 1980). In addition, labor and capacity can also be related. With widespread nursing and therapist s hortages, some facilities may not be able to increase the capacity due to labor limitations. Porters Generic Strategies Besides strategic group models, this study will examine the strategic-performance relationship using a conceptual sc heme that indicate the strategies that are available to firms. Porter (1980) introduced the concept that there are certain generic t ypes of strategies that can be used by firms to outperform other organizations in the industry. Po rters generic strategies are perhaps the most widely used and heavily researched. Firms have different types of environmentally determined functional demands and t hus choose different strategies in an effort to gain competitive advantage (Porter 1980). Originally Porter (1980) argued that a firm could choose one of three strategies to compete at the business level. One of those strategies is differe ntiation where a nursing home differentiates itself from other nur sing homes in the market. One way facilities can differentiate themselves from their competitors is providi ng a high quality product. In the nursing home industry this equates to better care. By providing a high quality product, businesses are able to shield themselves from reductions in price-co st margins (Porter, 1980). Weech-Maldonado et al. 34

PAGE 35

(2003) study found that there was a positive rela tionship between quality of care and financial performance in nursing homes. A nursing home can also use cost leadership in which the nursing home attempts to have relative lower costs. One would expect that nursing homes that follow the cost leadership paradigm would be more effici ent at using their staff. The third strategy is a focus strategy, in which the firm concentrates on a particular group of customers, geographic markets, or product lin e segments. In the case of the nursing home industry, some facilities may choose to focus on re habilitation because of the higher revenues per bed that Medicare pays. Porter (1980) considered firms that attempte d to apply both a differe ntiating strategy and a cost leader strategy at the same time as stuck in the middle or muddlers. Stuck in the middle or muddlers are considered those organiza tion who have no coherent strategy. These organizations did not satisfy the demands of either cost leaders or differe ntiators (Marlin et al. 2002). In contrast, Hill et al. 1988 and Wright et al. 1987 have pos ited that some nursing homes can be successful at differentiating themselves from their competitors while controlling their costs at the same time. Marlin et al. 2002 coined this type of nursing home as best costs and found that the best cost strategy existed in the hospita l industry. Porter has acknowledged that on rare occasions, firms can be successful with bot h differentiating strategy and a cost leader strategy (Murray et al.1988). Marlin et al. (2002) used strategic groups as defined by Porters theory of competitive strategy in the hospital industry. Ma rlin et al. 2002 used an object ive classification procedure to classify four strategic groups : differentiation, low cost, best costs and the muddlers. The variables for the cluster analysis were selected by a panel of experts. Thos e variables selected to 35

PAGE 36

capture cost leadership were measures that capture different types of cost like salaries and labor hours. The variables used to captu re the differentiation group were t ypes of services. Marlin et al. (2002) found that hospital strategi c groups could be defined using Porters generic strategies theory and that different strategic groups had different structure-performance relationships. Strategic Group Model Using Scope (Technology) Commitment, Resources Commitment and Porters Generic Strategy Theory Figure #2 is a strategic group model of the nursing homes industry. The horizontal axis represents the scope commitment of the nursing home as measured by the amount of technology provided by the facility. Figure 2-2. Strategic Group Model Based on Scope and Resource commitment The vertical axis represents the resources that the facilities commit to provide its technology. The circles represent clusters as defined by Porters stra tegies (Differentiator, Cost Leadership, Best Cost and Focus). Those faciliti es that do not have a strategy are defined as being groups that lack strategy 36

PAGE 37

The center of the circle depicts the centroid of the cluster. To be in the cluster, a nursing home is closer to the centroid of its cluster than the centroid of any other cluster in the industry. All nursing homes within the groups are located within the outer boundaries of each cluster. A nursing home is not able to be in two separate clusters. The upper scope boundary is a vertical line th at passes through the cen troid of the cluster that has the highest scope (technology) value. The lower resour ce boundary is a horizontal line that passes through the centroid th at has the lowest resources. Ther e are nursing homes that fall beyond both these boundaries; however the centroids of the clusters ca nnot fall beyond these boundaries. Differentiator For a nursing home to differentiate itself from its competitors, it must provide superior processes of care. By taking on the strategy of providing more service than the competitors in their local markets, these nursing homes are more likely to attract residents to their facility resulting in higher occupancy ra tes and choose residents that can provide them with higher margins. A differentiating nursing home would pr ovide a high volume of service relative to the average nursing home in the market. In Figure #2, the differentiating strategic group would be expected to be in the upper right quadrant of the graph. This is because the differentiating group is defined as the highest provider of servic e (technology). If a high volume of service is provided, it would be expected that facilities in the differentia tion strategic groups would also have to commit a greater amount of resources. The centroid of the cluster demarks the upper scope boundary. Cost Leadership Some nursing homes will be focused on keepi ng their costs low. Since staffing is the significant cost driver in the nursing home industry, k eeping lower staffing levels is one obvious 37

PAGE 38

method that facilities can use to control cost. RN s cost significantly more than CNAs and LPNs and therefore by manipulating staffing levels to a lower skill mix, cost le adership nursing homes are expected to keep their costs low. In Figure #2, the cost leadership group would be expected to be loca ted in the bottom left quadrant of the graph. This is because the cost le adership group has the lowest use of resources. It is expected that because of the lower availabi lity of resources, cost le adership strategic groups will provide less processes of care. The centroid of the cluster demarks the lower resource boundary. Best Cost Some nursing homes may seem like a hybrid of the cost leadership group and the differentiation group relative to the lack of strategy group. The best cost group represents efficient organizations. In Figure #2, the best cost group would be expected to have its centroid between the cost leaders and th e differentiators. More specifi cally, the best cost group is expected to provide fewer services than the differentiating group but use fewer resources. It would also be expected to provide more servic es than the cost leadership group using more resources. Focus Like any industry, it is possible for a nursing home to focus their strategy on a smaller segment of the industry. By allocating resources to that area, it may be possible to perform well financially. In Figure #2, the centr oid of the focus group could be located anywhere in the graph within the upper scope boundary and the lowest resource boundary. However it is likely that the focus group would provide more services than the cost leadership gr oup because it would be expected that it is focused on pr oviding a specific type of service like rehab. There is more than one type of technology that a nursing home can focus on providing. It is therefore possible that 38

PAGE 39

within the nursing home industry, there exists more than one focus group. For example, a nursing home could focus their atten tion on providing nursing care while a different one could focus their attention on providing rehabilitati on services. Lack of Strategy Facilities that use high levels of resources but provide low levels of technology fall into the lack of strategy group. This group will also not have any focus of services and special allocation of resources. Measuring Performance Financial performance metrics are the most commonly used when comparing different strategic groups. Some of the more common measures of performance include total margin and operating margin. Total margin is the total revenu e divided by the total expense. This is an important measure because any organization that is unable to sustain positive margins will cease to exist. The total margin not only includes th e operating costs and revenue, but other costs (capital costs and interest) and other revenue (like endowments and donations). Operating margin only includes operating revenue and opera ting cost. Operating margins are important because they measure efficiency. The higher the operating margin, the more profitable a company's core business. Operating margin is re lated to the day-to-day decisions that managers make such as pricing strategy, pri ces for raw materials, and / or la bor costs. Operating margin is also a measure of managerial flexibility a nd competency of the firm (Gapenski 2000). In the health care industry (inc luding nursing homes), patient cost s is also widely used as a financial metric (Weech-Maldonado et al. 200 3, Shen et al. 2007). Since public payers like Medicare and Medicaid have been placing significant pressure on a f acilities ability to generate revenue, it has been necessary for facilities to co ntrol their costs in order to achieve positive margins. 39

PAGE 40

In most industries, success is most commonly defined by financial metrics; however, Venkatraman et al. 1986 have suggested that othe r measures such as market share, new product introduction and product quality can be equally important. One measurement that has recently gained popularity in business is the balanced scorecard (BSC) approach. A major characterist ic of a BSC is that it combines long-range strategic financial goals with day-to day operations. Kaplan a nd Norton (2001) stated that the customers perspective and internal operations are important areas that should be considered along with the financ ial perspective. Taking the customers perspective, the ability of a nursing home to provide quality care is crucial if a facility is going to survive over the long term. Poor quality can result in higher costs (Weech-Maldonado et al. 2003) as well as lower reve nue streams due to lower occupancy rates. Poor quality can also result in moratorium on new admissions and even a facilitys closure. Quality in nursing homes has also generated c oncern for policymakers and the public ever since the release of the Institute of Me dicine (IOM) report in 1986. It is therefore very important that quality indicators be used in conjunction with financial variables when measuring performance. There are three key reasons why measures othe r than financial perfor mance are needed in the nursing home industry. First, the government programs like Medicare and Medicaid were established to pay for the care of the program s beneficiaries. Alt hough enough funds must be allocated to keep companies viable, the purpose of the program is not to give organizations above normal profits. Second, not a ll health care organizations are fo r profit. For example, nearly one third of nursing homes in the US are either go vernment or not for profit facilities (Harrington et al. 2001) and it is important to consider wh at health care organiza tions do. The presence of not-for-profit in the industry is another reason why performan ce objectives other than profit 40

PAGE 41

maximization are needed (Scanlon, 1980). Fina lly, health care organi zations (like nursing homes) are often charged with th e responsibility of providing care for one of the most vulnerable populations. The direness of poor quality health car e is so great that usin g financial performance as the sole measure of success (like in othe r industries) is just not adequate. For the aforementioned reasons, the measurement success in the health care industry has to be based on quality of services provided, cost of th at service and financial performance. New Contributions To date, no study has examined if strategic groups can be identifie d by the technology choices made by the nursing home management. The purpose of this study is to determine if strategic groups can be classified using Port ers generic strategies based on how nursing home management allocate their resources and how they select their technology. The study will also determine if these groups are succinct and if they differ with respect to quality, financial performance and cost. What and how much of technology a nursing home administrator chooses to provide is an important stra tegy choice that can ha ve significant repercussions. The study is different from prior studies because it uses the industrys technology as the scope variables to determine the strategic groups and will define th ose groups using Porters generic strategies. Hypothesis Hypothesis #1: Nursing homes th at have a strategy can be categorized into one of four strategic groups: Differentiator, cost leadership, focus or best cost. Hypothesis #2: Nursing homes in the diffe rentiator strategic gr oup will provide higher quality care than othe r strategic groups. Hypothesis #3: Nursing homes in focus stra tegic group will provide higher quality in the area that they are focused on than other strategic groups. 41

PAGE 42

42 Hypothesis #4: Nursing homes in the cost leadership strategic group will have lower costs than other strategic groups. Hypothesis #5: Nursing homes that lack a strategy will have lower financial performance than nursing homes that are in a strategic group.

PAGE 43

43 CHAPTER 3 METHODOLOGY Data This study combined the On-line Survey Ce rtification of Automated Records (OSCAR), the Minimum Data Set (MDS), the Area Resource File (ARF) and the Medicare Cost Reports. The OSCAR dataset is a facility level dataset th at is routinely collect ed through the Medicare and Medicaid certification process conducted by st ate licensure and certifi cation agencies. This data is updated annually as part of the recertif ication survey. OSCAR pr ovides a snapshot about the characteristics and care processes occurri ng within the nursing hom es at the time of the survey. It includes variables like staffing levels, ownership status, number of residents, facility acuity level as well as quality outcome measures lik e state survey care deficiencies and pressure sore prevalence. The Omnibus Budget Reconciliation Act of 1987 (OBRA) was the first major legislative step to improve the quality of care in nursing homes. It established the Resident Assessment Instrument (RAI) which was implemented in 1991. The RAI is the basis of the Minimum Data Set (MDS) which is a dataset that records info rmation about nursing home residents. The MDS includes demographic information, health status as well as amount and types of services that a resident receives while they reside at the nursi ng home. Each long term care resident is assessed when first admitted to a nursing home, then each quarter thereafter. The post acute resident is assessed on the 5th day, the 14th day, the 30th day, 60th day and 90th day (for as long they remain in the facility). The CMS MDS is the most comprehensive dataset available for examining quality outcomes in the long-term care setting. OSCAR and Quality Indicators (QI) generated from the MDS are used by the CMS as part of its Nursing Home Compare website. The MDS was acquired using a reuse agreement from the CMS. Data is currently located behind a firewall in a fully secured server of the College of Public Health and Health Professions

PAGE 44

44 building operated by the IT departme nt at the University of Florida. It was originally acquired by University of South Florida. The Area Resource File (ARF) contains count y level data on supply and demand factors, such as composition of health care workfo rce, and socioeconomic and demographic characteristics of the population. The Medicare Cost Report data is a public a ccess relational dataset that can be acquired from the CMS via the CMS website. All CMS certified nursing homes that have Medicare beds are required to submit the Medicare cost report on an annual basi s. MedPAC has repeatedly been using this dataset to report the financial information of nursing homes to congress over the past 7 years (MedPAC 2007). This dataset provides financ ial data like total patient revenue, total patient costs, and total revenue. Population This study included all nursing homes in the United States that had Center for Medicaid and Medicare certification and that were included in the OSCAR, MDS and Medicare cost report datasets in 2004. There were 16000+ nursing homes in the United States in 2004. However, a total of 4500 facilities were omitte d because they were either hospital based (2000 facilities) or government based (700 facilities), did not have Medicare beds (1200 facilities) or were the only nursing home in a county and would not have any market competition (500 facilities). Hospital based facilities were omitted because they may behave differently from free standing facilities (Weech-Maldonado et al. 2003, Weech-Maldonado et al. 2004). For example, hospitals may place patients in their own subacute beds that th ey cannot place in other nursing homes due to the level of acuity of the patient or because the patient lacks insurance. Government facilities may behave differently than non-government fac ilities because of their different financing structure. They are less influenced by market forces because of funding that they may receive directly from public sources. Finally, some fac ilities that may be included in the OSCAR may

PAGE 45

45 not have any Medicare residents or even Medicare beds availabl e and therefore they would not have to file a Medicare cost re port. Therefore, this study is li mited to nursing homes that are non government, free standing and have Medicare certifie d beds. The total number of facilities in the study is approximately 11601 facilities in the year 2004. Overview of Methodology To identify the strategic groups and the strategy-performance relationships of the strategic groups, the study perf ormed a factor analysis on the scope and resource commitment variables. The resulting variables we re used in a cluster analysis to identify the different clusters. Once clusters were identified, ordinary least s quares regression (OLS), negative binomial and ordered logit was used to determine if the diffe rent groups had different financial and quality strategic-performance relationships. Hypothesis #1: Nursing homes that have a st rategy can be categorized into one of four strategic groups: Differentiator, cost leadership, focus or best cost. Scope Similar to Zinn et al. (1994) and Marlin et al. (1999), this study used the scope commitment and resource commitment (also kno wn as resource deployment) dimensions to identify the strategic groups. The scope of this study included technologies used by nursing homes to take care of the residents. The tec hnology of a nursing home can be classified into two main categories: nursing t echnology and rehab technology. Nursing Technology Nursing technology includes many different types of processes. The overall care of the residents can be divided into several different component technologies that are specifically designed to address the varied needs of the nursing home resident. Each of these technologies has a specific purpose and if not applied properly will have an undesired outcome. Some of the essential technologies include skin car e technology, restorative technol ogy, incontinence reducing technologies and nutritional technologies.

PAGE 46

46 Skin care technology is designed to prevent pr essure ulcers from occurring and to treat pressure ulcers that already ex ist. Failure to apply preventativ e skin care technology properly can result in pressure ulcers. Failure to treat alrea dy existing pressure ulcers can result in pressure ulcer worsening. Restorative nurs ing technologies engage the resi dent in functional activities. Residents functional abilities can decline in the nursing home se tting and restorative technology can be used to help maintain their functional levels Incontinence reducing technologies assist residents to maintain or regain their continence through toileting programs or bladder retraining programs. Nutritional technologies help residen ce maintain a healthy we ight. In the nursing home setting, residents are at risk of either ga ining weight or losing weight depending on their medical condition. Weight management programs and therapeutic diets assist with proper nutrition while the use of plate gua rds and other devices ensures th at residents adequately feed themselves. Skin care technology : All processes of care that involve either preventative skin care or treatment skin care are cons idered skin care technology. Pressure relieving seat : This is a continuous facility le vel variable that represents the proportion of residents in the nursi ng home that are given a pre ssure relieving seat. Pressure relieving seats are an important intervention when trying to prev ent a new pressure sore from occurring or to prevent an alre ady acquired pressure sore from worsening. Includes gel, air (e.g., Roho), or other cushioning placed on a chair or wheelchair. Include pressure relieving, pressure reducing, and pressure re distributing devices. Pressure relieving bed : This is a continuous facility le vel variable that represents the proportion of residents in the nur sing home that are given a pr essure relieving bed. Pressure relieving beds are an important intervention to prevent new pressure sores from occurring or to prevent an already acquired pressure sore from worsening. Includes air fluidized, low air loss

PAGE 47

47 therapy beds, flotation, water, or bubble mattre ss or pad placed on the bed. Include pressure relieving, pressure reducing, and pr essure redistributing devices. Turning / Repositioning : This is a continuous facility le vel variable that represents the proportion of residents in the nursi ng home that are placed on a turning-repositioning schedule. Turning-repositioning schedules are an important intervention to prevent a new pressure sore from occurring or to prevent an already acquire d pressure sore from worsening. It includes a continuous, consistent program for changing th e residents position and realigning the body. Program is defined as a specific approach th at is organized, planned, documented, monitored, and evaluated. Ulcer care: This is a continuous facility level va riable that represents the proportion of residents in the nursing home that are receiving ulcer care trea tment. Residents who get ulcer care already have ulcers. Surgical care: This is a continuous facility level vari able that represents the proportion of residents in the nursing home th at are receiving surgical care treatment. Residents who get surgical care already have ulcers. Ointment: This is a continuous facility level va riable that represents the proportion of residents in the nursing home th at are receiving ointment car e. It includes ointments or medications used to treat a skin conditi on (e.g., cortisone, anti fungal preparations, chemotherapeutic agents, etc.). Dressing care: This is a continuous facility level vari able that represents the proportion of residents in the nursing home th at are receiving dressing care. Residents who get dressing care already have ulcers. Skin nutrition program care: This is a continuous facility le vel variable that represents the proportion of residents in the nursing home that are on a skin nutrition program. Dietary measures received by the resident for the purpo se of preventing or treating specific skin

PAGE 48

48 conditions e.g., wheat-free diet to prevent alle rgic dermatitis, high calorie diet with added supplements to prevent skin breakdown, high protein supplements for wound healing. Vitamins and minerals, such as Vitamin C and Zinc, whic h are used to mange a potential or active skin problem, should be coded here. Other skin care : This is a continuous fac ility level variable that represents the proportion of residents in the nursing home that are receivi ng other skin care. Residents who get other skin care already have ulcers. Restorative technology : All processes of care provided by the nursing staff that involves having the patient engage in an activity that is directed towards accomplishing an ADL goal that will result in improved health status. Restorative ambulation : The restorative ambulation variable is facility level and is continuous. It represents the av erage number of days that a resident receiv es restorative ambulation while remaining in the facility. It is generated by dividi ng the resident level restorative variable that represents the number of days of ambulation provided in the 7 days before the assessment date by the total number of residents in the facility. Restorative ambulation has the nursing staff walk the residents in order to improve the distance that the resident can ambulate. It is a routine service and does not requ ire the skill of the therapist. However the goals of the program need to be generated by the therapist. Restorative Passive Range of Motion (PROM) : The restorative PROM variable is facility level and is continuous. It represents the averag e number of days that a resident receives restorative PROM while remaining in the facilit y. It is generated by divi ding the resident level restorative variable that represents the number of days of PROM provided in the 7 days before the assessment date by the total number of reside nts in the facility. Restorative PROM has the nursing staff range the residents in order to improve the PROM of the resident. It is a routine

PAGE 49

49 service and does not require the skill of the therap ist. However the goals of the program need to be generated by the therapist. Restorative Active R ange of Motion (AROM) : The restorative AROM variable is facility level and is continuous. It represents the averag e number of days that a resident receives restorative AROM while remaining in the facilit y. It is generated by dividing the resident level restorative variable that represents the number of days of AROM provided in the 7 days before the assessment date by the total number of reside nts in the facility. Restorative AROM has the nursing staff range the residents in order to improve the AROM of the resident. It is a routine service and does not require the skill of the therap ist. However the goals of the program need to be generated by the therapist. Restorative transfers : The restorative transfer variable is facility level and is continuous. It represents the average number of days that a resident receives rest orative transfers while remaining in the facility. It is generated by dividing the resident level restorative variable that represents the number of days of transfers provided in the 7 days before the assessment date by the total number of residents in the facility. Re storative transfers have the nursing staff transfer the residents in order to improve the residents ab ility to transfer from one position to another. It is a routine service and does not require the sk ill of the therapist. However the goals of the program need to be gene rated by the therapist. Restorative bed mobility : The restorative bed mobility vari able is facility level and is continuous. It represents the aver age number of days that a resident receives restorative bed mobility while remaining in the facility. It is ge nerated by dividing the resident level restorative variable that represents the number of days of bed mobility provided in the 7 days before the assessment date by the total number of residents in the facility. Restorative bed mobility has the nursing staff transfer the residents in order to improve the reside nts ability to tr ansfer from one

PAGE 50

50 position to another in bed. It is a routine servi ce and does not require the skill of the therapist. However the goals of the program need to be generated by the therapist. Restorative dressing : The restorative dressing variable is facility level and is continuous. It represents the average number of days that a resident receives restorative dressing while remaining in the facility. It is generated by divi ding the resident level restorative variable that represents the number of days of dressing provided in the 7 days before the assessment date by the total number of residents in the facility. Re storative dressing have th e nursing staff help dress the residents in order to improve the residents ability to dress by themselves. It is a routine service and does not require the skill of the therap ist. However the goals of the program need to be generated by the therapist. Restorative eating : The restorative eating variable is facility level and is continuous. It represents the average number of days that a re sident receives restorativ e eating while remaining in the facility. It is generated by dividing the resident level restora tive variable that represents the number of days of eating provide d in the 7 days before the assessment date by the total number of residents in the facility. Rest orative eating have the nursing sta ff help the residents with eating in order to improve the residents ability to eat by themselves. It is a routine service and does not require the skill of the therapist. However the goals of the program need to be generated by the therapist. Restorative splinting : The restorative splinting variable is facility level and is continuous. It represents the average number of days that a resident receives restorative splinting while remaining in the facility. It is generated by dividing the resident level restorative variable that represents the number of days of splinting provided in the 7 days before the assessment date by the total number of residents in the facility. Re storative splinting have the nursing staff assist with the donning and doffing of splints in order to improve the residents ability to splint

PAGE 51

51 themselves. It is a routine service and does not require the skill of the therapist. However the goals of the program need to be generated by the therapist. Restorative amputee intervention : The restorative amputee intervention variable is facility level and is continuous. It represents the averag e number of days that a resident receives restorative amputee intervention while remaining in the facility. It is generated by dividing the resident level restorative variab le that represents the number of days of amputee intervention provided in the 7 days before the assessment date by the total number of residents in the facility. Restorative amputee intervention have the nursing staff assist with th e donning and doffing of prosthetic in order to improve the residents abil ity to don and doff their prosthetic by themselves. It is a routine service and does not require the skill of the therapist. However the goals of the program need to be generated by the therapist. Restorative communication : The restorative communication vari able is facility level and is continuous. It represents the av erage number of days that a resident receiv es restorative communication while remaining in the facility. It is generated by dividing the resident level restorative variable that represents the number of days of communication provided in the 7 days before the assessment date by the total number of residents in the facility. Restorative communication has the nursing staff assist with the communication in order to improve the residents ability to communicate. It is a routin e service and does not re quire the skill of the therapist. However the goals of the program need to be generated by the therapist. Restorative other treatment : The restorative other treatment variable is facility level and is continuous. It represents the aver age number of days that a resident receives other restorative while remaining in the facility. It is generated by dividing the resident level restorative variable that represents the number of days of other provid ed in the 7 days before the assessment date by the total number of residents in the facility.

