Resident and caregiver behaviors associated with falls in a long-term psychiatric residential treatment facility

MISSING IMAGE

Material Information

Title:
Resident and caregiver behaviors associated with falls in a long-term psychiatric residential treatment facility
Alternate title:
Resident & caregiver behaviors associated with falls in a long-term psychiatric residential treatment facility
Physical Description:
vi, 175 leaves : ill. ; 29 cm.
Language:
English
Creator:
Warren, Jacquelyn K
Publication Date:

Subjects

Subjects / Keywords:
Research   ( mesh )
Residential Facilities   ( mesh )
Accidental Falls   ( mesh )
Risk Factors   ( mesh )
Caregivers   ( mesh )
Mentally Ill Persons   ( mesh )
Department of Nursing thesis Ph.D   ( mesh )
Dissertations, Academic -- College of Nursing -- Department of Nursing -- UF   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D)--University of Florida, 2003.
Bibliography:
Bibliography: leaves 159-174.
Statement of Responsibility:
by Jacquelyn K. Warren.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002960439
notis - APK2157
System ID:
AA00011790:00001

Full Text










RESIDENT AND CAREGIVER BEHAVIORS ASSOCIATED
WITH FALLS IN A LONG-TERM PSYCHIATRIC
RESIDENTIAL TREATMENT FACILITY














By

JACQUELYN K. WARREN


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


2003














ACKNOWLEDGEMENTS

Special appreciation is extended to Dr. Faye Gary, chairperson of my supervisory

committee, for her tirelessness and compassion. She knew how to motivate me when no

one else could. I would also like to thank the other members of my supervisory

committee for their special contributions: Dr. Martha J. Hardman for her enthusiasm and

energy; Dr. Hossein Yarandi for his assistance with the data analysis portion of the

project; Dr. Lois Malasanos, who always knows how to keep things in perspective; and

Dr. Terry Mills for his willingness to see this through.

I would also like to extend my deep and humble appreciation and love to my

family. Jennifer, Emily, and Danny have been supportive and encouraging. They make

sure that I find humor in all that surrounds me, and remind me that I do not always need

to take myself so seriously.

I would also like to take this opportunity to thank Honors student, Tricia Hawkins

for her patience, perseverance, and sense of humor. She will undoubtedly do great

things.

Finally, I want to thank the residents and direct caregivers of Wellspring for their

contributions. Without their support and assistance, this project would have never come

to fruition.














TABLE OF CONTENTS

Page


A CKN OW LED GEM ENTS ....................................................................................... ......... ii

ABSTRA CT .......... ........................................................................................................ v

1 INTRODU CTION .................................................................................................. 1

Falls: M yth vs. Reality........................................................................................... 2
The Department of Children and Families Mental Health Treatment Authority.... 3
Statem ent of the Problem ..................................................................................... 4
Conceptual Fram ew ork........................................................................................ 4
Research Hypothesis............................................................................................ 8
Significance of the Study..................................................................................... 9
Definition of Term s............................................................................................ 12
A ssum ptions.......................................................................... ....................... 15
Lim itations...................................................................................................... 16

2 REVIEW OF THE LITERA TURE ................................................................... .. 18

Conceptualization of Falls .................................... .......................................... 18
Types of Falls........................................................................................................ 19
Risk Factors A associated w ith Falls ....................................................................... 20
Consequences of Falls......................................................................................... .. 22
The Political Econom y Perspective ...................................................................... 26
Caregiver Stress ...................... ..................................................... ................... 28

3 M ETH OD OLOGY .................................................................................. ............. 30

Research D esign.................................................................................................... 30
The Research Setting ............................ ................................................... ...... 31
Individuals Who Have Fallen and Individuals Who Have No Reported
History of Falling and are Residing in a Long-Term State Mental
Health Treatm ent Facility............................................. ............. .......... 33
Population ............................................................................................................ 37
D ata Collection Procedures.................................................................................. 41
Instrum ents Section.............................................................................................. 46
D ata Analysis....................................................................................................... 59
Protection of Hum an Subjects ................................................ ............................. 60








4 RESULTS AND DISCUSSION........................................................................ 61

Overview of the Residential Facility Sample .................................................... 61
Resident Sample Characteristics........................................................................ 64
Differences Between Two Resident Sample Groups......................................... 72
Research Hypotheses ......................................................................................... 75
Direct Caregivers ........................................................................... .................... 80
Summary........................................................................... ................................... 84

5 DISCUSSION, CONCLUSIONS, PRACTICE IMPLICATIONS, AND
RECOMMENDATIONS...................................................................................... 87

Discussion of Findings.......................................................................................... 87
Conclusions........................................................................................................... 98
Implications for Practice ............................................................................ ... 100
Implications for Mental Health Policy................................................................ 102
Recommendations ............................................................................................... 103

APPENDIX

A INSTRUMENTS ...................... ......................................................... 107

B METHODOLOGY DIAGRAM .............................................................. 151

C INSTITUTIONAL REVIEW BOARD STUDY ACCEPTANCE
AND PROTECTION OF HUMAN SUBJECTS.................................... 154

REFERENCES ............................................................................................................... 159

BIOGRAPHICAL SKETCH ............................................... ..................................... 175














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

RESIDENT AND CAREGIVER BEHAVIORS ASSOCIATED WITH FALLS IN A
LONG-TERM PSYCHIATRIC RESIDENTIAL TREATMENT FACILITY

By

Jacquelyn K. Warren

May 2003

Chair: Faye A. Gary
Major Department: Nursing

A review of the literature reveals that persons residing in long-term psychiatric

residential treatment facilities are at risk for falls. Therefore, the purpose of this research

was to describe two groups of residents, those with a recorded history of falls during the

previous 12 months and those with no recorded history of falls during the previous 12

months. The specific objectives of the study were to (1) describe characteristics of two

resident groups; (2) identify relationships among characteristics of residents, Cardinal

Risk Factor scores (CRS) and the incidence of falls; and (3) analyze relationships among

these characteristics. An adaptation of Andersen's Behavior Model for Vulnerable

Populations provided the theoretical framework for this research.

A series of standardized instruments were used for data collection. In addition,

residents were asked several open-ended questions about recent fall experiences, how

falling made them feel, and how they thought the fall occurred. Similarly, direct

caregivers (N=20) were asked to participate in interviews with the aim of gleaning salient








information about the functional status of residents as part of the Cardinal Needs

Schedule. They were also asked several open-ended questions regarding resident falls,

such as whether they knew how to avoid falls, what types of training they received, and

whether they found the training to be helpful.

Of the 58 residents that were interviewed, a typical portrait of a resident in a long-

term psychiatric residential treatment facility included in this study can be described as a

49-year old, never married, Caucasian woman who has graduated from high school and

has been a resident of Wellspring for approximately 53 months. She has a diagnosis of

schizophrenia and has had at least one prior long-term psychiatric residential treatment

facility admission.

A stepwise multiple regression analysis was used resulting in two variables that

significantly explained the incidence of falls in the study: the resident's perceived health

deficits as measured by the SF-36v2 Health Survey and CRS as measured by the

Cardinal Needs Schedule. Most residents deny having fallen, even when there is

documented evidence to suggest otherwise. More research would be beneficial to

explore this phenomenon and to understand its implications.














CHAPTER 1
INTRODUCTION

Falls have been identified as the second leading cause of accidental deaths in the

United States, and 75% of those falls occur in the elderly population (Morse, 1997a,

1997b). Among older adults, falls are also the most common cause of serious injury,

including broken hips. Annually, in the United States, as many as 200,000 hip fractures

are the result of fall incidents, with less than half of the hip fracture victims returning to

their previous level of functioning. Hospital falls represent a leading cause of adverse

events, accounting for 25 to 89% of all reported inpatient incidents (Jones & Smith, 1989;

Maciorowski et al., 1988; Tideiksaar, 1996; Tideiksaar & Kay, 1986). Falls account for

22% of all hospitalizations for persons of Medicare age, which is 65, and over (Centers

for Disease Control, 2001).

Historically, the responsibility for falling has, for the most part, been borne by the

person who falls, typically referred to as the individual who has fallen (Tideiksaar, 1998).

Popular misconceptions about falls abound and include the belief that falls are

attributable to either individual carelessness or the natural process of aging. The public

belief that only "old" persons fall is undergirded in the notion that falls are a

manifestation of a general physical and cognitive decline that is a naturally occurring

phenomenon. This belief system extends to include those with multiple disorders in

which one disorder after another culminates to a negative outcome: the fall (Tideiksaar,

1998). Many caregivers have dealt with falls and their adverse consequences for so long








that they have become hardened: they no longer identify falls as problems with solutions,

other than to restrict the individual's mobility. Moreover, health care providers may be

reluctant to explore the possibility of alternate solutions because of their philosophy that

could parallel public belief and myths (Tideiksaar, 1996).