PAGE 52

52 Incontinence reducing technologies : All processes of care pr ovided by the nursing staff intended for the patient engage in a program that is directed towards accomplishing continence. Toileting program : This is a continuous facility level variable that represents the proportion of residents in the nursing home that have a to ilet program. Toileting programs include such items like toileting schedules. Bladder retraining : This is a continuous facility level variable that represents the proportion of residents in the nur sing home that have a bladde r retraining program. Bladder training programs include such items like toileting schedules. Nutrition technology: All processes of care provided by the nursing staff intended for the patient engage in a program that is dir ected towards maintaining a healthy weight. Weight management program : This is a continuous facility level variable that represents the proportion of residents in the nursing home that have a weight management program. Resident is receiving a program of which the documented purpose and goal are to facilitate weight gain or loss (e.g., double portions; high calorie supplement s; reduced calories; 10 grams fat). Therapeutic diet: This is a continuous facility level variable that measures the proportion of residents that are on a thera py diet. A diet ordered to manage problematic health conditions. Examples include calorie-specific, low-salt, lo w-fat lactose, no added sugar, and supplements during meals. Plate guard : This is a continuous facility level va riable that represents the proportion of residents in the nursing home havi ng a plateguard. This includes a ny type of specialized, altered, or adaptive equipment to facilitate the reside nts involvement in self performance of eating. Rehab Technology Rehab technology is utilized to improve the residents health stat us. It is provided by licensed therapy staff. There are di fferent types of therapist that have different functions. These

PAGE 53

53 processes of care include physical therapy for mobility issues and occupational therapy for ADL issues other than mobility. Long term care rehabilitation: All processes of care provided by the therapy staff that involves rehabilitation services to the long term care residents. Mean PT minutes long te rm care per resident : Facility level variable that represents the average number of minutes of phys ical therapy per resi dent provided in the 7 days prior to the quarterly assessment date. The numer ator is the number of minutes of PT and the denominator is the number of residents in the nurs ing home. This is a MDS resident level variable that is raised to the facility level. Mean OT minutes long te rm care per resident : Facility level variable that represents the average number of minutes of o ccupational therapy per resident pr ovided in the 7 days prior to the quarterly assessment date. The numerator is the number of minut es of OT and the denominator is the number of residents in th e nursing home. This is a MDS resident level variable that is raised to the facility level. Post acute rehabilitation: All processes of care provided by th e therapy staff that involves rehabilitation services to the long term care residents. Mean PT minutes post acute per resident: Facility level variable that represents the average number of minutes of Physical Therapy th e 7 days prior to the 14 day assessment date. The numerator is the number of minutes of PT and the denominato r is the number of residents in the nursing home. This is a MDS resi dent level variable that is raised to the facility level. Mean OT minutes post acute per resident : Facility level variable that represents the average number of minutes of Occupational Therapy provided in the 7 days prior to the 14 day assessment date. The numerator is the number of minutes of PT and th e denominator is the number of residents in the nursing home. This is a MDS resident level variable that is raised to the facility level.

PAGE 54

54 Percent Medicare beds : Facility level variable that represents the percent of the total number of resident beds reimbursed by Medicare. Medicare only reimburses for resident who are at the facility for short term stay. This variable is used because it is necessary to account for the proportion of the patients that ar e receiving rehabilitation at th e facility. It is generated by dividing number of Medicare residents by the total number of beds for that facility. Resource Commitment Resource commitment refers to the commitment of resources to the different technology areas that are needed to gain and maintain competitive advantag e in targeted market segments (Zinn et al. 1994). The primary resour ce is staffing levels (the nurses and the therapist) since it is the health care staff that pr ovides the technology. Therefore, organizations manipulate their staffing levels to control costs and ensure qual ity in an attempt to perform well financially. This study divided the resource commitment into two separate subtypes because they are mutually exclusive of each other and do not overlap. Resource commitment nursing: Resource commitment nursing subtype includes all the licensed and non licensed nursing staff. They are responsible for all the nursing technologies RN FTE / Total Residents : Registered Nurse (RN) per resi dent is a continuous variable representing the ratio of the tota l number of RN FTEs divided by the total number of residents in the facility. RNs are involved in establishing processes and supervisi ng other nursing staff. LPN FTE / Total Residents : Licensed practical nurse (LPN) per resident is a continuous variable representing the ratio of the total numbe r of LPN FTEs divided by the total number of residents in the facility. Both measures are from the OSCAR dataset. LPNs are able to provide some skilled nursing processes of care like dispensing medication. They have to be under the supervision of a nurse. CNA FTE / Total Residents : Certified Nursing Assistant (CNA) per resident is a continuous variable representing the ratio of the total number of CNA FTEs divided by the total

PAGE 55

55 number of residents in the f acility. Both measures are from the OSCAR dataset. CNAs are mostly involved with the direct care of patients. Percent Licensed : Percent nursing staff that are li censed is a continuous staffing mix variables that measures the proportion of nur sing staff that have a license. Higher percent licensed will give the nursing home a higher ca pacity to have know ledge and intangible resources. Percent of nursing staff that are RNs : Percent of nursing staff that are RNs is a continuous staffing mix variable that measur es the proportion of the nursing staff that are RNs. Higher RN staffing mixes are advantageous since RNs can us e their professional skills and tacit knowledge to improve the quality of patient care through their involvement in establishing processes and supervising other nursing sta ff (Weech-Maldonado et al. 2003). Resource commitment rehab : Resource commitment rehab includes all therapists and therapy assistants. Only licensed rehab staff is included because the use of rehab aides for therapy treatments is minimal due to state and federal practice guidelines. PT FTE / Total Residents : Physical Therapist (PT) per resident is a continuous variable represents the ratio total number of PT FTEs divided by the total number of residents in the facility. PTs evaluate the patient and oversee th e plan of care. They supervise the physical therapy assistants PTA FTE / Total Residents : Physical Therapy Assistants (P TA) per resident is a continuous variable represents the ratio total number of PT A FTEs divided by the to tal number of residents in the facility. PTAs provide the treatment to the patients under the supervision of the physical therapist. Percent PT: Percent of the physical th erapy staff that are PTs is a continuous staffing mix variable. Higher PT staffing mixes are advantageous since PTs can use their professional skills

PAGE 56

56 and tacit knowledge to improve the quality of patient care through their involvement in establishing processes and supervis ing other physical therapy staff. OT FTE / Total Residents : Occupational Therapist (OT) per resident is a continuous variable represents the ratio total number of OT FTEs divided by the total number of residents in the facility. OT evaluates the pa tient and oversees the plan of car e. They supervise the Certified Occupational Therapy Assistant COTA FTE / Total Residents : Certified Occupational Ther apy Assistant (COTA) per resident is a continuous variable represents the ratio total numbe r of COTA FTEs divided by the total number of residents in the facility. COTAs provide the treatme nt to the patients under the supervision of the occupational therapist. Percent OT: Percent of the occupational therapy staff that are OTs is a continuous staffing mix variable. Higher OT staffing mixes are adva ntageous since OTs can use their professional skills and tacit knowledge to improve the quality of patient care through their involvement in establishing processes an d supervising other occu pational therapy staff. Wages: Nursing homes average wages is a cont inuous variable derived by dividing the total salary expense by the total number of working hours reported in the Medicare costs reports. Wages are considered resources measures because nursing homes utilize the wages to control their cost. Hypothesis #1 Analysis Since there was a total of 68 va riables of interest (technology and resource variables), it was necessary to reduce the num ber of variables into composite variables by using factor analysis. The internal consistency of the newly created compos ite variables was tested using Cronbachs (1951) alpha methodology. Some variables that were not included in the composite scores were still included in the clus ter analysis as individual variables.

PAGE 57

57 The composite (or index) and the appropriate individual variables were transformed into cluster variables using z-scores. Wards mini mum variance method was used to identify the number of clusters and k-means cluster analysis to identify the different strategic groups. The analysis of variance and covariance (ANOVA) statistical test with Tukeys comparison test was used for each variable of interest to determine if the strategic groups were significantly different from each other. The Multivariate anal ysis of variance and covariance (MANOVA ) test with Wilks Lambda was used to determine if the ove rall clusters were signi ficantly different from each other. Each variable was given a weighted rank score for each cluster. The weighted rank score was then aggregated by type of commitme nt (scope or resource). Using the aggregated weighted rank score, the clusters were classified based on Porters generic strategies as defined in the technology, resources and strate gic group model presented in Figure #2. Factor Analysis Factor analysis is a variable reducing technique. It is a fa mily of procedures designed to remove the redundancy from a set of highly correla ted variables and representing the variables in a smaller set of derived variables or factors (Kachigan 1991). Factor analys is can be seen as the grouping of similar variables (Kachigan 1991). Thes e groupings of variables (or factors) can be used to describe a particular construct. For example, in the nursing home industry, two of the treatments used in an attempt to decrease pre ssure sores include placing a pressure relieving device in a chair and placing a pressure relieving device in a bed. It would be expected that these two processes of care are going to be highly correlat ed since both processes are similar in nature. The study used exploratory factor analysis to identify scope commitment and resource commitment variables that are correlated to each other. These variables were combined into new index variables that were a measure of what was considered a technology or resource factor. In the above example, if the factor analysis dete rmined that the two pressure relieving device

PAGE 58

58 variables are correlated and need to be combine d, the new variable may be considered a measure of the construct Pressure Sore Preventative Interventions. The study used the interaction between the scope and the reso urce dimensions to classify the strategic groups. It was theref ore necessary to perform separate exploratory factor analysis for scope commitment and for resource commitment. Similarly, the staffing intensity and staffing mix variables were also analyzed separately. Th is is because it is necessary to separate the intensity and the skill mix of the staff to be able to identify the relationship between licensed and non licensed staff. Staffing mix vari ables cannot be used as a substi tute for the staffing intensity (i.e. nurse per resident) variables because th ey do not explain the volume of nursing care provided. For example, in the case of nursing, C NAs can be negatively correlated with the RNs in one cluster while being positively correlated in another cluster. As a result, the three separate factor analysis include technology variables, sta ffing intensity variables and skill mix variables. Variables included in the cluste r analysis but t not in the f actor analysis were wages and percent Medicare beds. Although av erage wage paid is a resource commitment measure and is related to the staffing mix variab les (higher RN mix represents higher average wages), it was not included in any exploratory factor analysis. Wages are used by nursing homes as an incentive for hiring new staff and retaining thei r staff and therefore were incl uded separately in the cluster analysis. Percent Medicare beds va riable is related to scope because it indicates the proportion of the facility residents getting re hab. It was included in the clus ter analysis independently to identify the total volume of therapy services prov ided. For example, a facility that provided a lot of minutes of therapy to a small number of re sidents could be misclassi fied as being a rehab facility if the total amount of residents receiving therapy were not accounted for. Factor Matrix In the factor matrix, the columns represen t the derived factors while the rows are the variables. The cells represent the factor loadings which represents the de gree that the variables

PAGE 59

59 relate to the derived factor. Fact or loadings are used to reveal the extent that the variables contribute to the meaning of the variable. Factor loading are the correlation coefficients between the variables and the factors. A factor loadi ng score of greater than 0.4 was the inclusion criterion to be considered a part of that factor (Kachigan 1991). Number-of-Factors Problem One concern when using factor analysis is how many factors are reta ined. Each factor accounts for a certain amount of variance in the data. Typically, the first f actor will extract the most variance; the second factor will extract the second most and so on. Once the variance of each successive factor extracts is known, it is possible to de termine how many factors can be retained. There are two primary methods that can be used to determine the amount of factors retained: 1) Eigenvalue is a measure of the variation and represents the equiva lent number of variables that the factor represents. The Kaiser criterion is a commonl y used standard that includes only factors with eigenvalues greater than 1 (Kachigan 1991). The rationale for using such a threshold is that unless a factor extracts at least as much variation as the equivalent of one original variable, it should be dropped. This criterion was pro posed by Kaiser (1960), and is probably the one most widely used standard in determining how many factors should be used. 2) Scree plot is another method that can be used to id entify the number of factors. It is also possible to identify how many factors to retain by plotting the incremental variance accounted for by each successive factor. The idea is that the ta il of the curve represents mostly random error variance and therefore the factor solution is to in clude all factors that exist just prior to the leveling off of the curve (Kachigan 1991). Rotation of the Factor Structure We could plot the factor loadings shown above in a scatter plot. In that plot, each variable is represented as a point. In this plot we could rotate the axes in any direction without changing

PAGE 60

60 the relative locations of the poi nts to each other; however, the act ual coordinates of the points, that is, the factor loadings would of course change (Kachigan 1991). Rotational strategies: There are various rota tional strategies that have been proposed. The goal of all of these strategies is to obtain a clear pattern of loadings, th at is, factors that are somehow clearly marked by high loadings for so me variables and low loadings for others. Fabrigar et al. 1999 describe two commonly used types of ro tational strategies: orthogonal and oblique. Although orthogonal is conceptually simpler, factors have to be uncorrelated. Oblique rotation on the other hand do es not have this limitation. Si nce the resource factors have the potential of being correlated to the technology that they are used for (for example, therapy resources will be correlated with rehab technol ogy), this study used the oblique rotation. Internal Consistency of Measures It was necessary to determine whether the vari ables of the generated factors have internal consistency. Cronbach alpha tested the correlation between the variables of the construct. The higher the alpha value, the closer the variables are correlated (C ronbach 1951). This value should be above 0.7 (Allen 2002). If th e alpha was lower than 0.7, the factor was dropped because the variables did not correlate well together. To dete rmine how much impact the individual variables had on the overall alpha, multiple tests will provide a score of the resulting alpha. Any variable whose exclusion causes the alpha score to rise above 0.7 were be dropped. Similarly, any variable whose inclusion causes the alpha score to drop below 0.7 were dropped. Cluster Analysis Cluster analysis is an explorat ory data analysis tool which ai ms at sorting different objects into groups in a way that the degree of association between two objects is maximal if they belong to the same group and minimal otherwise (Kach igan 1991). Computationa lly, cluster analysis could be thought of as analysis of variance (ANOVA) "in reverse." The program starts with k random clusters, and then move objects between those clusters with the goal to 1) minimize

PAGE 61

61 variability within clusters and 2) maximize vari ability between clusters In other words, the similarity rules apply maximally to the members of one cluster and minimally to members belonging to the rest of the cluste rs. Cluster analysis can also be described as a set of techniques for partitioning a set of objects into relative ly homogenous subsets based on inner objects similarities. The process begins by measuring a set number of objects (nursing homes) on each of the variables of interest. A measure of difference between each pair of nursing homes is obtained. Porters competitive strategy typology wa s used to define the strategic groups of nursing homes. Similar to Marlin et al. 1999 and Zinn et al. 1994, this study used the scope and resource commitment measures in the cluster analysis. Prior studies argue that due to the large interstate variation in legislation, they could only use one state for their analysis (Zinn et al. 1994, Marlin et al 1999). Although it is true that states have different governing laws, different reimbursement systems and different survey methodologies, limiting the analysis to one state lim its the generalizability of the results. Rather than limiting to one state, this study standardi zed all variables using z-score based on the mean of the nursing home market (c ounty). Regulation and reimburseme nt may vary not only by state but also by county. Similarly income and demogr aphics can vary by county. All these factors may influence the nursing homes strategy on how to use their resources and how much technology to provide. Therefore it is necessary to compare nursing homes to others in their county. Z-score or normal score is a dimensionless quantit y derived by subtracting the population mean from an individu al raw score and then dividi ng the difference by the population standard deviation. In this study, the county mean and the county standard deviation were used. This study used an inductive methodology to determine the strategic group clusters. Kmeans is a common method used to determine cl usters using inductive reasoning (Ketchen 1996, Harrigan 1981). A significant numbe r of studies have used k-me ans to identify the strategic groups (Marlin et al. 1999, Ford et al. 1998, Zinn et al. 1994, Short et al 2002, Dess et al. 1984).

PAGE 62

62 K-means cluster analysis was specifically desi gned for large datasets (Rudolph et al. 1991). Since this study began with over 11500 observati ons, K means cluster analysis was the most appropriate. K-means cluster analys is requires the researcher to es tablish the number of clusters a priori. This presents a challenge because it is not known if the number of clusters selected a priori is the most well defined clusters. Using theory to determine the number of clusters is considered to be a valid method to identify the k-means a priori (Ketchen et al. 1996 ). However, Ketchen et al. (1994) also r ecommended using a quantitative method in conjunction with the inductive method to increase the robustness of th e number of clusters se lected. One such method is the Wards minimum variance method which has been used by numerous strategic group studies (Marlin et al. 1999, Ford et al. 1998, Zinn et al. 1994, S hort et al. 2002). One problem with the Wards minimum varian ce method is that it functions poorly with very large datasets and it is very susceptible to outliers. FASTCLUS wa s used as technique to reduce the variables into a smaller number of cl usters that would be more manageable for the Wards minimum variance method (SAS 1990). FAST CLUS uses k-means cluster analysis to reduce the number of observati ons to a level that the Wards minimum variance can manage. In this study, the initial FASTCLUS reduced the data set to 100 cluster/observations. The study also removed outlier observations that were greater than five standard deviations. Wards minimum variance method was used for determining the number of clusters to be used. The study used the following decision criteria to identify the optim al cluster solutions (SAS 1990). 1. a local peak in the cu bic clustering criterion 2. a local peak in the pseudo F statistic, 3. a small value of the pseudo t2 st atistic and a larger pseudo t2 statistic for the next cluster solution level, 4. an additional cluster increasing the over all fit by less than 5 percent, and 5. The clusters obtained explaining at leas t 65 percent of the overall variance.

PAGE 63

63 These decision criteria are consis tent with those used in prior strategic group research (i.e., Fiegenbaum and Thomas 1990; Mehra 1996; Marlin 1999). The number of clusters identified by the Ward s minimum variance method was used in the k-means cluster analysis using the original dataset. Once clusters were identified, the means for each one of the variables used to identify the cluster. These mean values were used to determ ine which of the Porters generic strategies the cluster could be categorized into. Strategic Groups Defined As a result of using z scores for the cluster analysis, the variable va lues had a mean score of zero. In order for the model in Figure #2 to be utilized, it was necessary to convert the values into a positive scale. Therefore the clusters were ranked from highest to lowe st from 0 to the total number of clusters that were determined in the cluster analysis. For example, if there were a total of 3 clusters identified in the analysis, then the ranking would range from 0 to 3. The actual distance between clusters for each variable was not equal and therefore it was necessary to weight the rank sc ore relative to the total varian ce of the variable. The ranking protocol of each variable was determined as follows: Determine the total variance value The total variance for each variable was determined by subtracting the lowest mean value from the cluster with the highest mean value. Fo r example, if for a give n variable, the cluster with the lowest values was -1.5 and the cluster with the highest value was +1.5. Therefore, the total variance value was 3. Cluster value For each variable, a cluster value was dete rmined by subtracting the cluster with the lowest value from each individual cluster variables value. Us ing the above example of the

PAGE 64

64 lowest value being -1.5 and a third cluster variab le having a value -0.5, the cluster value of this cluster value would be 1 (-0.5-(-1.5)). Cluster rank score The cluster value was divided by the total va riance value. This ratio was then multiplied by the total number of clusters Using the above information and assuming only three clusters, the cluster rank score would be 1/3 multiplied by 3 for a rank score of 1. For this case, the rank scores would be 0 for the lowest rank, 1 for the middle rank and 3 for the top rank. Table A-8 in the appendix includes all th e total variance and cluster values for each variable. All resource variables and all scope variables were summed to generate a cumulative resource utilization score and cumulative scope variable score each cluster. These total values were then plotted on the model as illustrated in Figure 2. Each cluster was classified into strategies using the parameters defined below. Differentiator : For a nursing home to differentiate itself from the other facilities in its market, it would provide the greatest amount care. It is expected that a differentiating nursing home would be highly ranked in the skin car e technology, restorative nursing technology, other care technology such as bladder tr aining or weight program, and highest rehab care relative to the average nursing home in their county. To be c onsidered a differentiator, the cluster would have to have the highest cumulative scope score. Cost leadership : Some nursing homes will be focused on keeping their costs low. Since staffing is the significant cost driver in the nursing home indus try, strategies to have lower staffing levels can be used as a measure of cont rolling cost. Therefore it was expected that the cost leadership group would have the lowest cumulative rank scores for therapy resources, nursing resources and wages. Best cost : Some nursing homes may seem like a hybr id of the cost leadership group and the differentiator group relative to the lack of strategy group. It was expected that this group

PAGE 65

65 would have higher cumulative rank scores for re sources and scope when compared to the cost leadership group and would have lower cumula tive resources and scope scores than the differentiator. In Figure #2, the cen troid best cost cluster would ha ve fallen in the area between the boundaries set by the cost leaderships centr oid and the differentiati ng clusters centroid. Focus : Like any industry, it is possible for a nursing home to focus their strategy on a smaller market segment. The focus group can be identified by pr oviding a high amount of technology and resources in a spec ific area like rehab services. For example, if a cluster had high rank scores in post acute reha b, long term care rehab and ther apy resources, but low scores in every other category; they w ould be considered a rehab focu s group. There was the possibility that other focus groups also existed. Lack of strategy: A facility that uses more resources but provides less technology than a cost leader was considered as a facility that lacks a strategy. Analysis of Variance This study used Analysis of Variance (ANOVA) to determine if the component variables of the newly identified strategic groups are significantly different from each other. The study also used Tukeys method to make multiple pa ir wise comparisons of the means for each variable. The ANOVA used the F test statistic which is the ratio of the between error estimates divided by the within error estimates. If Ho is false, the between estimates tend to overestimate the variance, so it tends to be larger than the within estimates. This can make the F statistic significantly greater than 1.0 (Agresti and Finlay 1997). The Ho Hypothesis is that the mean of each variable of interest (for example RNs per resident) was the same for each strategic group; that is Differentiator = Cost Leadership = Best Cost = Focus = Lack of Strategy. To determine whether the entire clusters are significant from each other, the study also used the MANOVA test with Wilks' Lambda. Similar to ANOVA, this statistic evaluates the hypothesis that the