Falls: Myth vs. Reality

Contrary to popular myth, falls, to a large degree, rarely "just happen"--they are

neither accidental nor random events-but are predictable and are the outcome of a

multitude of intrinsic and extrinsic factors that occur either alone or in combination with

one another (Morse, 1997a, 1997b; Tideiksaar, 1996, 1998). Indeed, they are the

outcome of a multitude of risk factors, either host-related or environmental. It is

important to note that a majority of these factors are oftentimes amenable to interventions

(Morse, 1997a, 1997b; Tideiksaar, 1996, 1998). Hence, the minimization or elimination

of these risk factors can perhaps reduce or eliminate falls.

Falls occur among 50 to 67% of nursing home residents (Downton, 1992). The

high prevalence rate of falls in long-term care settings is attributed to the greater degree

of frailty and dependency among the residents and to more accurate reporting of falls by

staff (Tinetti, Williams, & Mayewski, 1986; Tinetti, 1987).

In the past two decades, there has been an effort to reduce the number of fall

incidents in long-term care facilities and to better prepare residents for community living.

These practices have a potential to help to reduce morbidity and mortality rates related to

falls among the elderly (Morse, 1997a, 1997b). Although recent fall-related quality

improvement programs highlighting risk assessment, intervention and outcome

measurements abound (Hill-Westmoreland, Soeken, & Spellbring, 2002; Mosley,








Galindo-Ciocon, Peak, & West, 1998; Ryan & Spellbring, 1996), fall prevention remains

an elusive goal for most health care providers.

The Department of Children and Families Mental Health Treatment
Authority

All State of Florida Mental Health Treatment Facilities operate under the purview

of the Department of Children and Families which oversees activities in 15

geographically defined operating districts. One particular facility in the Department of

Children and Families, Wellspring,' is the focus of this research. It is located in the

northeast portion of the state with an aim to serve five Department of Children and

Families Districts in its structure. These five Districts, 3, 4, 7, 12, and 13, comprise 27

counties (40% of the state's land area) with a mean population of 4,514,624 (30% of

Florida's total population) (University of Florida, 1998). This facility provides

psychiatric services for adult populations from ages 18 upward.

Florida's Mental Health Program Office (PDADM), under the auspices of the

Department of Children and Families, establishes statewide outcome measures and

criteria to assess treatment performance within all of the State's mental health programs.

These Mental Health Program Office measures link directly to Wellspring's strategic

objectives. Specifically, the objectives are (1) management excellence (implementation

of a comprehensive outcome-oriented management system; (2) customer satisfaction

(improved function and delivery of services); (3) treatment excellence and efficiency

(increasing the number of people served); and (4) cost efficiency (reduction of cost of

controllable expenses). In an effort to achieve these strategic objectives, Wellspring has

focused on outcome measures related to the safety and security of the environment, as


' Wellspring is the fictitious name used to reference the long-term care psychiatric facility in Florida.








well as the improvement of the residents' overall functioning. Within this context at

Wellspring, residents' falls, their associated injuries, and overall health outcomes have

been identified as critical outcome measures (Department of Children & Families

[DC&F], 1999).

Statement of the Problem

There is a plethora of information regarding fall risk and fall prevention programs

that focus on prescriptive policy and procedures. That is to say, programs exist that are

broad and general, but they do not provide situation-specific interventions to prevent

falls. Moreover, little is known about individuals who have fallen while residing in long-

term care psychiatric facilities, their perceived risk factors, the direct caregivers'

perceptions of residents' risk factors, and stress experienced by the direct caregivers.

Even less is known about how the manifestations of psychiatric symptomatology, the

types of psychotropic medications prescribed, and the availability of other treatment

modalities impact the residents' quality of care and how these key factors link to the

incidence and prevalence of falls. Perhaps, if these data were available and well

formulated, nurses and other health care providers might develop context-specific

interventions as part of a fall prevention program.

Conceptual Framework

This study is guided by a conceptual model that demonstrates the effects of

numerous independent variables on the outcome variable, falls, in a population of

severely and chronically mentally ill individuals who are hospitalized in a state

psychiatric facility. The independent variables under investigation are age, body mass

index (BMI), perceived and documented deficits in residents' physical health, severity of








mental illness, length of stay, number and types of psychotropic medications, gender,

number of years of formal education, caregiver's perceived self-efficacy, and cardinal

risk factors, which are comprised of residents' perceived risk factors, REHAB indicators,

and primary caregiver stress.

The conceptual model identified in this study is an adaptation of the Behavioral

Model for Vulnerable Populations (Gelberg, Andersen, & Leake, 2000). The original

behavioral model, developed in the late 1960's, used a systems approach to integrate a

range of individual, environmental, and provider-related factors that can affect a person's

decision to seek health care (Andersen, 1995). Gelberg and colleagues (2000) expanded

this model to include vulnerable populations. Vulnerable populations, as conceptualized

by Gelberg et al. (2000), are those groups of people who have limited resources, and who

are at higher risk for disease and injury (Aday & Awe, 1997). Factors that make groups

of people vulnerable may also affect their use of health services and health status (Aday

& Awe, 1997; Aday, 1994; Gelberg, Andersen, & Leake, 1996). Three domains of the

Behavioral Model for Vulnerable Populations predict personal health practices, such as

safe sexual behaviors or the use of health services: These domains are (1) predisposing,

(2) enabling, and (3) need. The predisposing domain is comprised of one's age, gender,

ethnicity, literacy, living conditions, and cognitive ability. The enabling domain

encompasses environmental factors that potentiate one's vulnerability, such as

availability of social services or lack of financial resources. The need domain focuses on

an individual's perceived need, as well as clinical evaluations of need, regarding

conditions that are particularly relevant to vulnerable populations, such as tuberculosis,








acquired immunodeficiency syndrome (AIDS), and severe and persistent mental illness

(Gelberg et al., 2000).

The Behavior Model for Vulnerable Populations (Andersen, 1995) is well-suited

for adaptation to this research: three domains are clearly identified and are consistent

with Andersen's (1995) model. For example, the factors in the predisposing domain

include gender, education level, and age. The enabling domain consists of REHAB

indicators, perceived and documented deficits in physical health, body mass index (BMI),

primary caregiver stress, self-efficacy, length of stay, and number and types of

psychotropic medications. Finally, factors in the need domain are severity of psychiatric

illness and mental status and the resident's perceived risk factors. A schematic diagram

of the relationships between and among the three domains (predisposing, enabling, and

need) and outcome (falls) is depicted in Figure 1.

Gelberg and colleagues (2000) assert that the Behavior Model for Vulnerable

Populations focuses on health service utilization and its effect on health outcomes, such

as consumer satisfaction, compliance, and perceived and evaluated health status.

Conversely, health outcomes are postulated to influence subsequent predisposition,

enabling resources, need for care, and health related behaviors (Andersen, 1995).

Although it is not within the scope of this study to examine health service use

patterns among vulnerable populations, the application of Andersen's (1995) model is

particularly relevant when examining predisposing, enabling, and need factors that may

be associated with the incidence of falls in the severe and persistent mentally ill. For the

purposes of this particular study, the researcher also explored the hypothesized effects of

the identified variables on the outcome indicator, falls (see figure 2).













Predisposing Domain Need Domain
Gender Severity of Psychiatric Illness
Number of Years of Formal and Mental Status
Education Resident's Perceived Risk
Age Factors







Enabling Domain
REHAB Indicators
Perceived and Documented
Deficits in Physical Health
Body Mass Index (BMI) OUTCOMES
Primary Caregiver Stress Falls
Self-Efficacy
Length of Stay
Number and Types of
Psychotropic Medications



Figure 1. Application of the Andersen's (1995) Behavior Model for Vulnerable
Populations to the Current Research.