PAGE 66

66 population means are equal across groups; however, th is method uses a multivariate approach. It was used in Zinn et al. 1994 study that examined strategic groups. Hypothesis #2: Nursing Homes in the Different iator Strategic Group will have the Highest Quality. Dependent Variables: There were a total of 8 dependent variables that were qual ity measures; five of which relate to long term care quality, two of which measure post acute quality and one measure of the overall quality of the facility. The long term care (LTC) and post acute (PA) measures were quality indicator (QI) variables; they include d LTC bladder decline in continence, LTC bowel decline in continence, LTC 4 poi nt decline in ADL function, the LTC prevalence of the pressure sores, LTC restraint prevalence, PA walking impr ovement, and PA pressure sore incidence. The bladder and bowel decline variables are measures of the proportion of residents who had become incontinent. The ADL decline in function is a risk adjusted ratio of the proportion of the residents who experienced a 4 point ADL decline (out of 16 points) relative to the total number of residents in the nursing home. The restraint prevalence is the proportion of residents that are restrained at the facil ity. The pressure sores prevalence is a risk adjusted ratio of the proportion of residents who have a pressure sores relative to the total nu mber of residents in the nursing home. The walking improvement variable measures the change in ambulation status between the 5 day assessment and the 14 day assessment. Pressu re sores incidence variable is a measure of the amount of post acute residents who developed a pressure sore while remaining in the nursing home relative to the total number of post acute residents in the nursing home. The quality indicators were continuous variables calculated using the MDS based on the methodology as described by Abt Associates 2004. These quality measures are included on the Nursing Home Compare website and are currently used by soci al workers and consumers to differentiate between good and bad facilities. The eighth quality variable was th e total number of quality of care deficiencies. This is an overall measure of the facilitys quality. Quality of care deficiencies

PAGE 67

67 is a count variable that represen ts the number of citations issued by state survey agencies and is located in the OSCAR dataset. Deficiencies are categorized acco rding to level of seriousness from letter A to L. Deficiencies categori zed F and higher denote those with varying degrees of pervasion, which have the potential to, or actually ca use harm or immediate jeopardy to residents (ONeill et al. 2004). It also is on the Nursing Home Compare Website and is used by social workers and consumers to differe ntiate between good and poor facilities. Independent Variables The independent variables of interest were dichotomous variable (0 or 1) for each strategic group generated by the cluster analysis. The expected strategic group variables include: differentiator strategy, cost lead ership strategy, best cost stra tegy, focus strategy, and lack of strategy for those facilities th at did not have any strategy. Control Variables The control variables in the study included both organizational and market factors. They included size, ownership, chain affiliation, ma rket competition, occupancy, payer mix and case mix. Nursing home size is measured using the number of beds. Larger size provides more resources and gives an advantage when trying to recruit staff. Similar to size, chain affiliation gives more resources to draw upon which can also improve the facility efficiency. Ownership and chain affiliation in nursing homes has been associated with lower quality of care (Harrington et al. 2001). The percent Medica re and percent private pay be ds are outcomes of effective strategy because both give the facility an ab ility to increase their revenue streams and subsequently provide the facility with more resources to provide quality of care. Market competition and the occupancy of the facility can influence the behavior of a facility as they attempt to attract more profitable residents. A facilitys case mix is the measure of the acuity of the nursing homes residents. Highe r acuity of residents may be more costly and more difficult to care for and may influence quality measures like pressure ulcers and ADL decline that is beyond

PAGE 68

68 the control of the facility. Several studies have utilized the Acuindex as the case mix variable (Johnson et al. 2004, Laberge et al. 2005, Weech-Mal donado et al. 2006). The Acuindex is an acuity variable located in the OSCAR dataset. It is represented by di fferent factors including mobility issues like how much a ssistance the resident requires with functional activities like ambulation or transfers, or nur sing issues like propor tion of residents receiving suctioning or proportion of residents receiving intravenous thera py. This study also included other measures of acuity that included ADL index, tube feeds, respir atory care, suction, tracheotomy care, IVs, and injections. These measures were included because different strategic groups may have different types of acuity. For example, a strategic group may have a high proportion of the residents that need IVs, respiratory care and need injecti ons but who are independent with their ADLs. Although the Acuindex is a good overall measure of acuity, it cannot capture the intricacies of the different groups that could be related to the different stra tegies. These different nursing acuity measures were not used in the cluster an alysis because they represent the needs of the residents rather than the technol ogy that was provided to them. For example, knowing that half the residents need respiratory care is not the same as a measure that represents the amount of treatment or number of different kinds of tr eatment provided. Although th ey were not used in determining the different groups, these variables ar e needed to be included as controls because they can influence how the techno logy of the facility is provided. Analysis The study intended to use the ordinary least squares (OLS) regression to analyze seven of the quality of care variables (AREG function on STATA) (LTC bladder decline in continence, LTC bowel decline in continence, LTC 4 point de cline in ADL function, the LTC prevalence of the pressure sores, LTC restraint prevalence, PA walking improvement, and PA pressure sore incidence). Prior to the analysis, it was necessary to evaluate the dependent variables in order to ensure that they have normal distribution. If any of the variables were not normally distributed,

PAGE 69

69 the dependent variables were transformed by using the log function. If the initial log transformation did not result in proper skewne ss and kurtosis, other transformations were attempted. Other possible transfor mations included the cubed root the root, the square and the cube. All of them would be attempted until a transformed variable was identified that had a skewness and kurtosis closest to 0 and 3 respectively. If the skewness was not able to be resolved and the variable seemed to have a gamma distribution, the study used the generalized linear model with gamma distribution and a log link. After the OLS an alysis, it was necessary to determine if the regressions pass the OLS linearity assumptions. The linearity assumption was tested through observation of the residual-predicted values plots and the Hosmer-Lemeshow test. The quality of care deficienci es is a count variable and wa s analyzed using the negative binomial equation. The study used Whites heteroskedacity-consistent covariance matrix (1980) to correct for a heteroskedascity. It used county level fixed effects. Dependent variable outliers of 5 standard deviations and above were dropped from the dataset. Qual = + g1Group1 + g2Group1 + ..+ gnGroupN + cControl + feFE + Qual represents one of the se ven quality dependent variables. represents the constant value, while gn represents the effect of each strategic group on the quality variable. gc represents the effect of the contro l variables on the quality variable. fe represents the effect of the county level fixed effects. represents the random error. Hypothesis #3: Nursing Homes in the Focus Strategic Group will provide Higher Quality in the area that they are focused on compared to other Strategic Groups. Dependent Variables Hypothesis #3 used the same quality of care dependent variables as Hypothesis #2. Independent Variables As in hypothesis #2, the independent variables of interest included a dichotomous variable (0 or 1) for each strategic group gene rated by the cluster analysis. However, unlike

PAGE 70

70 hypothesis #2 where the differentiator was used as the reference group; this hypothesis used the focus strategic group variable as the reference group. Control Variables Organizational characteristics that were used as controls included size, for-profit status, chain affiliation, payer mix, and case mix as described in hypothesis #2. Analysis As in hypothesis #2, the planned analysis for the QIs was the OLS with county fixed effects (AREG function on STATA. The study used Whites heteroskedacity-c onsistent covariance matrix (1980) to correct for a heteroskedascity. It used county level fixed effects. Dependent variable outliers of 5 standard deviations and above were dropped from the dataset. Focus Qual = + g1Group1 + g2Group1 + .. +gnGroupN + cControl + feFE + Qual represents one of the se ven quality dependent variables. represents the constant value, gn represents the effect of each strategic group on the quality variable. gc represents the effect of the control variab les on the quality variable. fe represents the effect of the county level fixed effects. represents the random error. Hypothesis #4: Nursing homes in the Cost Le adership Strategic Group will have the lowest costs. Dependent Variables The patient cost equation used operating expe nse per resident day. It is determined by dividing the operating revenue by the total amoun t of residents and dividing it by 365 days. Independent Variables The independent variables of interest include d a dichotomous variable (0 or 1) for each strategic group generated by the cluster analysis. Since the focus of the cost leader was to have the lowest cost, the cost leadership strategic group was used as the reference group.

PAGE 71

71 Control Variables Organizational characteristics that will be analyzed are size, for-profit status, chain affiliation, payer mix, market competition and case mix as described in hypothesis #2. Analysis Similar to health care expenditure measures the cost per patient per day tends to be skewed to the right with some facilities having si gnificantly higher cost than the mean. Similar to prior studies dealing with health care cost, th e study used the genera lized linear model (GLM) with gamma distribution and a log link (Fleishman et al. 2006, Pagano et al. 2005). The study then determined the predicted value using the smearing method. The study used c ounty level fixed effects. The study used Whit es heteroskedacity-consistent covariance matrix (19 80) to correct for any heteroskedascity as found in STATA (2001). It used county level fixed effects. Dependent variables outliers of 5 standard deviations and above were dropped from the dataset. lnCost = + g1Group1 + g2Group1 + ..+ gnGroupN + cControl + feFE + lnCost represents one of the seven quality dependent variables. represents the constant value, gn represents the effect of each strategic group on the quality variable. gc represents the effect of the control variab les on the quality variable. fe represents the effect of the county level fixed effects. represents the random error. Hypothesis #5: Differentiation, Cost Leadership, Focus and Best Cost Strategic Groups will have better Financial Performance than Nursing Homes that Lack a Strategy Dependent Variables There were two variables used to examine fi nancial performance. The first of the two financial performance variables used, the operatin g margin, is calculated using the operating revenue minus the operating expense then divide d by the operating revenue. Operating revenue and expense were direct measures of doing business and therefore the operating margins may be more sensitive to nursing home strategies. The second financial performance variable included total profit margin. Profit margin not only includes the operating revenue and operating expense,

PAGE 72

72 but also includes the non patient revenues like charitable donations and non patient expenses like capital costs. Although profit margin is not as sensitive as operating margins to changes in the nursing homes strategies, it is still important becau se it does represent th e overall picture of the nursing homes financial performance. Financial performance is important because it strongly influences the nursing homes behavi or. If strategies that produces higher quality re sult in better financial performance, then customers needs and providers goals will be aligned. However, if a strategy provides higher quality of care but does not provide the facility with better financial performance, alternative strategies by th e nursing homes will likely be considered. Independent Variables The independent variables of interest include d a dichotomous variable (0 or 1) for each strategic group generated by the cluster analysis. The lack of strategy group was used as the reference group since it was expected that nursin g homes that had a strategy would have better financial performance than those that did not have a strategy. Control Variables Organizational characteristics that were utilized as controls included size, for-profit status, chain affiliation, Medicaid and case mix as de scribed in Hypothesis #3. Unlike the other hypothesis, hypothesis #5 only used Medicaid as a control vari able. Some nursing homes are trying to acquire the high er reimbursing Medicare as part of their strategy. By using the percent Medicare variable as a control variable, it would not be po ssible to determine the benefit of having a strategy directed towa rds acquiring rehab patients. Analysis Although the operating margin a nd the total margin were both continuous variables, neither of them is normally distributed and ther efore Ordinary Least Squares regression is not appropriate. By the nature of the mathematical equation that is used to determine them, their range is from negative infinity to positive one. This is because both equations use the revenue of

PAGE 73

73 the facility in the numerator and the denominator and both equations subtract the costs of the facility from the revenue in the numerator. Therefore the highest value that the operating and total profit margins can be is one ((revenue minus no costs) / revenue = 1). Since the variable is not normally distributed, and cannot be transformed in a parsimoni ous way, a different analytical tool is necessary. In addition, the focus of the study is to dete rmine if the total and the operating margins vary by strategic groups. As a result, determining the odds of being a higher financial performer is appropriate. Therefore, the study used an orde red logit with odds ratio analysis to determine the odds that the strategic group will be in the highest group. Weech-Maldonado et al. 2008 used the ordered logit model with odds ratio in their study that examined quality and financial performance. The study determined the predicted probabilities of a strategic group being in each financial tier for each strategic group. By providi ng the predicted probabilities, it was possible to identify which strategies have the best financ ial performance for opera ting margin and total margin. The study divided the operating margin and the total margin variables into three financial performance tier: low (bottom 10%), medium (middle 80%), and high (upper 10%). Similarly, since different counties can have si gnificantly different market forces between them, it was necessary to control for these count y differences by using co unty fixed effects. Predicted values using logit we re used to identify predicted probabilities by tier for each strategic group. Ologit ( ) = + g1Group1 + g2Group1 + ..+ gnGroupN + cControl + feFE + lnCost represents the co st dependent variable. represents the constant value, gn represents the effect of each stra tegic group on the quality variable. gc represents the effect of the control variables on the quality variable. fe represents the effect of the county level fixed effects. represents the random error

PAGE 74

74 CHAPTER 4 RESULTS Hypothesis #1 Identification of Different Factors: The study completed three separate factor analyses to reduce the total number of variab les. The newly generated compos ite variables represented valid constructs of nursing homes technology and resource commitments. Initially there were 57 technology variables and 11 resource variables that were used in the factor analysis. Factor analysis #1 (Technology) : The first factor analysis included only the variables that measured the technology of the nursing homes. Of the 57 reported variables, there were 20 factors that were derived reporting an eige nvalue greater than 1 (table A1 in the appendix), with ten of those factors having eigenvalues between 1.0 a nd 1.54. Although the factors just above one would qualify using the Kaiser (1960) cr iteria, the amount of variati on removed by each subsequent factor after 10 was marginally in adequate. Therefore it was necessa ry to use the scree plot to identify the ideal number of clusters. The scree pl ot indicated a point of inflection between 9 and 11 factors (Figure A-1 in the appendix). Although the eigenvalues could justify the use of 20 factors, the scree plot indicates that the change in the amount of variance contributed by using additional factors beyond 10 appeared to be nonsignificant. Theref ore, the factor analysis for technology only used the 10 factors. All variables with a factor loading above 0.4 were initially included in the each of the 10 factor variables (Table A-2 Appendix A). There we re six variables that were included in more than one factor. Since each of these six variab les can only be in one factor, it was necessary to determine the most appropriate factor for each vari able. The six variables included long term care pressure relieving chairs, long term care pressure relieving beds, long term care and post acute restorative PROM and AROM. The long term care pr essure relieving chairs and beds had loadings that were similar in both factors (0.6). One of the factors also included post acute pressure relieving devices. Therefore this factor generated a clear construc t of pressure re lieving devices.

PAGE 75

75 The ROM variables were included in the restor ative long term care and restorative post acute measures. The ROM variables were also included in a factor that included splinting; however this factor had a Chronbachs alpha score that was below the 0.7 leve l and therefore was dropped. The end result was 8 technology factors with variables that had factor loadings over 0.4 (Table A-3 Appendix A). Factor Analysis #2 (Resource) : The staffing intensity factor analysis identified three factors with eigenvalues greater than one (Table A-4 Appendix A) The scree plot appeared to demonstrate a point of inflection between 2 and 3. The three factors included licensed therapist (PT and OT), therapy assistants (PTA and COTA) and nursing staffing (including RN, LPN, and CNA) (Appendix A table 4). Factor Analysis #3 (Resource) : For the staffing mix, the eigenvalues were all below 1. Since each of the factors removed less variance th an one variable, it was more appropriate to include each of the staffing mix variables in dividually rather than as a composite. The factor analysis identifi ed 10 technology factors and 3 re source factors. These factors were tested for internal consistency using the Cronbachs alpha method. As illustrated in Table 41, 8 technology factors and 2 resource factors ha d Cronbachs alphas over the 0.7 threshold. These factors included long term care sk in care, post acute ca re skin care, long term care restorative process, post acute restorative processes, relie f devices, therapy minutes, toileting program, weight management program, therapist, thera py assistants. The restor ative splinti ng technology variable was included in the factor that was a ssociated with restorative AROM and PROM yet failed to have Cronbachs alpha score greater th an 0.7. The technology variab les therapy diet and plate guard made up the final factor. This factor was dropped because they failed to have an alpha score greater than 0.7. Nursing (RN, LPN and CNA) on the other hand did have an alpha scor e of 0.69; however it was decided not to use a total sta ffing intensity factor because a nursing home could, as a strategy,

PAGE 76

76 choose to substitute a licensed staff member with a non licensed staff in an attempt to keep costs low. By generating an index that included all th ree nursing variables, the ability of the cluster analysis to identify nursing homes with different types of staffing intensity strategies would be compromised. It was therefore deemed necessary to include the RN, LPN and CNA variable in the cluster analysis as i ndividual variables. Table 4-1. Internal consistency of the composite variables Cluster variable Alpha Variables used Long term care skin processes of care 0.82 Long term care ointment, dressing application process, surgical wound care, ulcer care, skin nutrition and turning repositioning program Post acute skin processes of care 0.72 Post acute ulcer care, dressing application process, Long term care restorative processes 0.86 Long term care restorative dressing, restor ative transfers, restorative bed mobility, restorative PROM, Restorative AROM, restorative ambulation and restorative eating Post acute restorative processes 0.91 Post acute restorative dressing, restorative transfers, restorative bed mobility, restorative PROM, Restorative AROM, restorative ambulation and restorative eating Relief Devices 0.81 Long term care pressure relieving bed and chair, Post acute pressure relieving bed and chair Therapy minutes 0.93 Long term care and Post acute physical and occupational therapy Toileting program 0.88 Long term care and Post acute toileting program Weight management prg 0.88 Long term care and post acute weight management program Therapist 0.91 Total physical and occupational therapist per resident Therapy assistants 0.77 Total physical and oc cupational therapy assistants per resident Table 4-2 lists the nursing variables and all the individual variables utilized in the cluster analysis. Other variables included the per cent of nurses that were RNs, percent of nurses that were Table 4-2. Individual va riables used in cluster analysis that were not included in a factor variable. Individual Variables Definition RN Registered Nurse Per Resident LPN Licensed Practical Nurse Per Resident CNA Certified Nurse Assistant Per Resident RN Percent Percent of Nurses are Registered nurse License Nurse Percent Percent of Nurses are Licensed PT Percent Percent of Physical Therapist are licensed OT Percent Percent of Occupa tional Therapist are licensed Wages Facility direct care average wage Percent Medicare Percent re sidents who are post acute licensed, percent of physical therapist, and the percent o ccupational therapist. As described in the methodology, wages and percent Medicare beds were also used as individual variables.

PAGE 77

77 Cluster Analysis: The 10 factor variables and the 9 indi vidual variables were standardized using z scores that were based on the county m ean. The dataset were then trimmed for outliers using the 5 standard deviations for each variable. This resulted in the redu ction of the number of nursing homes used in the cluster analysis from 11601 to 9726. The newly generated variables were used in the Wards Minimum Variance Clus ter Analysis and the kmeans cluster analysis. Ward's Minimum Variance Cluster Analysis identified 5 clusters as the best number of clusters. This was based on a local peak in the pseudo F stat istic, a small value of the pseudo t2 statistic followed by a larger pseudo t2 statistic for the ne xt cluster solution leve l, and that adding an additional cluster increased the overall fit by less than 5 percent (Table A-6 and A-7 Appendix A). No other cluster arrangement met the three criteria. The data was then analyzed using the k-means with an a priori of 5 clusters. Clusters: Table 4-3 represents the results of the cl uster analysis. Using the ranking protocol described in the methodology (Table A-8 Appendix), the different cl usters were defined using the rank score ranging from 0 to 5. 0 represented the lowest and 5 represented the highest value for both the resource variables and th e technology variables. For each cluster, the cumulative rank score was calculated by summing the values fo r the scope variables and summing the resource variables. These cumulative rank scores were plo tted on a graph to generate model as indicated in Figure 4-1. The scope score was represented by th e horizontal axis and the resource score was represented by the vertical axis. Table 4-4 provided some descriptive structural, market and case mix variables of each strategic group. The ANOVA test identified that the clusters we re significantly different from each other for each variable. In addition, the Tukeys test compared each pair of clusters for every variable and found that there were significant differences be tween each pair 79% of the time. Finally, the MANOVA test criteria (Wilks' Lambda) found that th e entire clusters were significantly different from each other (p <.000). Descriptives of a ll variables is located in table A-9 in the appendix.

PAGE 78

78 Table 4-3. Rank scores scope and resource variables Variables Cluster 1 Differentiator Cluster 2 Cost Leader Cluster 3 Rehab Focus Cluster 4 Specialty care Focus Cluster 5 Lack of Strategy Average Wages 1.985 0.813,4 4.042,5 5.001,2,5 0.001,2 F=13.8*** RN per Resident 1.70* 0.00* 2.72* 5.00* 0.40* F=744*** LPN per Resident 0.993,4 1.073,4 5.00* 0.00* 0.863,4 F=121*** CNA per Resident 4.362,3,4 3.59* 5.00* 0.00* 4.052,3,4 F=367*** Therapist per Resident 1.00* 0.00* 5.00* 1.75* 2.23* F=1110*** Therapy assistant per resident 1.19* 2.48* 5.00* 1.62* 0.00* F=1492*** Percent Licensed PT 2.542,3,5 0.00* 1.58* 2.442,3,5 5.00* F=1180*** Percent Licensed OT 2.612,3,5 0.00* 1.94* 2.602,3,5 5.00* F=906*** Percent RN 1.15* 0.00* 1.42* 5.00 0.25* F=1538*** Percent Licensed Nurse 0.41* 0.071,3,4 1.76* 5.00* 0.001,3,4 F=833*** Total Resource 17.91* 8.06* 33.49* 28.32* 17.78* F=186*** LTC Skin Process 3.23* 0.60* 5.00* 2.07* 0.00* F=186*** Acute Skin Process 4.682,4,5 1.56* 5.002,4,5 3.03* 0.00* F=58.35*** LTC Restorative Process 5.00* 0.101,3,4 0.89* 0.34* 0.001,3,4 F=688*** Acute Restorative Process 5.00* 0.141,3,4 0.00* 0.33* 0.141,3,4 F=1987*** Relief Devices 4.112,4,5 1.672,4,5 5.001,4,5 1.512,4,5 0.00* F=39.8*** Toilet Program 4.96* 1.49* 0.111,2 0.091,2 0.001,2 F=19.7*** Weight Program 5.003,5 3.733,5 0.001,2,4 2.533 1.031,2 F=8.62*** Therapy Minutes 0.00* 1.601,3,5 5.00* 1.681,3,5 1.13* F=386*** Percent Medicare 0.103,4 0.153,4 5.00* 0.74* 0.003,4 F=918*** Total Scope 32.1* 11.03* 26.10* 12.32 2.29 F=59*** *Significantly different from all clusters, Superscript is significantly di fferent than identified groups The descriptives values were not adjusted for th e county level fixed effects and therefore were not used in the determination of the clusters. The plotted clusters in Figure 4-2 were then categorized using Porters generic strategies. Based on the definition describe d in the methodology, Cluster #1 was initially classified as Differentiator because its scope score was the hi ghest (technology cumulative rank score). However this group was ranked the lowest in all the categories that was related to rehab (i.e. physical therapy minutes and percent Medicare beds). Therefore this cluster may be more appropriately classified as a nursing care focus group rather than a differentiator. Although these nursing homes provided a high degr ee of technology, they did not use the highest amount of resources. They tended to have median level wages, median level total RN staffing, and below median level of percent licensed therapist. They had low staffing mix as a result of their high

PAGE 79

79 CNA staffing levels. This group al so tended to have high acuity pr imarily driven by residents who were more dependent and patients who had tube feeds. Table 4-4. Rank scores of st ructure and market factors Variables Cluster 1 Differentiator N=990 Cluster 2 Cost Leader N=2752 Cluster 3 Rehab Focus N=1816 Cluster 4 Specialty care N=1634 Cluster 5 Lack of Strategy N=2534 Percent for profit 71.2% 75.7% 79.7% 76.9% 76.2% Chi2= 27 Percent chain affiliated 51.5% 60.0% 66.4% 63.0% 58.2% Chi2= 70 Market competition 1.43 0.00 0.71 5.00 4.29 F=0.17 Occupancy 5.00 3,4 4.53 3,4 0.00 1,2,5 1.98 1,2 3.16 3 F=12.15 Size 0.67 2 5.00* 1.15 2,5 1.86 2,5 0.00 2,3,4 F=59.23 Case Mix 4.57 2,4,5 1.06 1,3 5.00 2,4,5 2.18 1,3,5 0.00 1,3,4 F=25.95 DL index 4.73 2,4,5 1.14 1,3 5.00 2,4,5 2.08 1,3,5 0.00 1,3,4 F=21.45 Percent tube feed 4.22 2,5 1.19 1,3,4 4.51 2,5 5.00 2,5 0.00 1,3,4 F=9.65 Percent tracheotomy 3.485 1.49 4 3.15 5 5.00 2,5 0.00 1,3,4 F=14.63 Percent requiring suctioning 2.96 2,5 0.71 1,3,4 3.39 2,5 5.00 2,5 0.00 1,3,4 F=15.63 Percent requiring respiratory care 1.28 3 0.50 3,4 5.00* 2.67 2,3,5 0.00 3,4 F=42.13 Percent requiring IV 0.64 3,4 0.74 3,4 5.00* 1.87* 0.00 3,4 F=54.7 Percent requiring injection 0.98 3 0.78 3 5.00* 2.19 3,5 0.00 3,4 F=33.18 *Significantly different from all clusters, Superscript is significantly di fferent than identified groups Cluster #2 was classified as the cost leader because it used the least resources. The only nursing resource that the cost leader wa s not the lowest on was the CNA staffing. The lowe st RN staffing coupled with the median CNA staffing resulted in the lowest staffing mix. Cost leadership groups also had the lowest amount of th erapist with an above median level of therapy assistant. This resulted in the low therapy staffing mix. Cost le aders were significantly la rger that other nursing homes and tended to have lower acuity. These were all consistent with wh at would be expected with a cost leader strategic group. Cluster #3 was classified as the rehab focus cluster. It had higher wages and the highest staffing intensity. It had an above median level of RN staffing and above median percent licensed staff. Rehab focus nursing homes provided the highe st skin services and therapy. They had the highest use of resources and provided a significant amount of service second only to the differentiators. This cluster cannot be classified as best cost because although the nursing homes

PAGE 80

80 provided more service than the co st leadership group, they also used more resources that the differentiators. The rehab focus group had th e highest acuity in most categories. Cluster #4 was classified as specialty care focus group. It had the highest overall wages, overall RN staffing and RN proportion when compared to the other clusters. This cluster had skin care services and weight program, and overall cas e mix acuity in the middle range. This cluster did not use the least resources, or provide the most processes and therefore could not be classified as either cost leader or differen tiator. It could not be considered a best cost strategic group because although this cluster provided a greater amount of car e than the cost leader cluster, it used more resources than the differentiating cluster. Alt hough this group did not have the highest overall acuity level, it did have the highest amount of resi dents with specialty care needs like tube feeds and tracheotomies. However, these specialty care services have li ttle effect on the overall case mix score (acuindex) because these types of servi ces account for only 20% of the acuindex variable (while the ADL factors account for 80% of the variable). Cluster #5 was classified as the lack of strategy cluster. This cluster of nursing homes appeared to lack any specific st rategy. They had patients with th e lowest acuity, and they provide by far less services than other groups while their st affing levels still remained at a median level. All the clusters were plotted on the strategic group model (Figure 4-1) using the cumulative technology and cumulative resource scores from tabl e 4-3. Hypothesis #1 was partially supported. The results indicated that there exist four distin ct strategies in the nur sing home industry, however these groups were different then what was or iginally hypothesized. Based on the Figure #3, there was no best cost group; instead ther e was a fifth group that was classi fied as a specialty care focus group. Figure #3 depicts the existence of a differentiator group based on the methodology, however closer examination of the individual pr ocesses of care indicates a nursing care focus strategy.

PAGE 81

81 Figure 4-1. Strategic group mode l based on scope and resource commitment in the nursing home industry Normal distribution: All of the eight quality dependent variables had excessive skewness and non-normal kurtosis levels and therefore had to be transformed. After being transformed using the either the root or log func tion, the skewness and the kurtosis of the quality valuables were long term care (LTC) bladder decline (-0.03, 3.62), LTC bowel decline (0.15, 3.19), LTC 4 pt ADL decline (0.13, 3.77), LTC restrain ts (0.18, 2.52), LTC pressure ulce r prevalence (-0.09, 3.81), Post Acute (PA) walking improvement (0.315, 4.15) and PA pressure ulcer incidence (-0.78, 4.32). The PA measures both failed the normal distribution a ssumption even after transformation. Both post acute variables were therefor e analyzed using the generali zed linear model with gamma distribution. Linearity: By observing the predicted value and residual plot, all six quality variables that were used in the OLS regression model appeared to pass the linearity assumption. The study also confirmed these results with a negative Hosmer-Lemeshow test (non-significant) for all the dependent variables.

PAGE 82

82 All the above quality variables with skew ness under 4 were analyzed using OLS with county level fixed effects. The quality variables that had skewness over 4 were initially analyzed by using the generalized linear model (glm ) with gamma distri bution and a log link. Unfortunately, the glm was not able to find a solution when using post acute pressure ulcer incidence quality variable. Therefore the PA pr essure ulcer incidence was analyzed using OLS with fixed effects. Hypothesis #2 Hypothesis #2 was partially supported. Table 4-5 (f ull regression results ar e located in tables A-9 to A-16 in the appendix) illustrates that the differentiator strategic group had either the same or better long term care quality than all other strategic groups in all measures except long term care pressure ulcers. Differentiators had lower 4 point ADL decline than cost leaders and those nursing homes that lacked a stra tegy. Differentiators also had fe wer pressures sores than nursing homes that had a rehab focus or had a specialty care focus. However, with respect to the post acute measures, differentiators performed poorly. The differentiator strategic group generally had worse outcomes with less walking improvement th an all other groups. Differentiators had higher incidence of post acute pressure sore than lack of strategy and cost leaders. Table 4-5. Strategic group quality of care with Di fferentiator / nursing care focus cluster as the reference group Dependent Variables Cost Rehab focu s Specialty care Lack of Strategy Quality of care deficiencies 1.06 (0.09)* 0.99 (0.86) 1.00 (0.97) 1.12 (0.00)* Long Term care measures Bladder Decline 0.006 (0.04)* 0.012 (0.00)* 0.009 (0.01)* 0.005 (0.12) Bowel Decline 0.008 (0.02)* 0.009 (0.03)* 0.012 (0.00)* 0.005 (0.16) ADL 4 pt decline 0.008 (0.00)* 0.003 (0.39) 0.005 (0.10)* 0.004 (0.13) Restraint use 0.011 (0.01) 0.004 (0.46) 0.001 (0.19) 0.001 (0.19) Pressure Ulcers -0.001 (0.66) 0.012 (0.00)* 0.004 (0.14) -0.004 (0.09) Post acute measures Ambulation improvement 0.031 (0.10) 0.050 (0.02) 0.037 (0.08) 0.060 (0.00) Pressure Ulcers incidence -0.045 (0.02) -0.015 (0.46) -0.027(0.20) -0.076 (0.00) Quality of care results ar e incidence rate ratio Square root transformation LTC bladder decline, bowel decline pr essure ulcer prevalence, LTC 4 point ADL decline, restraints, P ost acute pressure sore incidence are OLS regression coefficien ts. Post acute walking improvement was gl m with gamma distribution. Represents s ignificant values in the desired direction. Represents significant values in the negative direction.

PAGE 83

83 Hypothesis #3 This hypothesis was partially supported. As de scribed above, nursing homes that focus on providing nursing care had good quality. However, th e other to focus groups did not have similar results. The rehab focus group had the highest re hab staff, provided a highest amount of therapy minutes, provided the highest amount of skin care and had the highest proportion of post acute beds. This group was expected to have the best post acute outcomes. The second focus groups had a high proportion of their nurses that were RNs and had the highe st proportion of patients with specialty care needs like tracheotomies, tube f eeds and patients who required suctioning. This group was expected to have better overall care than those nursi ng homes that lacked strategy and better quality outcomes than the cost leaders. Table 4-6 (full regression results are avai lable in table A-18 to A-25 in Appendix A) illustrates that nursing homes in the rehab focu s group had lower ADL 4 pt decline and lower restraint use than the cost leaders. In addition, the reha b focus had lower quality of care deficiencies than the cost leaders and the lack of strategy group. This group only had better PA outcomes than differentiators. This group generally had poorer LTC quality of care Table 4-6. Strategic group quality of care with rehab focus cluster as the reference group Dependent Variables Cost Differentiator / Nursing care focus Specialty care Lack of Strategy Quality of care deficiencies 1.07 (0.03)* 1.01 (0.86) 1.01 (0.80) 1.13 (0.00)* Long Term care measures Bladder Decline -0.006 (0.05) -0.012 (0.00) -0.003 (0.39) -0.007 (0.02) Bowel Decline -0.001 (0.77) -0.009 (0.03) 0.002 (0.52) -0.004 (0.21) ADL 4 pt decline 0.006 (0.03)* -0.003 (0.39) 0.002 (0.39) 0.002 (0.53) Restraint use 0.011 (0.01)* 0.004 (0.46) 0.006 (0.19) 0.006 (0.22) Pressure Ulcers -0.013 (0.00) -0.012 (0.00) -0.008 (0.00) -0.016 (0.00) Post acute measures Ambulation improvement -0.019 (0.28) -0.050 (0.02)* -0.014 (0.47) 0.014 (0.57) Pressure Ulcers incidence -0.030 (0.06) 0.015 (0.46) -0.012 (0.49) -0.061 (0.00) Quality of care results ar e incidence rate ratio Square root transformation LTC bladder decline, bowel decline pr essure ulcer prevalence, LTC 4 point ADL decline, restraints, P ost acute pressure sore incidence are OLS regression coefficien ts. Post acute walking improvement was gl m with gamma distribution. Represents significant values in the desired direction. represents significant values in the negative direction.

PAGE 84

84 especially in comparison to the differentiator and the lack of strategy grou p. This group had higher LTC prevalence and higher PA incidence of pressure ulcers than cost lead er and lack of strategy group. Table 4-7 (full regression results are avai lable in table A-26 to A-33 in Appendix A) illustrates that nursing homes in the specialty care focus group had lower quality of care deficiencies than the cost leaders and the lack of strategy group. They had less pressure ulcer prevalence than facilities in the rehab focus group. The specialty care focus group had higher prevalence of pressure ulcers than cost leader and lack of strate gy. It had higher bladder decline, bowel decline, ADL 4 point decline than differen tiators. It had lower restraint use than cost leaders. With respect to post acute incidence of pressure sores, the results were mixed with specialty care focus having a higher incidence of pressure ulcers than l ack of strategy group. And a higher prevalence of long term care pressure so res that cost leaders and lack of strategy. Table 4-7. Strategic group quality of care with Speci alty care focus cluster as the reference group Dependent Variables Cost Differentiator / Nursing care focus Specialty care Lack of Strategy Quality of care deficiencies 1.06 (0.05)* 0.99 (0.80) 1.00 (0.97) 1.12 (0.00)* Long term care measures Bladder Decline -0.003 (0.26) 0.003 (0.39) -0.009 (0.01) -0.004 (0.12) Bowel Decline -0.003 (0.27) -0.002 (0.52) -0.012 (0.00) -0.006 (0.03) ADL 4 pt decline 0.003 (0.16) -0.002 (0.39) -0.005 (0.10) -0.001 (0.78) Restraint use 0.005 (0.22) -0.006 (0.19) -0.002 (0.68) -0.001 (0.88) Pressure Ulcers -0.005 (0.01) 0.008 (0.00)* -0.004 (0.14) -0.008 (0.00) Post acute measures Ambulation improvement -0.005(0.74) 0.014 (0.47) -0.037 (0.08)* 0.024 (0.15) Pressure Ulcers incidence -0.018 (0.25) 0.012 (0.49) 0.027 (0.20) -0.049 (0.00) *Quality of care results are incidence rate ratio Square root transformation LTC bladder dec line, bowel decline pressure ulcer prevalence, LTC 4 point ADL decline, restraints, Post acute pressure sore incidence are OLS regression coeffi cients. Post acute walking im provement was glm with gamma distribution. Represents significant values in the desired direction. Represents significant values in the negative direction. Hypothesis #4 Hypothesis #4 was not supported. Table 4-8 illustra tes that cost leaders nursing homes had significantly lower costs that sp ecialty care and rehab focus groups but no significant difference between cost leaders and differentiators or with those facilities that l ack a strategy altogether.

PAGE 85

85 Table 4-9 provides the cost predicted values for each strategic group. The predicted values are the patient costs per resident day. Sp ecialty care focus and Rehab focus groups had significantly higher costs than the cost leader by 9 and 13 dollars per re sident day respectively. Table 4-8. Strategic groups and costs with cost leader as reference group regression results Lack of Strategy Differentiation / Nursing Care Specialty Care Focus Rehab Focus Costs per resident day -0.006 (0.60) 0.002 (0.90) 0.039 (0.00)* 0.080 (0.00)* Represents significant values compared to cost leadership cluster Table 4-9. Strategic groups and predicted costs Differentiator / Nursing Care Focus Cost Leader Rehab Focus Specialty Care Focus Lack of Strategy Costs per resident day 149.15 149.06 161.65* 155.03* 148.12 In dollars Hypothesis #5 The initial analysis that included all the case mix variables was unable to converge. It was therefore necessary for the study to use acuindex va riable rather than the 7 individual case mix variables. By dropping the 7 case-mix variables, th e ologit analyses were ab le to converge (A-34 and A-35 in Appendix A). The ologit for both financial performance meas ures (table A-34 and A-35 Appendix A) had variables that were completely determined. 53 of the observations for operating margin and 57 observations for total margin observations were co mpletely determined. There was possibility this was the result of multicollinearity between some of the county leve l fixed effects. There was over 1500 county level fixed effects variables with many of them having a very low number of nursing homes within the county. A correlation matrix wa s generated to determine if the completely determined variables were either the independent variables of interest or one of the control variables other than the county fixed effects variables. The correlation matrix (table A-36 appendix A) illustrated that none of the indepe ndent or control variables were completely determined. As a result of the high number of c ounty level fixed effects, the determined variables

PAGE 86

86 made up approximately 5% of all the county fixed effects and therefore likely have minimal impact the overall model. Hypothesis #5 was partially supported. Table 410 highlights that differentiator had a 20% greater odds of being in the highest tier for operating margin and a 31% greater odds of being in the highest tier for total margin relative to thos e nursing homes that lacked a strategy. Cost leaders had 25% greater odds of being in the highest tier for operating ma rgin and 22% greater odds of being in the highest tier for tota l margin relative to those nursing homes that lacked a strategy. Rehab focus had 46% greater odds of being in the highest tier for ope rating margin and 27% greater odds of being in the highest tier for to tal margin relative to those nursing homes that lacked a strategy. Differentiators, cost leaders and rehab focus had better total margins than the specialty care focus strategy but only the rehab focus group and the cost leadership group had better operating margins than the specialty care focus group. The specialty care focus group did not have significantly higher odds of being in th e highest tier with respec t to the nursing homes that lacked strategy. Table 4-10. Odds Ratio of Nursing Home Strategy Dependent Variable Cost Leader Differentiation Specialty care Focus Rehab Focus F test Operating Margin 1.25 (0.01)* 1.20 ( 0.09)* 1.04 (0.66) 1.46 (0.00)* F=9.7 (0.05) Total Margin 1.22 (0.01)* 1.31 (0.01)* 0.98 (0.80) 1.27 (0.01)* F=10.0 (0.03) *Significantly different from lack of strategy Table 4-11 and 4-12 illustrate the predicted pr obabilities of the operating and total margins. Both demonstrate that the nursi ng homes with differentiation, cost leadership and rehab focus strategy perform better than those nu rsing homes that lack strategy. Table 4-11. Probability of nur sing homes being in the hi ghest operating margin tier Dependent Variable Differentiator5 Cost Leader4,5 Rehab Focus4,5 Specialty care Focus2,3 Lack of Strategy1,2,3 Low 0.099 0.096 0.085 0.110 0.111 Mid 0.799 0.799 0.797 0.798 0.797 High 0.102 0.105 0.118 0.092 0.090 # Significant p>0.10

PAGE 87

87 Table 4-12. Probability of nursing homes being in the highest total margin tier Dependent Variable Differentiator4,5 Cost Leader4,5 Rehab Focus4,5 Specialty care Focus1,2,3 Lack of Strategy1,2,3 Low 0.089 0.094 0.094 0.112 0.109 Mid 0.799 0.799 0.799 0.798 0.799 High 0.112 0.106 0.106 0.090 0.092 # Significant p>0.05

PAGE 88

88 CHAPTER 5 DISCUSSION This study concurred with prior research because it demonstrated the ex istence of strategic groups in the nursing home industry. However, this study used a different conceptual framework to identify them. While earlier studies by Zinn et al. 1994 and Marlin et al. 2002 used resources and payer mix, this study used the business leve l decisions of resource allocation and technology provision. One advantage of using technology over payer mix variables to identify strategic groups is the applicability of th e results for managers. Management has control over how much technology to provide and how mu ch resource to use. How a nursing home decides to balance these two factors is an important part of that facility s overall strategy. The results from this study demonstrate that different strate gic groups provide their technology a nd utilize their resources in a different way. Using technology and resource allocation to iden tify the clusters makes it possible to define the groups using Porters generic strategies. There are no nursing home studies that have attempted to define strategic groups by Porters generic strategi es. However, Marlin et al. 2002 did find the existence of Porters strategies in the hosp ital industry. Unlike this study, Marlin et al 2002 explicitly used variables th at defined the differentiator strategic group. These variables included the total number of hos pital units, the number high technology units and ratio of intensive care beds to total beds. They did not use any measures that reflect how the care was provided. Although a hospital could have a high number of different specialty units, it did not necessarily provide a high degree of technology within those units. E ach specialty unit could have been poorly staffed and the patients could have received a very low level of services. Marlin et al. 2002 study was examining the governance level of decision making as opposed to the business level decisions. Business level op erations are significantly impacted by the day to day decisions of how to provide service and how to use resources.

PAGE 89

89 The study identified 5 of the strategic groups listed in the hypothesis (nursing care focus (differentiator), cost leader, rehab focus, specia lty care focus and lack of strategy). The study did not identify the best cost group and did not iden tify a well defined differentiator group. Instead, the study identified three focus groups; rehab focus, specialty care focus a nd a nursing care group. The rehab group from this study appeared similar to the high Medicare fa cilities in the prior nursing home strategic group studies This group not only had a high proportion of Medicare beds, but also provided a high level of therapy minutes and skin care. The specialty care focus group had a high RN intensity and had a high concentration of residents with specific care needs. Residents with specialty care needs like trache otomy care or suctioning require the nursing home that has a capacity to deliver a high level of skilled services. These conditions are also only minimally covered by Medicare in the post acu te setting and have long been considered underserved by MedPAC (MedPAC 2003, 2007). Be ginning in 2004, nursing homes could get more reimbursement for residents with these exte nded needs. This change in reimbursement only started in 2004 and it may be possible that the facilities had not been able to fully take advantage of it. One unexpected finding was that the defined differentiator stra tegic group provided the least therapy services. The fact that this group is defi ned as a differentiator when it provides the lowest amount of therapy brings up the question of the ap propriateness of using differentiator as this groups definition. The group clearl y provided the most nursing car e technology and therefore was reclassified as the nursing care focus strate gic group. Whether this group is considered a differentiator or a nursing care fo cus strategic group, this group still represents nursing homes that are trying to separate themselv es from their competition by pr oviding a high level of technology and therefore they can still be defined by using Porters generic strategies. Out of all the clusters identified, the cost l eader proved to be the best defined. The cost leader group had the lowest staffing levels and lower staffing mixes. Lower staffing mixes would

PAGE 90

90 indicate a cost leader because it indicates that the nursing home is substi tuting CNAs for RNs. In addition, on observation of the structur al features of this group, the cost leaders also tended to be larger and have lower acuity. It is clear that this groups strategy was to keep cost under control by providing a low amount of service. Finally, the lack of strategy group was identified because it used a median amount of resources with almost no technology. In addition, the nursing homes that lacked strategy appeared not to have any specialty services. By using technology and resource deployment, this study found that strategic groups not only exist, but can be classified using Porters generic strategies. The ne xt question is whether strategic-performance relationships exist between the different st rategic groups and quality, costs and financial performance. Quality This study identified that the nursing care focus (differentiator), sp ecialty care focus and rehab focus strategic groups had lower quality of care deficienci es than the nursing homes that were in the cost leader strate gic group or the nursing homes th at lacked strategy. This finding provides some evidence that focusing on providi ng quality either by providing technology or having better staffing mixes can result in some better outcomes. However, only quality of care deficiencies had consistent results through these three stra tegic groups. Quality of care deficiencies is a general measure of the facilities quality and can relate to anything from staff that is inadequately trained to poor care practices The Nursing Home Comp are website lists and describes all of facilities deficiencies from th e most recent state survey. It is a good general measure of quality and has been used in past research (Harri ngton et al. 2000, 2001). One limitation of deficiencies is that it does not hi ghlight specific care outcomes. Other measures like the QIs that were derived by Abt are specifically designed to identify the specific quality issues like pressure ulcer prevalence and ADL declin e. Knowing the specific quality areas would

PAGE 91

91 hopefully help managers address the problems direc tly. In this study the results of these measures were mixed. When compared to other groups, the nursing ca re focus strategic group generally provided better care to their long term car e residents. The nursi ng care focus, however, also had poor post acute outcomes. The fact that this group provided the least amount of minutes of therapy, had a lower proportion of post acute patients, and had the poorest post acute quality provides further support to classify this cluster as a nursi ng technology focus group ra ther than an overall differentiator group. These results suggest that there may be a re lationship between a strategic group focus and the quality of care provided. The other two focus strategic gr oups did not fair as well. Th e rehab focus strategic group had a high level of rehab staffing and provided high amount of rehab technology. The specialty care focus had a high level of RN staffing and provide d care to residents with specific care needs. It would be expected that the rehab focus group would have the bett er post acute care outcomes than other groups. However, rehab fo cus strategic group had a higher incidence of pressure sores and did not have better walking im provement than the specialty nurse, cost leaders or nursing homes that lacked strategy. This was surprising since the rehab focus did not have better outcomes even though they provided significantly more therapy. There are a number of possible explanations. The rehab focus strategic group had higher acuity than other strategi c groups. Although the quality measures used are risk adjusted and the case mix of the facility was controlled for, there may be factors that may not be captured. Mukame l et al. 2008 has determined that the current quality measures including pressure ulcers would bene fit with a greater degree of risk adjustment. The rehab focus group has a high proportion of Medi care patients and may have the expertise to handle patients with more co morbidities that may not be captured by the risk adjustments of the outcome variables. Similarly, different strategi c groups had different ty pes of acuity measures.