The relationships between and among variables in this conceptual model will be

explored in the context of Derivational Thinking, a theoretical framework proposed by

(Hardman, 1996). A tenet of this framework proposes that the way in which language is

constructed explicates a particular societal world view. For example, the manner in

which English is constructed and used in everyday life could be posited to support the

notion that power, competition and even domination are a part of the "normal" American

environment (Hardman, 1996). Indeed, the linguistic postulates, or features of English








language that are used repeatedly to organize the universe, demonstrate that human

beings are ranked in terms of gender, ethnicity, education, and income. Language

patterns that evoke images of power, dominance, and subordination are especially

prevalent in the healthcare arena: Images of power or violence are evoked when we

"battle" an illness or disease until we have "fought the good fight" and "conquered" it

(Hardman, 1996). Researchers have explored the use of language and language patterns

in a variety of healthcare settings and have delved into how language patterns influence

caregiver attitudes and behaviors, particularly among chronically ill persons (Breeze &

Repper, 1998; Gignac & Cott, 1998; Hardman, 1996).

Research Hypotheses

Seven research hypotheses are addressed in this study. First, persons who are

advanced in age will have higher Cardinal Risk Factor scores and more frequent

incidences of falls. Second, persons who have higher BMIs will have higher Cardinal

Risk Factor scores and a greater incidence of falls. Third, perceived and recorded deficits

in residents' physical health will have a positive effect on Cardinal Risk Factor scores

and the incidence of falls. Fourth, severity of mental illness and length of stay have a

positive effect on Cardinal Risk Factor scores and those persons are more likely to

experience greater incidences of falls. Fifth, the number and types of psychotropic

medications have a positive effect on Cardinal Risk Factor scores and the incidence of

falls. Sixth, gender, a dichotomous variable, will affect Cardinal Risk Factor scores and

the incidence of falls. Finally, the fewer number of years of formal education, the higher

the Cardinal Risk Factor scores and the greater the number of falls (see figure 2).







CARDINAL
RISK FACTORS

Age +

BMI

Perceived and Documented + Resident's Perceived Risk
Deficits in Physical Health Factors
t
Severity of Psychiatric Illness REHAB Indicators
and Mental Status t FALLS
+ Primary Caregiver Stress
Length of Stay -- I

Number and Types of Self Efficac
Psychotropic Medications + Y /

Gender

Number of Years of Formal
Education

Figure 2. Effects of Independent Variables and Their Cardinal Risk Factors on the
Incidence of Falls in a Population of Residents in a State Psychiatric Hospital

Significance of the Study

According to (Dayhoff, 1997a, 1997b), between 1993 and 1997, 213 articles were

generated on the subject of "prevention and control of falls." The majority of these

studies were identified as descriptive and retrospective and used chart records as the

primary source of data, which were described and "analyzed" in anecdotal form.

Furthermore, most subjects were from one type of facility, a rehabilitation hospital. This

type of facility, as a rule, does not include individuals with severe and persistent mental

illness. Hence, among populations with severe and persistent mental illness, little data








exist regarding perceptions of individuals who have fallen and their willingness to be

helped, or perceptions of staff regarding the risk factors of these residents who fall.

Gaps in existing knowledge are particularly provocative when we consider the

current trend to reduce numbers of hospital beds through reintegrating individuals into

the community. Researchers have suggested that the current trend to deinstitutionalize

persons with severe and persistent mental illness is leading to the neglect of their needs

(Lelliott, Audini, & Darroch, 1995; Powell & Hollander, 1994). One possible

explanation for the neglect of deinstitutionalized persons with severe and persistent

mental illness is the competition for scarce resources. The political economy of aging is

a theoretical explanation that can be used to explain this phenomenon. Indeed, if the

theoretical constructs gleaned from the political economy of aging paradigm are applied

to those persons afflicted with severe and persistent mental illness, a number of parallels

become evident: Socially constructed notions of severe and persistent mental illness

define how the chronically mentally ill are perceived regardless of whether or not these

perceptions have been empirically tested (Estes, Linkins, & Binney, 1996). One of these

notions purports that persons with severe and persistent mental illness "repeatedly fail to

respond to treatment" and use disproportionate amounts of scarce clinical resources

(Gerson, 1994). Hence, these societally constructed perceptions and notions guide

decisions about how the scarce resources will be allocated to the public (Estes et al.,

1996).