PAGE 92

92 Some of the case mix variables may have greater influence on the quality outcomes than others. The rehab focus groups may be penalized by taking on residents that are more prone to pressure ulcers and whose risk factors are not captured by the risk adjustment s of the quality measures or control variables used in the analysis. In the case of the walking improvement, th e rehab focus group provided nearly 25% more minutes then any other group and yet did not attain better outcomes than the specialty care focus, cost leaders or lack of strate gy groups. There are a few reasons that may explain why the rehab focus group does not have better walking impr ovement outcomes. First, rehab focus nursing homes may be downgrading the progress of their residents. Rehab focus strategic groups are accustomed to providing care to rehab patients and may be more rigorous in the application of walking improvement scales. A second reason is that there may be an upper therapy minute threshold where adding more minutes of therapy does not translate into better outcomes. The rehab group provided significantly mo re minutes of therapy when compared to the specialty care focus, the cost leader groups and the lack of strategy group ye t did not have better therapy outcomes. The fact that differentiators provide considerably less therapy and have worst outcomes when compared to all other groups supports the no tion that there is at l east a minimal threshold level of minutes of therapy that needs to be achieve d in order for basic outcomes to occur. Another possibility is that si nce revenue streams are enhanced by providing higher intensity of therapy for a longer period of time, these faciliti es may be more conservative with their 14 day assessments in an effort to increase the length of stay. By having a lower 14 day assessment value, the facility could be able to ju stify keeping the resident at the facility for a longer period time. Over the past 7 years, MedPAC have issued stat ements of concern that the current post acute reimbursement schematic promotes therapy overutiliz ation. More research is needed to determine if the lack of post acute quality outcomes differ ence between the rehab group and the cost leader,

PAGE 93

93 lack of strategy and the specialty care groups is the result of some therapy intensity upper boundary, or if the rehab focus groups are manipulating from the current reimbursement system. It would be expected that the specialty care focus strategic group would have better quality of care outcomes in the areas where high RN staffing levels are needed. Generally, besides the quality of care deficiencies and in comparison to the differentiator, the sp ecialty care group results were somewhat mixed. The fact that differentiators had better outcomes is not surprising since the outcome measures are more closely related to the technology provided by the differentiators. Similar to the differentiator and the rehab fo cus group, the specialty care group had a higher prevalence of pressure ulcers with their long term care residents than the cost leader and the lack of strategy. This was contrary to Weech-Maldon ado et al. 2004 that id entified RN staffing intensity and staffing mix as related to quali ty outcomes like lower pressure ulcers. One explanation for the high pressure sore prevalence could be endogeneity. Nursing homes with a high prevalence of pressure sores may employ mo re RNs so they can treat those residents effectively. Cost Rehab focus and specialty care focus group have significantly higher costs than the other groups. This is not surprising since both groups ha ve a higher amount of licensed staff. What was unexpected is that the nursing ca re focus group (differentiator group) had similar costs to the cost leader group. These results are unexpected be cause nursing care focus group had high acuity and provided the most technol ogy at the same cost as the cost leaders and those nursing homes that lacked a strategy. Past studies have found that be tter quality translates to lower cost in nursing homes (Weech-Maldonado et al. 2003). In the case of this study, there was a relationship between the nursing care focus group and be tter long term care outcomes. Although the nursing care focus group did not have the lowest staffing costs, it may compensate by saving money by providing more effective and efficient care.

PAGE 94

94 Financial Performance The study found that rehab focus, nursing care focus group and cost leaders had better financial performance than the nursing homes that lacked a strategy and the specialty care focus group. The poor financial performance of the specialty care focus group could be related to higher costs of having more RNs without the benefit of generating higher revenue streams. Unlike the rehab group where the more therapy results in mo re revenue, the specialty care focus groups may have limited ability to increase revenue. Medica re and some Medicaid programs can (depending on the state) provide more payment for nursing care; however the reimbursement is provided based on the utilization of services and not based on the staffing levels of the facility. In addition, these groups provided a care to pati ents with highly specific a nd costly conditions. MedPAC 2007 has expressed concern that the cu rrent Medicare reimbursement syst em does not cover the costs of care adequately to patients with needs over than therapy. Rehab focus nursing homes would be able to generate the greatest am ount of revenues since they had th e most Medicare rehab patients. This would explain why even though the rehab fo cus group had some of the highest costs, they still managed to be one of the best financial pe rformers. These results are providing some support with prior research that found that nursing homes with higher Medicare residents and private pay residents had the best financial performance (Zi nn et al. 1994, Marlin et al. 1999). However, in this study there were other groups that also perf ormed well. Differentiators managed to have good financial performance without a high degree of reha b. One explanation is th at by 2004, most states through Medicaid used case mix reimbursement that would compensate facilities that provided certain types of nursing technology (e.g. restorative care). Nursing homes that provide services that can generate revenue and lower costs via be tter quality can increas e their odds of achieving better financial performance. The hypothesis of this study stated that nursi ng homes that lacked a strategy would not perform as well as nursing homes that had a strategy. The results suggest that this is true with one

PAGE 95

95 exception. The specialty care focus group appeared to be a poor strategy with respect to financial performance. This outcome presents the possibility that not all strategies are created equal. The strategic groups rehab focus, cost leader and nursing care focus group (differentiator) all had better total margins than the other nur sing homes with no signifi cant difference between these strategies. One aspect of strategic group theory is that it is possible that there are multiple strategies within the same industry that are effective but not necessarily better than each other. This study supports this aspect of strategic group theory. Porters (1980) original theory described three effective strategies: cost leaders, differentiators and focus. This study provides support for these stra tegies even if the line between the differentiator and the focus strategy is blurre d. Porters generic stra tegies can help nursing homes identify strategies that allow them to achieve good financial performance. Managerial Implications Management is interested in performance. Providing managers with a way of identifying effective strategies that can im prove their facilities performance should be well received. Cool and Schendel (1987) defined strategic groups as sets of firms competing within an industry on the basis of similar combinations of scope and re source deployment commitments. Marlin et al. 2002 argued that for health care administrators, scop e decisions refer to the selection of market segments to compete in and the types of services to offer in these market segments. This study differed from prior research because it examin ed if the business level decision of technology commitment could be used to identify different st rategies and whether those strategies would have different performance outcomes. For managers that find themselves in facilities that would be considered to have poor or no strategy, this study provides some interesting findings. Strategic group theory posits that there exist mobility barriers that restrict a facility from moving from one strategic to another. The size of the investment needed for a nursing home to move from one strategic groups dictate the height

PAGE 96

96 of the barrier. This study partia lly supports that theory. Certainl y a nursing home would have to incur significant cost to enter the rehab focus gr oup thereby protecting this group with a relatively high mobility barrier. However, the mobility barrier to the differentiator may not be as high since these groups staffing levels, wages and cost are considerably low. Since it appears that nursing homes that provide higher levels of technology can have low costs and be profitable, managers with low performing strategy can improve their performance providing more care with the staff that they have. The study found that business level decisions of technology and resource utilization can be used to identify strategies th at lead to greater fi nancial performance and lower cost. The study identified cost leaders, rehab focus and diffe rentiators groups as ha ving significantly better financial performance than the RN focus strate gy and nursing homes that had no strategy. The rehab focus group had the highest proportion of Medicare beds and other payer beds. Although this supports prior studies that found that those facilities that had higher Medicare beds and private pay beds perform well fi nancially (Marlin et al. 1999 a nd Zinn et al. 1994), it found that this strategy was not significantly better than the differentiation or the cost leadership. Therefore, managers may be able to choose from one of thes e three alternatives and still have their facility perform well financially. One of the primary functions of managers is to control costs. With differentiators able to provide a high degree of care at the same costs as the cost leaders raises the question of whether there are cost savings in keeping staffing levels and staffing mixes lo w. Further research is needed to know how the differentiators were able to pr ovide all the services yet have similar costs. Policy Implications State and federal governments have a consid erable interest in nur sing homes since they provide over 60% of the funding. In return for thei r investment, they expect quality care for the residents. Therefore, it is in the governments best interest to know if there are some stra tegies that

PAGE 97

97 are better than others with resp ect to ensuring qualit y of care. This study has found some evidence that nursing homes with strategi es directed toward providing care have better quality of care than those nursing homes focused on cost savings. Th e study also found some evidence that nursing homes that provide a high level of nursing care technology to their long te rm care residents have better quality of care outcomes for that population. The study al so identified a group of nursing homes that provide care to patient with needs ot her than rehabilitation that may not be receiving adequate funding. The specialty care group provi des service to a population that is very vulnerable. Future policies should be directed to provide incentives that increase the amount of the important processes of care provided and not just address structural factors like staffing. Another factor is that is of interest to government agencies like CMS is that facilities that provided a significantly higher amount of therapy did not have bett er post acute outcomes with respect to walking improvement. At the same time, the strategic group that provided significantly less therapy than the other groups had worse walk ing improvement. From these results, it appears that there is a minimum amount of therapy required for adequate improvement, but there is limit beyond which more therapy does not lead to any mo re significant benefit. With the high costs of post acute care which is almost solely pa id for by the government, understanding how much therapy is sufficient is be very valuable. Further research is necessary to determine the minimum amount of therapy that is necessary. Similar to managers, policymakers are inte rested to know that a strategy of providing a high degree of service does not ne cessarily costs more. The fact that some facilities are able to provide a high level of service and yet keep th eir cost down can guide policymakers to provide incentives that focus on care prov ided as well as staffing. Further research is needed examining the relationship between processe s of care and the outcomes that they would be expected to influence.

PAGE 98

98 Conclusion The study indicates that strate gic groups exist in the nursi ng home industry that can be defined by Porters generic strategy. The study also demonstrates that strategic groups can be determined using the business leve l strategies of what technol ogy to provide and how to deploy resources. With respect to quality, nursing hom es that focused on providing technology (rehab or nursing) or had higher RN staffing levels had less quality deficien cies than facilities that did not have a strategy at all. Only the strategic group that provided a broad array of nursing technology had better quality in the area that the technology was focused. Th e study found that nursing homes that provide a high level of nursing technology do not necessarily have higher costs. The study further demonstrate that Porters cost leader, nursing care focus and rehab focus strategies have better financial performance than those facilities that have poor strategy or do not have a strategy at all. Limitations There are a number of limitations to this study. First, the results of the study apply only to free standing, non government faciliti es that have Medicare cost re port and are locat ed in a county that has at least 2 nursing homes. Facilities that only had private pay residents or facilities that only had Medicaid residents were not included in the sample. This limits the generalizability to the 9700 nursing homes that have all these charac teristics. Another limitation includes the possible e ndogeneity between pressure ulcer quality measures and the technology variable for skin technology. The skin technology variable includes both preventative and treatment processes. As a resu lt, especially in the case of the long term care skin technology variable, nursing ho mes with a high prevalence of pr essure ulcers are forced to provide treatments. This fact increases the li kelihood of an endogenous relationship between the quality variable and the strategic groups that pr ovide a high level of long term care skin processes and may explain the why the rehab focus group had the highest preval ence and incidence of

PAGE 99

99 pressure sores and also provided the most skin care. Another con cern about the post acute pressure ulcer incidence variable was that it did not have a normal distribution and an adequate transformation was not achieved. When the GLM an alysis was attempted, the analysis could not converge. Therefore the results of this measure may not be reliable. Another limitation of the study was the magnitude of the impact of some of the dependent variables. 4 point ADL decline a nd post acute improvement in walking had very small regression coefficients relative to the cons tant. Although these variables presen ted statistical significance, the small size of the effect may limit the relevance. The study used cross-sectional data and therefore caution shou ld be taken when trying to infer a causal relationship. Finally, the datase ts included in the study are administrative data. Although the nursing homes can be held accountable for the accuracy of the data, some inconsistencies can occur dependi ng on the individual who is enteri ng the data. For example, the MDS is entered by at least one nurse in each of the 16000 nursing homes in the United States.

PAGE 100

100 APPENDIX BACKGROUND TABLES Table A-1. Eigenvalues of scope (technology) factor analysis Eigenvalue Difference Proportion Cumulative 1 7.03 2.134 0.126 0.126 2 4.89 1.138 0.087 0.213 3 3.76 1.074 0.067 0.280 4 2.68 0.391 0.048 0.328 5 2.29 0.080 0.041 0.369 6 2.21 0.416 0.040 0.408 7 1.80 0.017 0.032 0.440 8 1.78 0.241 0.032 0.472 9 1.54 0.031 0.028 0.500 10 1.51 0.043 0.027 0.527 11 1.46 0.098 0.026 0.553 12 1.37 0.024 0.024 0.577 13 1.34 0.067 0.024 0.601 14 1.28 0.061 0.023 0.624 15 1.21 0.048 0.022 0.646 16 1.17 0.047 0.021 0.666 17 1.12 0.057 0.020 0.686 18 1.06 0.022 0.019 0.705 19 1.04 0.027 0.019 0.724 20 1.01 0.064 0.018 0.742 21 0.95 0.101 0.017 0.759 22 0.85 0.037 0.015 0.774 23 0.81 0.044 0.015 0.789 24 0.77 0.066 0.014 0.802 25 0.70 0.014 0.013 0.815 26 0.69 0.053 0.012 0.827 27 0.63 0.009 0.011 0.839 28 0.63 0.006 0.011 0.850 29 0.62 0.022 0.011 0.861 30 0.60 0.060 0.011 0.871 31 0.54 0.033 0.010 0.881 32 0.50 0.032 0.009 0.890 33 0.47 0.038 0.008 0.898 34 0.43 0.008 0.008 0.906 35 0.43 0.021 0.008 0.914 36 0.41 0.021 0.007 0.921 37 0.39 0.006 0.007 0.928 38 0.38 0.024 0.007 0.935 39 0.36 0.026 0.006 0.941 40 0.33 0.009 0.006 0.947 41 0.32 0.018 0.006 0.953 42 0.30 0.050 0.005 0.958 43 0.25 0.017 0.005 0.963 44 0.24 0.023 0.004 0.967 45 0.21 0.011 0.004 0.971 46 0.20 0.011 0.004 0.974 47 0.19 0.008 0.003 0.978 48 0.18 0.005 0.003 0.981 49 0.18 0.005 0.003 0.984 50 0.17 0.008 0.003 0.987

PAGE 101

101 Scree plot1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 1920 21 22 23 24 25 26 27 2829 30 31 32 33 3435 36 37 38 39 404142 43 44 45464748 49505152 53 54 5556 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 01 02 03 04 05 06 0 Series1 Figure A-1. Scree plot of scope (technology) factor analysis

PAGE 102

Table A-2. Factor analysis with 10 f actors of scope (technology) variables Rotated Factor Pattern Factor1 Factor2 Factor3 Factor4 Factor5 Factor6 Factor7 Factor 8 Factor9 Factor10 Long term care plateguard -0.01 0.13 0.07 0.02 -0.04 0.01 0.05 0.02 0.00 0.76 Long term care Weight program -0.01 -0.01 0.04 0.03 0.02 -0.03 0.03 0.02 0.92 0.03 Long term care Therapy Diet 0.01 0.17 0.07 -0.03 0.03 0.05 0.13 -0.05 -0.02 0.70 Long term care Bladder training -0.03 0.07 0.08 -0.11 0.02 0.05 0.12 0.30 -0.05 -0.23 Long term care toilet program 0.07 -0.05 0.02 0.20 -0.06 -0.02 -0.06 0.79 0.05 0.22 Long term care no skin treatment 0.00 0.09 0.00 -0.71 -0.24 -0.05 0.00 -0.11 -0.01 -0.06 Long term care other skin tx -0.03 0.63 0.14 0.24 0.05 -0.05 0.01 0.08 0.00 0.22 Long term care ointment -0.06 0.70 0.18 0.02 0.13 0.02 0.06 -0.02 -0.02 0.16 Long term care pressure relief bed -0.05 0.57 0.18 0.59 -0.19 0.04 0.02 -0.11 -0.05 0.04 Long term care pressure relief chair 0.00 0.61 0.12 0.51 -0.20 -0.05 0.06 0.00 -0.02 0.08 Long term care dressing application -0.06 0.77 0.11 -0.07 0.16 0.13 0.05 -0.18 -0.01 0.06 Long term care surgical wound care 0.06 0.40 -0.09 -0.07 0.08 0.35 0.06 -0.07 0.01 0.01 Long term care ulcer care -0.03 0.72 0.07 -0.06 0.15 0.11 0.03 -0.20 0.00 0.01 Long term care skin nutrition program 0.02 0.60 0.08 0.04 0.21 -0.09 -0.08 0.12 0.09 -0.06 Long term care turning repositioning -0.02 0.69 0.14 0.18 0.03 0.04 0.00 0.19 -0.03 -0.03 Long term care OT minutes 0.03 0.34 -0.11 -0.05 -0.11 0.75 0.05 -0.09 0.10 -0.02 Long term care PT minutes 0.03 0.38 -0.10 -0.05 -0.10 0.74 0.04 -0.09 0.09 0.01 Long term care other restorative care -0.07 0.15 0.47 0.01 -0.04 0.02 0.21 0.08 -0.02 -0.17 Long term care restorative communication 0.17 0.01 0.39 0.04 0.04 -0.01 -0.13 -0.07 0.01 0.02 Long term care restorative amputation -0.04 -0.03 0.38 0.05 -0.03 -0.02 0.05 0.01 0.00 -0.19 Long term care restorative eating 0.17 0.12 0.72 0.00 0.01 0.06 0.05 -0.01 0.03 0.08 Long term care restorative dressing 0.30 0.11 0.74 -0.02 0.01 0.02 0.03 0.03 0.03 0.13 Long term care restorative transfers 0.31 0.14 0.73 0.00 0.00 0.01 0.06 0.01 0.00 0.09 Long term care restorative bed mobility 0.39 0.06 0.64 0.00 0.03 0.00 -0.03 0.01 0.03 -0.01 Long term care restorative splinting 0.02 -0.03 0.18 0.05 0.23 0.14 0.61 -0.03 0.00 -0.03 Long term care restorative Passive ROM 0.12 0.31 0.42 0.03 -0.03 -0.03 0.59 0.03 0.00 0.17 Long term care restorative Active ROM 0.10 0.28 0.47 0.01 -0.09 0.01 0.51 0.05 -0.01 0.17 Long term care restorative ambulation 0.05 0.33 0.62 0.02 -0.07 0.03 0.35 0.07 -0.02 0.22 Post acute plateguard 0.03 0.02 -0.01 0.06 -0.03 -0.14 0.12 0.12 0.10 0.24 Post acute Weight program -0.02 0.01 0.03 0.03 0.05 -0.03 0.03 0.01 0.93 0.02 Post acute Therapy Diet -0.01 -0.04 -0.04 0.04 0.10 -0.15 0.29 0.02 0.05 0.10 Post acute Bladder training -0.01 -0.04 0.06 -0.05 0.14 0.05 0.15 0.39 -0.07 -0.20 Post acute toilet program 0.11 -0.06 -0.05 0.18 -0.03 -0.07 -0.03 0.80 0.09 0.12 102

PAGE 103

Table A-2. Continued Post acute no skin treatment 0.01 -0.08 -0.04 -0.79 -0.31 -0.03 -0.02 -0.09 -0.02 -0.08 Post acute other skin tx 0.03 0.19 -0.01 0.32 0.31 -0.06 0.00 0.20 0.05 0.15 Post acute ointment -0.03 0.16 0.02 0.04 0.66 -0.02 0.12 0.03 0.02 0.00 Post acute pressure relief bed 0.01 0.04 0.03 0.84 0.01 0.05 0.03 -0.04 -0.01 -0.06 Post acute pressure relief chair 0.10 0.14 -0.04 0.65 -0.06 -0.09 0.06 0.13 0.06 -0.02 Post acute dressing application -0.04 0.07 0.01 0.03 0.77 0.06 0.07 -0.07 -0.03 0.09 Post acute surgical wound care 0.01 -0.06 0.00 0.12 0.42 0.15 -0.07 0.08 -0.12 0.27 Post acute ulcer care 0.00 0.07 -0.05 0.01 0.71 -0.03 0.15 -0.13 0.05 -0.13 Post acute skin nutrition program 0.06 0.27 -0.03 0.09 0.52 -0.13 0.00 0.17 0.14 -0.16 Post acute turning repositioning 0.07 0.07 -0.06 0.26 0.39 0.01 0.02 0.35 0.02 -0.18 Post acute OT minutes -0.14 -0.13 0.14 0.06 0.05 0.69 -0.10 0.06 -0.10 -0.02 Post acute PT minutes -0.10 -0.19 0.14 0.08 0.10 0.67 -0.08 0.11 -0.12 0.02 Post acute other restorative care 0.23 0.04 0.11 -0.02 -0.03 -0.02 0.18 0.07 -0.01 -0.19 Post acute restorative communication 0.49 -0.03 0.14 0.01 0.02 -0.01 -0.12 -0.07 0.00 -0.01 Post acute restorative amputation 0.31 -0.02 -0.05 0.03 0.04 0.01 0.12 -0.01 -0.01 0.00 Post acute restorative eating 0.65 -0.03 0.29 0.03 0.05 0.00 -0.01 0.02 0.01 0.02 Post acute restorative dressing 0.86 -0.01 0.26 0.00 -0.03 -0.01 -0.01 0.04 0.01 0.04 Post acute restorative transfers 0.90 -0.02 0.19 0.01 -0.03 -0.03 0.02 0.05 -0.01 0.02 Post acute restorative bed mobility 0.88 -0.03 0.15 0.01 0.00 -0.01 0.00 0.03 0.00 0.00 Post acute restorative splinting 0.30 -0.04 0.01 0.03 0.09 -0.03 0.52 0.05 0.01 0.01 Post acute restorative Passive ROM 0.58 -0.02 -0.04 0.00 0.03 -0.10 0.54 0.02 0.02 -0.03 Post acute restorative Active ROM 0.65 0.01 0.01 0.01 -0.07 -0.08 0.46 0.05 0.01 -0.03 Post acute restorative ambulation 0.79 0.02 0.10 0.01 -0.08 -0.08 0.15 0.10 -0.01 0.05 103 Values over 0.4 are highlighted and represent the initial variables used to define the factor

PAGE 104

Table A-3. Factor analysis with 10 f actors of scope (technology) variables Rotated Factor Pattern Post acute restorative LTC skin care LTC restorative Pressure devices Post acute skin care Therapy minutes Toilet program Weight program Long term care plateguard -0.01 0.13 0.07 0.02 -0.04 0.01 0.02 0.00 Long term care Weight program -0.01 -0.01 0.04 0.03 0.02 -0.03 0.02 0.92 Long term care Therapy Diet 0.01 0.17 0.07 -0.03 0.03 0.05 -0.05 -0.02 Long term care Bladder training -0.03 0.07 0.08 -0.11 0.02 0.05 0.30 -0.05 Long term care toilet program 0.07 -0.05 0.02 0.20 -0.06 -0.02 0.79 0.05 Long term care no skin treatment 0.00 0.09 0.00 -0.71 -0.24 -0.05 -0.11 -0.01 Long term care other skin tx -0.03 0.63 0.14 0.24 0.05 -0.05 0.08 0.00 Long term care ointment -0.06 0.70 0.18 0.02 0.13 0.02 -0.02 -0.02 Long term care pressure relief bed -0.05 0.57 0.18 0.59 -0.19 0.04 -0.11 -0.05 Long term care pressure relief chair 0.00 0.61 0.12 0.51 -0.20 -0.05 0.00 -0.02 Long term care dressing application -0.06 0.77 0.11 -0.07 0.16 0.13 -0.18 -0.01 Long term care surgical wound care 0.06 0.40 -0.09 -0.07 0.08 0.35 -0.07 0.01 Long term care ulcer care -0.03 0.72 0.07 -0.06 0.15 0.11 -0.20 0.00 Long term care skin nutrition program 0.02 0.60 0.08 0.04 0.21 -0.09 0.12 0.09 Long term care turning repositioning -0.02 0.69 0.14 0.18 0.03 0.04 0.19 -0.03 Long term care OT minutes 0.03 0.34 -0.11 -0.05 -0.11 0.75 -0.09 0.10 Long term care PT minutes 0.03 0.38 -0.10 -0.05 -0.10 0.74 -0.09 0.09 Long term care other restorative care -0.07 0.15 0.47 0.01 -0.04 0.02 0.08 -0.02 Long term care restorative communication 0.17 0.01 0.39 0.04 0.04 -0.01 -0.07 0.01 Long term care restorative amputation -0.04 -0.03 0.38 0.05 -0.03 -0.02 0.01 0.00 Long term care restorative eating 0.17 0.12 0.72 0.00 0.01 0.06 -0.01 0.03 Long term care restorative dressing 0.30 0.11 0.74 -0.02 0.01 0.02 0.03 0.03 Long term care restorative transfers 0.31 0.14 0.73 0.00 0.00 0.01 0.01 0.00 Long term care restorative bed mobility 0.39 0.06 0.64 0.00 0.03 0.00 0.01 0.03 Long term care restorative splinting 0.02 -0.03 0.18 0.05 0.23 0.14 -0.03 0.00 Long term care restorative Passive ROM 0.12 0.31 0.42 0.03 -0.03 -0.03 0.03 0.00 Long term care restorative Active ROM 0.10 0.28 0.47 0.01 -0.09 0.01 0.05 -0.01 Long term care restorative ambulation 0.05 0.33 0.62 0.02 -0.07 0.03 0.07 -0.02 Post acute plateguard 0.03 0.02 -0.01 0.06 -0.03 -0.14 0.12 0.10 Post acute Weight program -0.02 0.01 0.03 0.03 0.05 -0.03 0.01 0.93 Post acute Therapy Diet -0.01 -0.04 -0.04 0.04 0.10 -0.15 0.02 0.05 Post acute Bladder training -0.01 -0.04 0.06 -0.05 0.14 0.05 0.39 -0.07 Post acute toilet program 0.11 -0.06 -0.05 0.18 -0.03 -0.07 0.80 0.09 104

PAGE 105

105 Post acute restorative LTC skin care LTC restorative Pressure devices Post acute skin care Therapy minutes Toilet program Weight program Table A-3. Continued Post acute no skin treatment 0.01 -0.08 -0.04 -0.79 -0.31 -0.03 -0.09 -0.02 Post acute other skin tx 0.03 0.19 -0.01 0.32 0.31 -0.06 0.20 0.05 Post acute ointment -0.03 0.16 0.02 0.04 0.66 -0.02 0.03 0.02 Post acute pressure relief bed 0.01 0.04 0.03 0.84 0.01 0.05 -0.04 -0.01 Post acute pressure relief chair 0.10 0.14 -0.04 0.65 -0.06 -0.09 0.13 0.06 Post acute dressing application -0.04 0.07 0.01 0.03 0.77 0.06 -0.07 -0.03 Post acute surgical wound care 0.01 -0.06 0.00 0.12 0.42 0.15 0.08 -0.12 Post acute ulcer care 0.00 0.07 -0.05 0.01 0.71 -0.03 -0.13 0.05 Post acute skin nutrition program 0.06 0.27 -0.03 0.09 0.52 -0.13 0.17 0.14 Post acute turning repositioning 0.07 0.07 -0.06 0.26 0.39 0.01 0.35 0.02 Post acute OT minutes -0.14 -0.13 0.14 0.06 0.05 0.69 0.06 -0.10 Post acute PT minutes -0.10 -0.19 0.14 0.08 0.10 0.67 0.11 -0.12 Post acute other restorative care 0.23 0.04 0.11 -0.02 -0.03 -0.02 0.07 -0.01 Post acute restorative communication 0.49 -0.03 0.14 0.01 0.02 -0.01 -0.07 0.00 Post acute restorative amputation 0.31 -0.02 -0.05 0.03 0.04 0.01 -0.01 -0.01 Post acute restorative eating 0.65 -0.03 0.29 0.03 0.05 0.00 0.02 0.01 Post acute restorative dressing 0.86 -0.01 0.26 0.00 -0.03 -0.01 0.04 0.01 Post acute restorative transfers 0.90 -0.02 0.19 0.01 -0.03 -0.03 0.05 -0.01 Post acute restorative bed mobility 0.88 -0.03 0.15 0.01 0.00 -0.01 0.03 0.00 Post acute restorative splinting 0.30 -0.04 0.01 0.03 0.09 -0.03 0.05 0.01 Post acute restorative Passive ROM 0.58 -0.02 -0.04 0.00 0.03 -0.10 0.02 0.02 Post acute restorative Active ROM 0.65 0.01 0.01 0.01 -0.07 -0.08 0.05 0.01 Post acute restorative ambulation 0.79 0.02 0.10 0.01 -0.08 -0.08 0.10 -0.01 Values over 0.4 are highlighted and represent the initial variables used to define the fact

PAGE 106

Table A-4. Eigenvalues of scope (technology) factor analysis Eigenvalues of the Correlation Matrix: Eigenvalue Difference Proportion Cumulative 1 3.16 1.77 0.45 0.45 2 1.39 0.32 0.20 0.65 3 1.08 0.43 0.15 0.80 4 0.65 0.30 0.09 0.90 5 0.35 0.10 0.05 0.95 6 0.24 0.11 0.03 0.98 7 0.13 0.02 1 1 2 3 4 5 6 7 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 012345678 Series1 Figure A-2. Scree plot of resource sta ffing intensity from factor analysis Table A-5. Factor analysis of resource staffing intensity with three factors Total therapist Total therapy assistants Total nursing Total PT per Resident 0.94987 0.13329 0.09479 Total OT per Resident 0.93244 0.10457 0.13311 Total PTA per Resident -0.02059 0.92192 0.05157 Total OTA per Resident 0.39775 0.78854 0.21303 RN per Resident 0.0587 0.03395 0.79055 LPN per Resident 0.15186 0.56287 0.61385 CNA per Resident 0.13473 0.15559 0.86025 Values over 0.4 are highlighted and represent the initial variables used to define the factor 106

PAGE 107

107 Table A-6. Ward's minimum variance cluster analysis The CLUSTER Procedure Ward's Minimum Variance Cluster Analysis Eigenvalues of the Covariance Matrix Eigenvalue Difference Proportion Cumulative 1 6.24393174 0.63097545 0.2280 0.2280 2 5.61295630 1.53129364 0.2050 0.4330 3 4.08166266 0.58262014 0.1491 0.5821 4 3.49904251 1.67386173 0.1278 0.7099 5 1.82518078 0.40031900 0.0667 0.7766 6 1.42486178 0.27397392 0.0520 0.8286 7 1.15088786 0.14931403 0.0420 0.8706 8 1.00157383 0.45163154 0.0366 0.9072 9 0.54994229 0.12809678 0.0201 0.9273 10 0.42184551 0.01558279 0.0154 0.9427 11 0.40626272 0.06700601 0.0148 0.9575 12 0.33925671 0.10344911 0.0124 0.9699 13 0.23580760 0.08800114 0.0086 0.9785 14 0.14780646 0.02048422 0.0054 0.9839 15 0.12732224 0.02697285 0.0047 0.9886 16 0.10034939 0.01066020 0.0037 0.9923 17 0.08968920 0.02449037 0.0033 0.9955 18 0.06519883 0.03495529 0.0024 0.9979 19 0.03024354 0.01013037 0.0011 0.9990 20 0.02011316 0.01323966 0.0007 0.9997 21 0.00687351 0.0003 1.0000 Root-Mean-Square Total-Sample Standard Deviation = 1.141862 Root-Mean-Square Distance Between Observations = 7.400109 Table A-7. Cluster History NCL ------Clusters Joined-----FREQ SPRSQ RSQ ERSQ CCC PSF PST2 e 10 CL22 CL29 3 0.0250 .645 .605 2.70 13.1 4.3 9 CL12 CL13 37 0.0258 .619 .583 2.42 13.4 5.4 8 CL15 CL31 3 0.0286 .591 .558 2.13 13.8 5.2 7 CL17 CL11 7 0.0330 .558 .530 1.79 14.3 3.8 6 CL9 CL15 57 0.0466 .511 .497 0.78 14.4 9.1 5 CL22 CL23 5 0.0550 .456 .458 -.12 14.7 8.4 4 CL6 CL7 64 0.0887 .367 .400 -1.7 13.7 13.6 3 CL4 CL10 67 0.0942 .273 .320 -2.3 13.5 11.7 2 CL3 CL8 70 0.1260 .147 .188 -2.2 12.6 13.1 1 CL2 CL5 75 0.1472 .000 .000 0.00 12.

PAGE 108

Table A-8. Determination of rank scores table RN focus n=1634 Rehab Focus n=1816 Differentiator n=990 Cost leader n=2752 Lack of strategy n=2554 V LV TV R V LV TV R V LV TV R V LV TV R V LV TV R Zwages 0.15 -0.10 0.25 5.0 0.10 -0.10 0.20 4.0 0.00 -0.10 0.10 2.0 -0.06 -0.10 0.04 0.8 -0.10 -0.10 0.00 0.0 ZRN per res 1.16 -0.56 1.72 5.0 0.38 -0.56 0.94 2.7 0.02 -0.56 0.58 1.7 -0.56 -0.56 0.00 0.0 -0.42 -0.56 0.14 0.4 ZLPN per res -0.26 -0.26 0.00 0.0 0.57 -0.26 0.83 5.0 -0.10 -0.26 0.16 1.0 -0.08 -0.26 0.18 1.1 -0.12 -0.26 0.14 0.9 ZCNA per res -1.02 -1.02 0.00 0.0 0.46 -1.02 1.48 5.0 0.27 -1.02 1.29 4.4 0.04 -1.02 1.06 3.6 0.18 -1.02 1.20 4.0 Ztherapy -0.07 -0.81 0.74 1.8 1.31 -0.81 2.12 5.0 -0.39 -0.81 0.42 1.0 -0.81 -0.81 0.00 0.0 0.14 -0.81 0.95 2.2 Ztherapy assist -0.19 -0.91 0.72 1.6 1.33 -0.91 2.24 5.0 -0.38 -0.91 0.53 1.2 0.20 -0.91 1.11 2.5 -0.91 -0.91 0.00 0.0 ZPT percent 0.06 -0.79 0.85 2.4 -0.24 -0.79 0.55 1.6 0.10 -0.79 0.89 2.5 -0.79 -0.79 0.00 0.0 0.95 -0.79 1.74 5.0 ZOT percent 0.07 -0.76 0.83 2.6 -0.14 -0.76 0.62 1.9 0.07 -0.76 0.83 2.6 -0.76 -0.76 0.00 0.0 0.84 -0.76 1.60 5.0 ZRN percent 1.67 -0.58 2.25 5.0 0.06 -0.58 0.64 1.4 -0.06 -0.58 0.52 1.2 -0.58 -0.58 0.00 0.0 -0.47 -0.58 0.11 0.3 ZLicense perc. 1.47 -0.48 1.95 5.0 0.21 -0.48 0.69 1.8 -0.32 -0.48 0.16 0.4 -0.45 -0.48 0.03 0.1 -0.48 -0.48 0.00 0.0 Total Resource 28 34 18 8 18 zltcskinpr~s 0.05 -0.31 0.36 2.1 0.56 -0.31 0.87 5.0 0.25 -0.31 0.56 3.2 -0.20 -0.31 0.11 0.6 -0.31 -0.31 0.00 0.0 Zrestorative 14d -0.21 -0.43 0.22 0.3 -0.43 -0.43 0.00 0.0 2.93 -0.43 3.36 5.0 -0.34 -0.43 0.09 0.1 -0.34 -0.43 0.09 0.1 Zrestorative Q -0.17 -0.31 0.14 0.3 0.05 -0.31 0.36 0.9 1.73 -0.31 2.04 5.0 -0.27 -0.31 0.04 0.1 -0.31 -0.31 0.00 0.0 Zrelief device -0.05 -0.17 0.12 1.5 0.24 -0.17 0.41 5.0 0.17 -0.17 0.34 4.1 -0.03 -0.17 0.14 1.7 -0.17 -0.17 0.00 0.0 Ztherapy min -0.08 -0.57 0.49 1.7 0.88 -0.57 1.45 5.0 -0.57 -0.57 0.00 0.0 -0.11 -0.57 0.46 1.6 -0.24 -0.57 0.33 1.1 Zacute skin care 0.07 -0.23 0.30 3.0 0.26 -0.23 0.49 5.0 0.23 -0.23 0.46 4.7 -0.08 -0.23 0.15 1.6 -0.23 -0.23 0.00 0.0 Ztoilet program -0.06 -0.07 0.01 0.1 -0.06 -0.07 0.01 0.1 0.29 -0.07 0.36 5.0 0.04 -0.07 0.11 1.5 -0.07 -0.07 0.00 0.0 Zweight prog. 0.01 -0.10 0.11 2.5 -0.10 -0.10 0.00 0.0 0.12 -0.10 0.22 5.0 0.06 -0.10 0.16 3.7 -0.05 -0.10 0.05 1.0 Zpercent Mcare -0.13 -0.39 0.26 0.7 1.37 -0.39 1.76 5.0 -0.36 -0.39 0.03 0.0 -0.34 -0.39 0.05 0.2 -0.39 -0.39 0.00 0.1 Total scope 12 26 32 11 2.3108 V is the actual Z score value LV is the lowest value for the variable among clusters TV is the total variance is the difference between the highest value among the clusters and the lowest values among clusters R is the final rank from 0 to 5. Determined ((V-LV)/TV*5)

PAGE 109

Table A-9. Descriptives of dependent variables, independent variables and control variables Cluster 1 Differentiator N=990 Cluster 2 Cost Leader N=2752 Cluster 3 Rehab focus N=1816 Cluster 4 Specialty care N=1634 Cluster 5 Lack of strategy N=2534 Avg Wages 12.61 12.52 12.99 12.76 12.45 RN per Resident 0.06 0.04 0.07 0.08 0.05 LPN per Resident 0.14 0.14 0.17 0.14 0.14 CNA per Resident 0.44 0.42 0.47 0.31 0.44 Therapist per Res. 0.02 0.01 0.04 0.02 0.02 Therapist asst per Res. 0.01 0.02 0.03 0.01 0.01 Percent Licensed PT 0.62 0.39 0.54 0.6 0.84 Percent Licensed OT 0.65 0.41 0.6 0.64 0.86 Percent RN 0.09 0.07 0.1 0.17 0.08 LTC Skin Process 0.83 0.72 0.87 0.78 0.69 Acute Skin Process 0.91 0.83 0.91 0.86 0.81 Relief device 1.95 1.8 1.98 1.79 1.72 LTC Restorative Process 4.28 1.8 2.03 1.88 1.67 Acute Restorative Process 3.69 0.47 0.39 0.58 0.44 Therapy minutes 324.7 357.8 439.6 357.3 349.6 Toilet Program 0.68 0.59 0.58 0.57 0.57 Weight Program 0.34 0.32 0.28 0.31 0.29 Percent Medicare 10.7 10.82 22.2 11.88 10.47 Percent Other 23.39 21.7 26.56 24.68 22.65 Percent Medicaid 65.91 67.48 51.24 63.45 66.88 Acuindex 10.39 10.2 10.45 10.25 10.17 ADL index 10.13 10.02 10.23 10.03 9.98 Percent tube feed 0.07 0.06 0.06 0.06 0.06 Percent tracheotomy 0.01 0.01 0.01 0.01 0 Percent requiring suctioning 0.01 0.01 0.01 0.01 0.01 Percent requiring respiratory care 0.11 0.1 0.12 0.11 0.1 Percent requiring IV 0.01 0.01 0.02 0.01 0.01 Percent requiring injection 0.15 0.15 0.17 0.16 0.15 Quality care deficiency 1.782 1.928 1.749 1.777 1.929 Bladder decline 0.186 0.193 0.209 0.192 0.187 Bowel decline 0.169 0.176 0.188 0.176 0.168 ADL 4 point decline 0.165 0.172 0.17 0.169 0.17 Restraint 0.076 0.079 0.075 0.076 0.077 Long term care pressure ulcer 0.08 0.076 0.093 0.081 0.075 Post acute walking improvement 0.232 0.238 0.246 0.24 0.246 Post acute pressure ulcer 0.187 0.173 0.177 0.174 0.169 Ownership 0.71 0.76 0.8 0.77 0.76 Market Competition 0.13 0.14 0.14 0.16 0.14 Chain 0.52 0.6 0.66 0.63 0.58 Total beds 114 130 116 116 111 109

PAGE 110

Table A-10. Negative binomial of quali ty of care deficiencies with di fferentiators as reference group IRR Std. Err. z P>z Cost leaders 1.059 0.036 1.72 0.09 Rehab focus 0.993 0.038 -0.18 0.86 Specialty care focus 1.001 0.037 0.04 0.97 Lack of strategy 1.124 0.038 3.43 0.00 Ownership 1.099 0.029 3.59 0.00 Percent Medicaid 1.005 0.001 7.69 0.00 Percent Medicare 1.005 0.001 3.75 0.00 Market competition 0.773 0.367 -0.54 0.59 Occupancy 0.830 0.065 -2.38 0.02 Total beds 1.003 0.000 14.74 0.00 Chain 1.024 0.021 1.18 0.24 Acuity index 1.040 0.013 3.20 0.00 ADL index 0.952 0.029 -1.63 0.10 Tube feed care 1.009 0.237 0.04 0.97 Respiratory care 0.972 0.157 -0.18 0.86 Suctioning care 0.811 0.454 -0.37 0.71 IV therapy care 3.826 1.304 3.94 0.00 Tracheostomy care 1.142 0.745 0.20 0.84 Need injection 1.189 0.141 1.46 0.14 _Ifips_2 1.206 0.470 0.48 0.63 Table A-11. OLS regressions of root transformed bladder decline with differentiators as ref. group Coef. Std. Err. t P>t Cost leaders 0.006 0.003 2.03 0.04 Rehab focus 0.012 0.004 3.27 0.00 Specialty care focus 0.009 0.004 2.68 0.01 Lack of strategy 0.005 0.003 1.56 0.12 Ownership 0.006 0.002 2.46 0.01 Percent Medicaid 0.000 0.000 -5.35 0.00 Percent Medicare 0.001 0.000 3.84 0.00 Market competition 0.077 0.064 1.21 0.23 Occupancy 0.014 0.008 1.70 0.09 Total beds 0.000 0.000 3.66 0.00 Chain 0.004 0.002 1.97 0.05 Acuity index 0.009 0.001 6.66 0.00 ADL index 0.021 0.003 6.68 0.00 Tube feed care -0.061 0.027 -2.24 0.03 Respiratory care -0.040 0.014 -2.99 0.00 Suctioning care 0.097 0.067 1.44 0.15 IV therapy care 0.004 0.040 0.09 0.93 Tracheostomy care -0.245 0.074 -3.32 0.00 Need injection??? 0.004 0.012 0.33 0.74 _cons 0.241 0.017 13.91 0.00 110

PAGE 111

Table A-12. OLS regressions of root transformed bowel decline with differentiators as ref. group Coef. Std. Err. t P>t Cost leaders 0.008 0.004 2.31 0.02 Rehab focus 0.009 0.004 2.22 0.03 Specialty care focus 0.012 0.004 2.92 0.00 Lack of strategy 0.005 0.004 1.41 0.16 Ownership 0.005 0.003 1.83 0.07 Percent Medicaid 0.000 0.000 -3.76 0.00 Percent Medicare 0.001 0.000 3.82 0.00 Market competition 0.040 0.050 0.79 0.43 Occupancy 0.013 0.009 1.49 0.14 Total beds 0.000 0.000 2.94 0.00 Chain 0.004 0.002 1.79 0.07 Acuity index 0.011 0.001 7.26 0.00 ADL index 0.025 0.003 7.43 0.00 Tube feed care -0.039 0.029 -1.34 0.18 Respiratory care -0.040 0.015 -2.74 0.01 Suctioning care 0.014 0.077 0.18 0.86 IV therapy care 0.016 0.044 0.36 0.72 Tracheostomy care -0.101 0.085 -1.19 0.23 Need injection 0.020 0.013 1.55 0.12 _cons 0.176 0.017 10.11 0.00 Table A-13. OLS regressions of r oot transformed ADL decline with differentiators as ref. group Coef. Std. Err. t P>t Cost leaders 0.008 0.003 2.94 0.00 Rehab focus 0.003 0.003 0.86 0.39 Specialty care focus 0.005 0.003 1.67 0.10 Lack of strategy 0.004 0.003 1.50 0.13 Ownership 0.001 0.002 0.50 0.62 Percent Medicaid 0.000 0.000 0.10 0.92 Percent Medicare 0.000 0.000 2.97 0.00 Market competition 0.083 0.053 1.57 0.12 Occupancy 0.006 0.007 0.84 0.40 Total beds 0.000 0.000 -0.71 0.48 Chain -0.001 0.002 -0.72 0.47 Acuity index 0.002 0.001 1.46 0.14 ADL index -0.003 0.003 -1.01 0.31 Tube feed care -0.028 0.026 -1.10 0.27 Respiratory care -0.005 0.013 -0.42 0.67 Suctioning care -0.054 0.057 -0.95 0.34 IV therapy care -0.013 0.033 -0.39 0.70 Tracheostomy care -0.072 0.066 -1.10 0.27 Need injection -0.008 0.011 -0.69 0.49 _cons 0.374 0.016 23.87 0.00 111

PAGE 112

Table A-14. OLS regressions of l og transformed restraints with differentiators as ref. group Coef. Std. Err. t P>t Cost leaders 0.011 0.004 2.48 0.01 Rehab focus 0.004 0.006 0.74 0.46 Specialty care focus 0.006 0.005 1.32 0.19 Lack of strategy 0.006 0.005 1.24 0.22 Ownership 0.020 0.004 5.50 0.00 Percent Medicaid 0.000 0.000 4.06 0.00 Percent Medicare 0.000 0.000 1.16 0.25 Market competition 0.053 0.070 0.76 0.45 Cost leaders -0.010 0.012 -0.84 0.40 Total beds 0.000 0.000 7.90 0.00 Chain -0.007 0.003 -2.24 0.03 Acuity index 0.011 0.002 6.07 0.00 ADL index 0.016 0.004 3.80 0.00 Tube feed care -0.049 0.041 -1.20 0.23 Respiratory care -0.012 0.022 -0.54 0.59 Suctioning care -0.138 0.101 -1.37 0.17 IV therapy care 0.028 0.054 0.51 0.61 Tracheostomy care 0.060 0.110 0.54 0.59 Need injection -0.055 0.018 -3.12 0.00 _cons -0.004 0.023 -0.19 0.85 Table A-15. OLS regressions of r oot transformed LTC ulcers with differentiators as ref. group Coef. Std. Err. t P>t Cost leaders -0.001 0.002 -0.45 0.66 Rehab focus 0.012 0.003 4.19 0.00 Specialty care focus 0.004 0.003 1.47 0.14 Lack of strategy -0.004 0.002 -1.71 0.09 Ownership 0.006 0.002 3.32 0.00 Percent Medicaid 0.000 0.000 -0.71 0.48 Percent Medicare 0.001 0.000 6.94 0.00 Market competition -0.001 0.029 -0.04 0.97 Occupancy -0.035 0.006 -5.55 0.00 Total beds 0.000 0.000 3.31 0.00 Chain -0.002 0.002 -1.57 0.12 Acuity index 0.008 0.001 8.42 0.00 ADL index 0.007 0.002 3.11 0.00 Tube feed care 0.109 0.024 4.6 0.00 Respiratory care 0.031 0.011 2.74 0.01 Suctioning care 0.045 0.060 0.74 0.46 IV therapy care 0.150 0.032 4.7 0.00 Tracheostomy care -0.031 0.065 -0.47 0.64 Need injection 0.046 0.010 4.74 0.00 _cons 0.151 0.012 12.84 0.00 112

PAGE 113

Table A-16. GLM with gamma distribu tion with PA walking improve with differentiators as ref. group Coef. Std. Err. z P>z Cost leaders 0.031 0.019 1.65 0.10 Rehab focus 0.050 0.022 2.32 0.02 Specialty care focus 0.037 0.021 1.76 0.08 Lack of strategy 0.060 0.019 3.17 0.00 Ownership -0.087 0.014 -6.04 0.00 Percent Medicaid -0.004 0.000 -10.18 0.00 Percent Medicare -0.002 0.001 -2.81 0.01 Market competition -0.425 0.248 -1.71 0.09 Occupancy -0.160 0.043 -3.73 0.00 Total beds 0.000 0.000 -0.53 0.60 Chain -0.032 0.012 -2.74 0.01 Acuity index -0.028 0.007 -4.07 0.00 ADL index -0.110 0.017 -6.49 0.00 Tube feed care -0.666 0.143 -4.66 0.00 Respiratory care 0.139 0.081 1.71 0.09 Suctioning care 0.444 0.419 1.06 0.29 IV therapy care 0.240 0.213 1.13 0.26 Tracheostomy care -0.238 0.449 -0.53 0.60 Need injection 0.038 0.068 0.56 0.58 _cons 0.38 0.283 0.77 0.44 Table A-17. OLS regressions of root transformed PA skin ulcers w ith differentiators as ref. group Coef. Std. Err. t P>t Cost leaders -0.045 0.020 -2.30 0.02 Rehab focus -0.015 0.021 -0.74 0.46 Specialty care focus -0.027 0.021 -1.28 0.20 Lack of strategy -0.076 0.020 -3.81 0.00 Ownership 0.034 0.014 2.38 0.02 Percent Medicaid 0.001 0.000 3.39 0.00 Percent Medicare 0.001 0.001 0.70 0.48 Market competition -0.031 0.308 -0.10 0.92 Occupancy 0.014 0.046 0.31 0.75 Total beds 0.000 0.000 5.08 0.00 Chain -0.021 0.012 -1.76 0.08 Acuity index 0.034 0.007 4.88 0.00 ADL index 0.027 0.019 1.42 0.16 Tube feed care 0.568 0.156 3.64 0.00 Respiratory care 0.088 0.082 1.07 0.28 Suctioning care 0.535 0.413 1.30 0.20 IV therapy care 0.270 0.194 1.40 0.16 Tracheostomy care -0.160 0.431 -0.37 0.71 Need injection 0.183 0.069 2.63 0.01 _cons -2.513 0.098 -25.77 0.00 113

PAGE 114

Table A-18. Negative binomial of qua lity of care deficiencies with rehab focus as reference group IRR Std. Err. z P>z Cost leader 1.067 0.032 2.14 0.03 Differentiator 1.007 0.038 0.18 0.86 Specialty care focus 1.008 0.033 0.25 0.80 Lack of strategy 1.132 0.035 4.03 0.00 Ownership 1.099 0.029 3.59 0.00 Percent Medicaid 1.005 0.001 7.69 0.00 Percent Medicare 1.005 0.001 3.75 0.00 Market competition 0.773 0.367 -0.54 0.59 Occupancy 0.830 0.065 -2.38 0.02 Total beds 1.003 0.000 14.74 0.00 Chain 1.024 0.021 1.18 0.24 Acuity index 1.040 0.013 3.20 0.00 ADL index 0.952 0.029 -1.63 0.10 Tube feed care 1.009 0.237 0.04 0.97 Respiratory care 0.972 0.157 -0.18 0.86 Suctioning care 0.811 0.454 -0.37 0.71 IV therapy care 3.826 1.304 3.94 0.00 Tracheostomy care 1.142 0.745 0.20 0.84 Need injection 1.189 0.141 1.46 0.14 _Ifips_2 1.206 0.470 0.48 0.63 _Ifips_3 1.025 0.198 0.13 0.90 Table A-19. OLS regressions of r oot transformed bladder decline with rehab focus as ref. group Coef. Std. Err. t P>t Cost leader -0.006 0.003 -1.97 0.05 Differentiator -0.012 0.004 -3.27 0.00 Specialty care focus -0.003 0.003 -0.86 0.39 Lack of strategy -0.007 0.003 -2.32 0.02 Ownership 0.006 0.002 2.46 0.01 Percent Medicaid 0.000 0.000 -5.35 0.00 Percent Medicare 0.001 0.000 3.84 0.00 Market competition 0.077 0.064 1.21 0.23 Occupancy 0.014 0.008 1.70 0.09 Total beds 0.000 0.000 3.66 0.00 Chain 0.004 0.002 1.97 0.05 Acuity index 0.009 0.001 6.66 0.00 ADL index 0.021 0.003 6.68 0.00 Tube feed care -0.061 0.027 -2.24 0.03 Respiratory care -0.040 0.014 -2.99 0.00 Suctioning care 0.097 0.067 1.44 0.15 IV therapy care 0.004 0.040 0.09 0.93 Tracheostomy care -0.245 0.074 -3.32 0.00 Need injection 0.004 0.012 0.33 0.74 _cons 0.253 0.017 14.72 0.00 114

PAGE 115

Table A-20. OLS regressions of r oot transformed bowel decline with rehab focus as ref. group Coef. Std. Err. t P>t Cost leader -0.001 0.003 -0.29 0.77 Differentiator -0.009 0.004 -2.22 0.03 Specialty care focus 0.002 0.003 0.64 0.52 Lack of strategy -0.004 0.003 -1.26 0.21 Ownership 0.005 0.003 1.83 0.07 Percent Medicaid 0.000 0.000 -3.76 0.00 Percent Medicare 0.001 0.000 3.82 0.00 Market competition 0.040 0.050 0.79 0.43 Occupancy 0.013 0.009 1.49 0.14 Total beds 0.000 0.000 2.94 0.00 Chain 0.004 0.002 1.79 0.07 Acuity index 0.011 0.001 7.26 0.00 ADL index 0.025 0.003 7.43 0.00 Tube feed care -0.039 0.029 -1.34 0.18 Respiratory care -0.040 0.015 -2.74 0.01 Suctioning care 0.014 0.077 0.18 0.86 IV therapy care 0.016 0.044 0.36 0.72 Tracheostomy care -0.101 0.085 -1.19 0.23 Need injection 0.020 0.013 1.55 0.12 _cons 0.186 0.017 10.67 0.00 Table A-21. OLS regressions of r oot transformed ADL decline with rehab focus as ref. group Coef. Std. Err. t P>t Cost leader 0.006 0.003 2.18 0.03 Differentiator -0.003 0.003 -0.86 0.39 Specialty care focus 0.002 0.003 0.86 0.39 Lack of strategy 0.002 0.003 0.63 0.53 Ownership 0.001 0.002 0.50 0.62 Percent Medicaid 0.000 0.000 0.10 0.92 Percent Medicare 0.000 0.000 2.97 0.00 Market competition 0.083 0.053 1.57 0.12 Occupancy 0.006 0.007 0.84 0.40 Total beds 0.000 0.000 -0.71 0.48 Chain -0.001 0.002 -0.72 0.47 Acuity index 0.002 0.001 1.46 0.14 ADL index -0.003 0.003 -1.01 0.31 Tube feed care -0.028 0.026 -1.10 0.27 Respiratory care -0.005 0.013 -0.42 0.67 Suctioning care -0.054 0.057 -0.95 0.34 IV therapy care -0.013 0.033 -0.39 0.70 Tracheostomy care -0.072 0.066 -1.10 0.27 Need injection -0.008 0.011 -0.69 0.49 _cons 0.377 0.016 23.89 0.00 115

PAGE 116

Table A-22. OLS regressions of log transformed restraint with rehab focus as ref. group Coef. Std. Err. t P>t Cost leader 0.011 0.004 2.48 0.01 Differentiator 0.004 0.006 0.74 0.46 Specialty care focus 0.006 0.005 1.32 0.19 Lack of strategy 0.006 0.005 1.24 0.22 Ownership 0.020 0.004 5.50 0.00 Percent Medicaid 0.000 0.000 4.06 0.00 Percent Medicare 0.000 0.000 1.16 0.25 Market competition 0.053 0.070 0.76 0.45 Occupancy -0.010 0.012 -0.84 0.40 Total beds 0.000 0.000 7.90 0.00 Chain -0.007 0.003 -2.24 0.03 Acuity index 0.011 0.002 6.07 0.00 ADL index 0.016 0.004 3.80 0.00 Tube feed care -0.049 0.041 -1.20 0.23 Respiratory care -0.012 0.022 -0.54 0.59 Suctioning care -0.138 0.101 -1.37 0.17 IV therapy care 0.028 0.054 0.51 0.61 Tracheostomy care 0.060 0.110 0.54 0.59 Need injection -0.055 0.018 -3.12 0.00 _cons -0.004 0.023 -0.19 0.85 Table A-23. OLS regressions of r oot transformed skin ulcer with rehab focus as ref. group Coef. Std. Err. t P>t Cost leader -0.013 0.002 -5.69 0.00 Differentiator -0.012 0.003 -4.19 0.00 Specialty care focus -0.008 0.002 -3.34 0.00 Lack of strategy -0.016 0.002 -6.89 0.00 Ownership 0.006 0.002 3.32 0.00 Percent Medicaid 0.000 0.000 -0.71 0.48 Percent Medicare 0.001 0.000 6.94 0.00 Market competition -0.001 0.029 -0.04 0.97 Occupancy -0.035 0.006 -5.55 0.00 Total beds 0.000 0.000 3.31 0.00 Chain -0.002 0.002 -1.57 0.12 Acuity index 0.008 0.001 8.42 0.00 ADL index 0.007 0.002 3.11 0.00 Tube feed care 0.109 0.024 4.60 0.00 Respiratory care 0.031 0.011 2.74 0.01 Suctioning care 0.045 0.060 0.74 0.46 IV therapy care 0.150 0.032 4.70 0.00 Tracheostomy care -0.031 0.065 -0.47 0.64 Need injection 0.046 0.010 4.74 0.00 _cons 0.163 0.012 13.90 0.00 116

PAGE 117

Table A-24. GLM of PA walking improve with gamma distribution with rehab focus as ref. group Coef. Std. Err. z P>z Cost leader -0.019 0.018 -1.09 0.28 Differentiator -0.050 0.022 -2.32 0.02 Specialty care focus -0.014 0.019 -0.73 0.47 Lack of strategy 0.010 0.018 0.57 0.57 Ownership -0.087 0.014 -6.04 0.00 Percent Medicaid -0.004 0.000 -10.18 0.00 Percent Medicare -0.002 0.001 -2.81 0.01 Market competition -0.425 0.248 -1.71 0.09 Occupancy -0.160 0.043 -3.73 0.00 Total beds 0.000 0.000 -0.53 0.60 Chain -0.032 0.012 -2.74 0.01 Acuity index -0.028 0.007 -4.07 0.00 ADL index -0.110 0.017 -6.49 0.00 Tube feed care -0.666 0.143 -4.66 0.00 Respiratory care 0.139 0.081 1.71 0.09 Suctioning care 0.444 0.419 1.06 0.29 IV therapy care 0.240 0.213 1.13 0.26 Tracheostomy care -0.238 0.449 -0.53 0.60 Need injection 0.038 0.068 0.56 0.58 _Ifips_2 0.218 0.283 0.77 0.44 _Ifips_3 0.209 0.188 1.11 0.27 Table A-25. OLS regressions of root transformed PA skin ulcers with rehab focus as ref. group Coef. Std. Err. t P>t Cost leader -0.030 0.016 -1.91 0.06 Differentiator 0.015 0.021 0.74 0.46 Specialty care focus -0.012 0.017 -0.69 0.49 Lack of strategy -0.061 0.016 -3.80 0.00 Ownership 0.034 0.014 2.38 0.02 Percent Medicaid 0.001 0.000 3.39 0.00 Percent Medicare 0.001 0.001 0.70 0.48 Market competition -0.031 0.308 -0.10 0.92 Occupancy 0.014 0.046 0.31 0.75 Total beds 0.000 0.000 5.08 0.00 Chain -0.021 0.012 -1.76 0.08 Acuity index 0.034 0.007 4.88 0.00 ADL index 0.027 0.019 1.42 0.16 Tube feed care 0.568 0.156 3.64 0.00 Respiratory care 0.088 0.082 1.07 0.28 Suctioning care 0.535 0.413 1.30 0.20 IV therapy care 0.270 0.194 1.40 0.16 Tracheotomy care -0.160 0.431 -0.37 0.71 Need injection 0.183 0.069 2.63 0.01 _cons -2.528 0.097 -26.06 0.00 117

PAGE 118

Table A-26. Negative binomial of qua lity of care deficiencies with specialty care as reference group IRR Std. Err. z P>z Cost leader 1.058 0.030 2.00 0.05 Rehab focus 0.992 0.033 -0.25 0.80 Differentiator 0.999 0.037 -0.04 0.97 Lack of strategy 1.123 0.032 4.02 0.00 Ownership 1.099 0.029 3.59 0.00 Percent Medicaid 1.005 0.001 7.69 0.00 Percent Medicare 1.005 0.001 3.75 0.00 Market competition 0.773 0.367 -0.54 0.59 Occupancy 0.830 0.065 -2.38 0.02 Total beds 1.003 0.000 14.74 0.00 Chain 1.024 0.021 1.18 0.24 Acuity index 1.040 0.013 3.20 0.00 ADL index 0.952 0.029 -1.63 0.10 Tube feed care 1.009 0.237 0.04 0.97 Respiratory care 0.972 0.157 -0.18 0.86 Suctioning care 0.811 0.454 -0.37 0.71 IV therapy care 3.826 1.304 3.94 0.00 Tracheostomy care 1.142 0.745 0.20 0.84 Need injection 1.189 0.141 1.46 0.14 _Ifips_2 1.206 0.470 0.48 0.63 Table A-27. OLS regressions of root transformed bladder decline with specialty care focus as ref. group Coef. Std. Err. t P>t Cost leader -0.003 0.003 -1.13 0.26 Rehab focus 0.003 0.003 0.86 0.39 Differentiator -0.009 0.004 -2.68 0.01 Lack of strategy -0.004 0.003 -1.55 0.12 Ownership 0.006 0.002 2.46 0.01 Percent Medicaid 0.000 0.000 -5.35 0.00 Percent Medicare 0.001 0.000 3.84 0.00 Market competition 0.077 0.064 1.21 0.23 Occupancy 0.014 0.008 1.70 0.09 Total beds 0.000 0.000 3.66 0.00 Chain 0.004 0.002 1.97 0.05 Acuity index 0.009 0.001 6.66 0.00 ADL index 0.021 0.003 6.68 0.00 Tube feed care -0.061 0.027 -2.24 0.03 Respiratory care -0.040 0.014 -2.99 0.00 Suctioning care 0.097 0.067 1.44 0.15 IV therapy care 0.004 0.040 0.09 0.93 Tracheostomy care -0.245 0.074 -3.32 0.00 Need injection 0.004 0.012 0.33 0.74 _cons 0.251 0.017 14.71 0.00 118

PAGE 119

Table A-28. OLS regressions of root transformed bowel decline with specialty care as ref. group Coef. Std. Err. t P>t Cost leader -0.003 0.003 -1.10 0.27 Rehab focus -0.002 0.003 -0.64 0.52 Differentiator -0.012 0.004 -2.92 0.00 Lack of strategy -0.006 0.003 -2.18 0.03 Ownership 0.005 0.003 1.83 0.07 Percent Medicaid 0.000 0.000 -3.76 0.00 Percent Medicare 0.001 0.000 3.82 0.00 Market competition 0.040 0.050 0.79 0.43 Occupancy 0.013 0.009 1.49 0.14 Total beds 0.000 0.000 2.94 0.00 Chain 0.004 0.002 1.79 0.07 Acuity index 0.011 0.001 7.26 0.00 ADL index 0.025 0.003 7.43 0.00 Tube feed care -0.039 0.029 -1.34 0.18 Respiratory care -0.040 0.015 -2.74 0.01 Suctioning care 0.014 0.077 0.18 0.86 IV therapy care 0.016 0.044 0.36 0.72 Tracheostomy care -0.101 0.085 -1.19 0.23 Need injection 0.020 0.013 1.55 0.12 _cons 0.188 0.017 10.99 0.00 Table A-29. OLS regressions of r oot transformed ADL decline with specialty care as ref. group Coef. Std. Err. t P>t Cost leader 0.003 0.002 1.39 0.16 Rehab focus -0.002 0.003 -0.86 0.39 Differentiator -0.005 0.003 -1.67 0.10 Lack of strategy -0.001 0.002 -0.28 0.78 Ownership 0.001 0.002 0.50 0.62 Percent Medicaid 0.000 0.000 0.10 0.92 Percent Medicare 0.000 0.000 2.97 0.00 Market competition 0.083 0.053 1.57 0.12 Occupancy 0.006 0.007 0.84 0.40 Total beds 0.000 0.000 -0.71 0.48 Chain -0.001 0.002 -0.72 0.47 Acuity index 0.002 0.001 1.46 0.14 ADL index -0.003 0.003 -1.01 0.31 Tube feed care -0.028 0.026 -1.10 0.27 Respiratory care -0.005 0.013 -0.42 0.67 Suctioning care -0.054 0.057 -0.95 0.34 IV therapy care -0.013 0.033 -0.39 0.70 Tracheostomy care -0.072 0.066 -1.10 0.27 Need injection -0.008 0.011 -0.69 0.49 _cons 0.379 0.016 24.46 0.00 119

PAGE 120

Table A-30. OLS regressions of r oot transformed restraint with specialty care as ref. group Coef. Std. Err. t P>t Cost leader 0.005 0.004 1.22 0.22 Rehab focus -0.006 0.005 -1.32 0.19 Differentiator -0.002 0.005 -0.41 0.68 Lack of strategy -0.001 0.004 -0.15 0.88 Ownership 0.020 0.004 5.50 0.00 Percent Medicaid 0.000 0.000 4.06 0.00 Percent Medicare 0.000 0.000 1.16 0.25 Market competition 0.053 0.070 0.76 0.45 Occupancy -0.010 0.012 -0.84 0.40 Total beds 0.000 0.000 7.90 0.00 Chain -0.007 0.003 -2.24 0.03 Acuity index 0.011 0.002 6.07 0.00 ADL index 0.016 0.004 3.80 0.00 Tube feed care -0.049 0.041 -1.20 0.23 Respiratory care -0.012 0.022 -0.54 0.59 Suctioning care -0.138 0.101 -1.37 0.17 IV therapy care 0.028 0.054 0.51 0.61 Tracheostomy care 0.060 0.110 0.54 0.59 Need injection -0.055 0.018 -3.12 0.00 _cons 0.002 0.023 0.08 0.93 Table A-31. OLS regressions of root transformed LTC skin ulcer with specialty care as ref. group Coef. Std. Err. t P>t Cost leader -0.005 0.002 -2.50 0.01 Rehab focus 0.008 0.002 3.34 0.00 Differentiator -0.004 0.003 -1.47 0.14 Lack of strategy -0.008 0.002 -3.97 0.00 Ownership 0.006 0.002 3.32 0.00 Percent Medicaid 0.000 0.000 -0.71 0.48 Percent Medicare 0.001 0.000 6.94 0.00 Market competition -0.001 0.029 -0.04 0.97 Occupancy -0.035 0.006 -5.55 0.00 Total beds 0.000 0.000 3.31 0.00 Chain -0.002 0.002 -1.57 0.12 Acuity index 0.008 0.001 8.42 0.00 ADL index 0.007 0.002 3.11 0.00 Tube feed care 0.109 0.024 4.60 0.00 Respiratory care 0.031 0.011 2.74 0.01 Suctioning care 0.045 0.060 0.74 0.46 IV therapy care 0.150 0.032 4.70 0.00 Tracheostomy care -0.031 0.065 -0.47 0.64 Need injection 0.046 0.010 4.74 0.00 _cons 0.155 0.012 13.38 0.00 120

PAGE 121

Table A-32. GLM with gamma distribution PA walking improve with specialty care as ref. group Coef. Std. Err. Z P>z Cost leader -0.005 0.016 -0.33 0.74 Rehab focus 0.014 0.019 0.73 0.47 Differentiator -0.037 0.021 -1.76 0.08 Lack of strategy 0.024 0.016 1.45 0.15 Ownership -0.087 0.014 -6.04 0.00 Percent Medicaid -0.004 0.000 -10.18 0.00 Percent Medicare -0.002 0.001 -2.81 0.01 Market competition -0.425 0.248 -1.71 0.09 Occupancy -0.160 0.043 -3.73 0.00 Total beds 0.000 0.000 -0.53 0.60 Chain -0.032 0.012 -2.74 0.01 Acuity index -0.028 0.007 -4.07 0.00 ADL index -0.110 0.017 -6.49 0.00 Tube feed care -0.666 0.143 -4.66 0.00 Respiratory care 0.139 0.081 1.71 0.09 Suctioning care 0.444 0.419 1.06 0.29 IV therapy care 0.240 0.213 1.13 0.26 Tracheostomy care -0.238 0.449 -0.53 0.60 Need injection 0.038 0.068 0.56 0.58 _Ifips_2 0.218 0.283 0.77 0.44 _cons -0.35 0.188 1.11 0.27 Table A-33. OLS regressions of root transformed PA skin ulcers w ith specialty care as ref. group Coef. Std. Err. T P>t Cost leader -0.018 0.016 -1.15 0.25 Rehab focus 0.012 0.017 0.69 0.49 Differentiator 0.027 0.021 1.28 0.20 Lack of strategy -0.049 0.016 -3.00 0.00 Ownership 0.034 0.014 2.38 0.02 Percent Medicaid 0.001 0.000 3.39 0.00 Percent Medicare 0.001 0.001 0.70 0.48 Market competition -0.031 0.308 -0.10 0.92 Occupancy 0.014 0.046 0.31 0.75 Total beds 0.000 0.000 5.08 0.00 Chain -0.021 0.012 -1.76 0.08 Acuity index 0.034 0.007 4.88 0.00 ADL index 0.027 0.019 1.42 0.16 Tube feed care 0.568 0.156 3.64 0.00 Respiratory care 0.088 0.082 1.07 0.28 Suctioning care 0.535 0.413 1.30 0.20 IV therapy care 0.270 0.194 1.40 0.16 Tracheostomy care -0.160 0.431 -0.37 0.71 Need injection 0.183 0.069 2.63 0.01 _cons -2.540 0.096 -26.36 0.00 121

PAGE 122

Table A-34. GLM regressions of nursing home cost with gamma distribution Coef. Std. Err. z P>z Differentiator 0.002 0.016 0.12 0.90 Rehab focus 0.080 0.015 5.42 0.00 Specialty care focus 0.039 0.014 2.91 0.00 Lack of strategy -0.006 0.012 -0.54 0.59 Ownership -0.108 0.012 -8.87 0.00 Percent Medicaid -0.009 0.000 -27.82 0.00 Percent Medicare 0.001 0.001 1.49 0.13 Market competition -0.188 0.222 -0.85 0.39 Occupancy -0.904 0.035 -26.03 0.00 Total beds -0.001 0.000 -11.94 0.00 Chain -0.043 0.010 -4.38 0.00 Acuity index -0.012 0.006 -2.09 0.03 ADL index 0.001 0.014 0.07 0.94 Tube feed care 0.012 0.120 0.10 0.92 Respiratory care 0.159 0.069 2.29 0.02 Suctioning care 0.330 0.334 0.99 0.32 IV therapy care 0.056 0.175 0.32 0.75 Tracheotomy care 0.636 0.360 1.77 0.07 Need injection 0.093 0.056 1.66 0.09 _Ifips_2 -0.407 0.235 -1.73 0.08 _cons 7.526 0.150 49.92 0.00 Table A-35. Odds of nursing home of being in the highest operating margin tier Odds Ratio Std. Err. Z P>z Cost leader 1.25 0.10 2.79 0.01 Rehab focus 1.46 0.13 4.12 0.00 Specialty care focus 1.04 0.09 0.44 0.66 Differentiator 1.20 0.13 1.71 0.09 Ownership 4.65 0.37 19.15 0.00 Percent Medicaid 1.00 0.00 2.76 0.01 Acuindex 1.04 0.03 1.27 0.21 Total beds 1.00 0.00 1.94 0.05 Market competition 0.92 1.23 -0.06 0.95 Chain Affiliation 1.39 0.09 4.92 0.00 _Ifips_2 131.45 180.57 3.55 0.00 _Ifips_3 153.83 145.15 5.34 0.00 Note: 53 observations completely determ ined. Standard errors questionable 122

PAGE 123

123 Table A-36. Odds of nursing home of being in the highest total margin tier Odds Ratio Std. Err. z P>z Cost leader 1.22 0.09 2.58 0.01 Rehab focus 1.27 0.11 2.74 0.01 Specialty care focus 0.98 0.09 -0.25 0.80 Differentiator 1.31 0.14 2.6 0.01 Ownership 1.64 0.12 6.62 0.00 Percent Medicaid 1.06 0.03 1.98 0.05 Acuindex 1.00 0.00 1.29 0.20 Total beds 0.92 1.25 -0.06 0.95 Market competition 1.04 0.07 0.58 0.56 Chain Affiliation 10.53 15.18 1.63 0.10 Ifips_2 18.39 18.84 2.84 0.00 Note: 57 observations completely determ ined. Standard errors questionable .

PAGE 124

Table A-37. Correlations of independent variables 124 Cost Leader RN Focus Rehab Focus Differentiate Lack of Strategy Ownership Case mix Size Occupancy Pct Medicaid Pct Medicare Competition System Cost Leader 1 RN Focus -0.282 1.000 Rehab Focus -0.301 -0.215 1.000 Differentiator -0.212 -0.151 -0.161 1.000 Lack of Strategy -0.373 -0.267 -0.284 -0.200 1.000 Ownership -0.010 0.006 0.038 -0.041 -0.001 1.000 Case mix -0.032 -0.005 0.069 0.033 -0.047 0.062 1.000 Size 0.126 -0.017 -0.020 -0.026 -0.079 -0.029 0.082 1.000 Occupancy 0.033 -0.046 -0.041 0.034 0.019 -0.084 0.106 -0.027 1.000 Pct Medicaid 0.124 0.000 -0.288 0.041 0.100 0.231 0.024 0.160 0.048 1.000 Pct Medicare -0.136 -0.051 0.434 -0.077 -0.149 0.101 0.098 -0.001 -0.039 -0.477 1.000 Competition -0.010 0.071 -0.025 -0.034 -0.004 0.023 -0.095 -0.178 -0.105 0.088 -0.089 1.000 System -0.004 0.024 0.059 -0.061 -0.026 0.171 0.015 -0.066 -0.046 0.038 0.101 0.093 1

PAGE 125

LIST OF REFERENCES Agresti A, Finlay B, Statistical methods for social sciences, third edition, New Jersey 1997 Allen, M.J., & Yen, W. M. (2002). Introduction to Measurement Theory. Long Grove, IL: Waveland Press. American Physical Therapy Association History of Therapy caps Available at: http://www.apta.org/AM/Template.cfm?Section=Media&TEMPLATE=/CM/ContentDisplay.cfm&CONTENT ID=47095 Accessed March 30th, 2008 Amit, R., & Schoemaker, P. f. H. 1993. Strategic assets and organizational rent. Strategic Management Journal, 14: 33-4B Ansoff, H. Igor, Corporate Strategy, McGraw-Hill, New York, 1965 Abt Associates (2004), National Nursing Home Quality Measures Users Manual, Available at: http://www.cms.hhs.gov/NursingHomeQuality Inits/downloads/NHQIQMUsersManual.pdf Accessed March 30th, 2008 Barney J B. 1991. Firm resources and sustained competitive advantage. Journal of Management, 17: 99-129. Britton L. C.; Clark, T. A. R.; Ball, D. F.. Modify or Extend? The Application of the Structure Conduct Performance Approach to Service Industries. By: Service Industries Journal, Jan92, Vol. 12 Issue 1, p34-43, 10p; Center for Medicare and Medicaid 2007. National Health Expenditures by type of service and source of funds, CY 1960-2006. Available at: http://www.cms.hhs.gov/NationalHealthExpendData/02_N ationalHealthAccountsHistorical.asp#TopOfPage Accessed March 30th, 2008 Center for Disease Control. 2007. NCHS faststats Nursing home care. Available at: http://www.cdc.gov/ Cohen nchs/fastats/nursingh.htm Accessed March 30th, 2007 Center for Medicare and Medicaid 2007 Me dicare Long Term Care Available at: http://www.medicare.gov/longtermcare/static/home.asp Accessed March 30th, 2008 Cool K Schendel, D STRATEGIC GROUP FORMATION AND PERFORMANCE: THE CASE OF THE U.S. PHARMACEUTICAL INDUSTRY, 1963-1982. Management Science ; Sep87, Vol. 33 Issue 9, p11021124, 23p Castle NG. Strategic groups and outcomes in nursi ng facilities. Health Care Manage Rev. 2003 JulSep;28(3):217-27. Cool, K., Schendel, D., (1988) Performance differences among strategic group members. Strategic Management Journal, 9 (3), pp 207-223 Caves, R., Porter, M., (1977), From entry barriers to mobility barriers: conjectural decisions and contrived deterrence to new competition. Quar terly Journal of Economics p.241-261 Centers for Medicare and Medicaid: RAI MDS Users Manual Available at: 125

PAGE 126

http://www.cms.hhs.gov/NursingHomeQualityInits/downloads/MDS20rai1202ch3.pdf Accessed March 30th, 2008 Centers for Medicare and Medicaid : State Operations Manual Appendix PP Guidance to Surveyors for Long Term Care Facilities Available at: http://www.cms.hhs.gov/manuals/Dow nloads/som107ap_pp_guidelines_ltcf.pdf Accessed March 30th, 2008 Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16(3), 297-334 Day, G., Strategic Market Planning, West Publishing Co., St. Paul, 1984 Dess, G., Davis, P., (1984), Porter's (1980) Generic Strategies as Determinants of Strategic Group Membership and Organizational Performance. Academy of Management Journal ; Sep84, Vol. 27 Issue 3, p467-488, 22p, 5 charts Fabrigar L. R. Wegener D. T. MacCallum R. C., & Strahan E. J. (1999). Evaluating the use of exploratory factor analysis in psychological research. Psychological Methods 4 2729. Feng Z Grabowski DC Intrator O Mor V The effect of state medicaid case-mix payment on nursing home resident acuity.Health Serv Res. 2006 Aug;41 (4 Pt 1):1337. Feng Z, Grabowski DC, Intrator O, Zinn J, Mor V. Medicaid payment rates, case-mix reimbursement, and nursing home staffing--1996-2004. Med Care. 2008 Jan;46(1):33-40. Fiegenbaum, A; Thomas, H., INDUSTRY AND STRATEGIC GROUP DYNAMICS: COMPETITIVE STRATEGY IN THE INSURANCE INDUSTRY, 1970-84. Journal of Management Studies, Jan93, Vol. 30 Issue 1, p69-105, 37p Fleishman JA, Cohen JW, Manning WG, Kosinski M. Us ing the SF-12 health status measure to improve predictions of medical expenditures. Med Care. 2006 May;44(5 Suppl):I54-63. Ford EW, Duncan JD, Ginter PM., The Structure of State Health Agencies: A Strategic Analysis. MCR&R 60:1 (March 2003) Fries BE, Schneider DP, Foley WJ, et al. Refining a case-mix measure for nursing homes: Resource Utilization Groups (RUG-III). Med Care.1994;32:668685. Gapenski, L., Understanding Healthcar e Financial Management, Third Ed ition, 2000 Health Administration Press GAO Report to Congressional Requesters (July 2001), Nursing Home Quality: Prevalence of Serious Problems, While Declining Reinforces Importance of Enhanced Oversight, http://www.gao.gov/new.items/d03561.pdf Grabowski DC, Castle NG. Nursing homes with pers istent high and low quality. Med Care Res Rev. 2004 Mar;61(1):89-115. Grabowski DC. A longitudinal study of Medicaid paymen t, private-pay price and nursing home quality. Int J Health Care Finance Econ. 2004;4:526. 126

PAGE 127

Harrington, C., Zimmerman, D., Karon, S.L. Robinson, J. & Beutel, P.(2000). Nursing home staffing and its relationship to deficiencies. Journal of Gerontology: Social Sciences, 55B, S278-S287 Harrington, C., Kovner, C., Mezey, M., Kayser-Jones, J., Burger, S., Mohler, M., Burke, S. & Zimmerman, D. (2000) Experts recommend minimum nurse staffing standards for nursing homes in the United States. The Gerontologist: 40, 5-16 Harrington C, Woolhandler S, Mullan J, Carrillo H, Hi mmelstein DU. Does investor ownership of nursing homes compromise the quality of care? Am J Public Health. 2001 Sep;91(9):1452-5. Hill, C.W. Differentiation Versus Low Cost or Di fferentiation and Low Cost: A Contingency Framework. Academy of Management Review 13 (1988): 401-12 Hitt M., Bierman L.,Shimizu K., Kochhar L. Direct and Moderating Effects of Human Capital on Strategy and Performance in Professional Service Firms: A Resource-Based Perspective Michael A. Hitt, Leonard Bierman, Katsuhiko Shimizu and Rahul Kochhar The Academy of Management Journal Vol. 44, No. 1 (Feb., 2001), pp. 13-28 Hofer, C. Schendel, D., Strategy Formulation: Anal ytical Concepts, West Publishin, St. Paul, 1978. Harrigan, C An Application of Clustering for Strategic Group Analysis. Strategic Management Journal ; JanMar85, Vol. 6 Issue 1, p55-73, 19p Hulin, C., Roznowski, M., (1985) Organizational Technol ogies: Effect on organizational characteristics and individuals responses. Research in Or ganizational behavior, vol7, p.39-85 Hunt, M.S. Competition in the Major Home App liance Industry, 1960-1970. Unpublished Doctoral dissertation, Harvard, 1972. Instititute of Medicine, 1986. Improving the qualit y of care of nursing homes. Washington DC: National Academy Press Johnson CE, Dobalian A, Burkhard J, Hedgecock DK, Harman J. Predicting lawsuits against nursing homes in the United States, 1997-2001. Health Serv Res. 2004 Dec;39(6 Pt 1):1713-31. Kachigan S.K., Multivariate Statistical analysis: A conceptual in troduction, Radius, New York, 1991 Kaiser, H. F. (1960). The application of electronic computers to factor analysis. Educational and Psychological Measurements, 20, 141-51 Kaplan, and Norton. 2001. Transforming the balanced scorecard from performance measurement to strategic management: Part I. Accounting Horizons (March): 87-104 Katz, R., Cases and Concepts in Corporate St rategy, Prentice-Hall, Englewood Cliffs, N.J.1970 Ketchen DJ, Shook CJ. The Application of Cluster Analys is In Strategic Management Research: An Analysis and Critique. Strategic Manageme nt Journal 17 (1996): 441-58 Kling, J A.; Smith, K A. Identifying Strategic Groups in the U.S. Airline Industry: An Application of the Porter Model. Transportation Journal, Winter95, Vol. 35 Issue 2, p26-34, 9p 127

PAGE 128

Laberge A, Weech-Maldonado R, Johnson CE, Jia H, Dewald L.Outsourcing vetera ns for long-term care: comparison of community and state veterans' nursing homes. J Health Hum Serv Adm. 2008 Spring;30(4):44167. Leask, G.; Parker, D. Strategic groups, competitive groups and perfo rmance within the U.K. pharmaceutical industry: Improving our understanding of the competitive process. Strategic Management Journal, Jul2007, Vol. 28 Issue 7, p723-745 Marlin D, Sun M, Huonker JW. Strategic groups and performance in the nursing home industry: a reexamination.Med Care Res Rev. 1999 Jun;56(2):156-76 Marlin D, Huonker JW, Sun M. An examination of th e relationship between strategic group membership and hospital performance. Health Care Manage Rev. 2002 Fall;27(4):18-29 Mukamel DB, Glance LG, Li Y, Weimer DL, Spector WD, Zinn JS, Mosqueda L. Does risk adjustment of the CMS quality measures for nursing homes ma tter? Med Care. 2008 May;46(5):532-41. Mehra, Ajay. RESOURCE AND MARKET BASED DETERMINANTS OF PERFORMANCE IN THE U.S. BANKING INDUSTRY. Strategic Management Journal, Apr96, Vol. 17 Issue 4, p307-322 Murray, A.L. A Contingency view of porters generic strategies. Academy of Management Review 13 (1988): 390-400 National Institute of Health, http://www.nlm.nih.gov/medlineplus/nursinghomes.html accessed March 20th, 2008 O'Neill C, Harrington C, Kitchener M, Saliba D. Quality of care in nursing homes: an analysis of relationships among profit, quality, and ownership. Med Care. 2003 Dec;41(12):1318-30. Pagano E, Bo S, Petrinco M, Rosato R, Merletti F, Gre gori D. Factors affecting hospitalization costs in Type 2 diabetic patients. J Diabetes Complications. 2009 Jan-Feb;23(1):1-6. Epub 2008 Apr 16. Pfeffer, J. 1994. Competitive advantage through people Boston: Harvard Business School Press Porter, M.E., Competitive Strategy, New York: Free press, 1980 Rhoades J, Sommers J. Expenses and Sources of Payment for Nursing Home Residents 1996. Rockville, MD: Agency for Healthcare Research and Quality; 2000. MEPS Research Findings No 13. AHRQ Pub. No 01-0010 SAS manual 1990, SAS Technical Report A-108 Scanlon, W. 1980. A theory of the nursing home market. Inquiry 17 (1):25-41. Schnelle JF. Determining the relati onship between staffing and quality. Gerontologist. Feb 2004;44(1):10-12 Shen YC, Eggleston K, Lau J, Schmid CH.Hospital owne rship and financial performance: what explains the different findings in the empirical litera ture? Inquiry. 2007 Spring;44(1):41-68 Short JC, Palmer TB, Ketchen DJ Jr. Resource-based and strategic group influences on hospital performance. Health Care Manage Re v. 2002 Fall;27(4):7-17. 128

PAGE 129

STATA. STATA reference manual release 7. College station: STATA Press; 2001 Swan JH, Harrington C Clemea W Pickard RB Studer L deWit SK Med Care Res Rev. 2000 Sep;57(3):361-78 Medicaid nursing facility reimbursement methods: 1979-1997 The Medicare Payment Advisory Commission 2007 Report to the Congress: Medicare Payment Policy (March 2007) Available at: http://www.MedPAC.gov/documents/Mar07_EntireReport.pdf Accessed March 30th, 2007 The Medicare Payment Advisory Commission 2003 Report to the Congress: Medicare Payment Policy (March 2007) Available at: http://www.MedPAC.gov/documents/Mar03_EntireReport.pdf Accessed March 30th, 2007 Tremblay, V. J., Strategic groups and the demand for beer. Journal of Industrial Ec onomics, Dec85, Vol. 34 Issue 2, p183, 16p United States General Accounting Office 2000 NUR SING HOMES: Aggregate Medicare Payments Are Adequate Despite Bankruptcies, GAO/T-HEHS-00-192 Venkatraman N., Vasudevan R. Measurement of Busin ess Performance in Strategy Research: A Comparison of Approaches The Academy of Management Review Vol. 11, No. 4. (Oct., 1986), pp. 801-814. Walshe K, Harrington C. Regulation of nursing facilities in the United States: an analysis of resources and performance of state survey agencies. Gerontologist. 2002 Aug;42(4):475-87. Weech-Maldonado R, Neff G, Mor V. Does quality of care lead to better financial performance?: the case of the nursing home industry. Health Care Manage Rev. 2003 Jul-Sep;28(3):201-16. Weech-Maldonado, R., Meret-Hanke, L., Neff, M., and Mor, V. (2004). Nurse staffing patt erns and quality of care in nursing homes. Health Care Management Review, 29 (2), 1-10. Weech-Maldonado, R., Laberge, A., Pradham, R., Johnson, C., Yang, Z., Hyers, K., Nursing home quality and financial performance. Working Paper White H. A. Heteroskedasticity-Consistent Covari ance Matrix Estimator and a Direct Test for Heteroskedasticity Econometrica 1980; 48 (4) :8178 Wodchis WP, Fries BE, Pollack H. Payer incentives and physical rehabilitation therapy for nonelderly institutional long-term care residents: evidence from Mi chigan and Ontario. Arch Phys Med Rehabil. 2004 Feb;85(2):210-7. Wright P., A Refinement of Porters Strategies Strategic Management Journal 8 (1987) 93-101 Zinn JS, Aaronson WE, Rosko MD. Strategic groups, performance, and strategic response in the nursing home industry. Health Serv Res. 1994 Jun;29(2):187-205. Zinn J, Feng Z, Mor V, Intrator O, Grabowski D. Re structuring in response to case mix reimbursement in nursing homes: a contingency approach. Health Care Manage Rev. 2008 Apr-Jun;33(2):113-23. 129

PAGE 130

130 BIOGRAPHICAL SKETCH Alexandre Laberge completed his honors bachelor s of physical therapy at the University of Western Ontario in London, Ontario in 1995. He worked as a physical therapist in a variety of settings including hospitals, skilled nursing facilities, assisted living facili ties and in the home health settings. He completed his Masters of Business Administration in informa tion technology at Goldey Beacom College in 2002. He completed his Ph.D. in health se rvices research, management and policy in May in 2009. He worked in skilled nursing facilities for over 6 years as a therapist as well as a rehab manager. In the 4.5 years that he worked as a research assi stant, he worked closely with Dr Weech-Maldonado on a variety of projects that examin ed the relationship between cost, quality and financial performance in nursing homes.