Consequently, many persons with severe and persistent mental illness live out

their lives in correctional facilities or on the streets (Murray, Walker, Mitchell, & Pelosi,

1996). In the United States, where economic growth and high profit margins are








perceived as favorable outcomes and "welfare" (financial assistance for the mentally ill)

is identified as encouraging dependency, the severe and persistent mentally ill become

increasingly disenfranchised. In 1996, Congress terminated Supplemental Security

Income (SSI) benefits to individuals disabled by substance abuse. Although many who

reported poor physical health were expected to continue benefits under another disability

category, 64% who reported severe and persistent mental illness were not reclassified and

did not receive public income assistance (Watkins, Podus, & Lombardi, 2001). Their

fate, too often, is homelessness and early deaths (Murray et al., 1996; Watkins et al.,

2001).

Economists and researchers suggest that since mental health care is expensive to

provide, resources should be targeted to those with the greatest need (Slade, Powell, &

Strathdee, 1997). Individuals with the greatest need consist of those in psychiatric

hospitals. Falls, a common occurrence in long-term psychiatric inpatient facilities, are

the most frequently cited cause of broken hips and other severe injuries, in that they

account for resident disability and morbidity that require an inordinate amount of medical

and nursing care, amounting to an estimated $10 billion annually (Poster, Pelletier, &

Kay, 1991; Sattin, 1992). Hence, it is imperative that an accurate assessment of risk

factors is developed which could provide the basis for knowledge to design context-

specific interventions to prevent falls among residents with severe and persistent mental

illness who reside in long-term psychiatric facilities.








Definition of Terms

The definitions of terms used in this study are as follows:

Cognitive function refers to the performance of intellectual tasks, such as

thinking, remembering, perceiving, communicating, orienting, calculating, and problem

solving (Gurland & Cross, 1982; Trzepacz & Baker, 1993).

Fall is an unanticipated change in body position in a downward motion that may

or may not result in physical injury (Hendrich, 1996).

An individual who has fallen refers to any Wellspring resident with a history of at

least one fall (Morse, 1997a). For the purpose of this research, any individual who has

fallen within the past 12 months is referred to as an individual who has fallen.

An individual who has no recorded history of falling is any resident who has

never fallen (Morse, 1997b). For the purpose of this research, any individual who has not

had a reported fall within the past 12 months is an individual who has no recorded history

of falling.

An individual who has fallen repeatedly, is any resident who has had more than

one fall. More specifically, in this study, this individual would have had two or more

falls per year or at least three falls during a consecutive two-month period within the past

year (Cutchins, 1991; Morse, Tylko, & Dixon, 1985; Tinetti, 1987).

An individual who falls and is prone to injury is a resident who is at risk for

sustaining serious injuries during or because of falls (Ginter & Mion, 1992).

Direct caregiver, also known as the primary caregiver, is a hospital

employee who provides direct care to the resident, such as assistance with the resident's

personal maintenance and other activities of daily living. Caregivers include Health








Service Workers (HSW I and II) and Unit Treatment and Rehabilitation Specialist

Supervisor I (UTRSS I). According to the State of Florida job requirements, specific

categories of caregivers are determined by level of education and/or experience and

supervisory responsibilities (State of Florida Classification and Pay Plan, October 14,

1998).

Fall rate is defined as the number of falls divided by the number of

total patient days within a specified time frame (i.e., month) times 1000 (Morse, 1997a,

1997b; Tideiksaar & Kay, 1986; Tinetti, 1987), or

The Number of Falls
Total number of Patient Days on Unit or in Facility X 1000

An example of how the Fall Rate is calculated at Wellspring follows:

39 Falls During the Month of June X 1000 = 2.6 Fall Rate for the Month of June
15,000 Patient Days in Facility

Long-term care psychiatric facility is a facility where the majority of residents

have an average length of stay that is greater than 180 days (Dhillon & Dollieslager,

2000). Long-term care psychiatric facilities offer more specialized services, such as

individual and group therapy sessions, art therapy programs, and other beneficial

psychosocial and rehabilitative activities to prepare the resident for reintegration into the

community (Dhillon & Dollieslager, 2000).

According to several researchers, severe and persistent mental illness can

be conceptualized in terms of a three-dimensional definition (Ruggeri, Leese,

Thornicroft, Bisoffi, & Tansella, 2000). The first component of this three-dimensional

definition is diagnosis of any non-organic psychosis, commonly referred to as a

functional disorder. The second component is a duration of psychiatric treatment lasting








two years or more. Finally, the third component is identified as severe psychiatric

dysfunction, as measured by the Global Assessment of Functioning (GAF) scale

(American Psychiatric Association & American Psychiatric Association Task Force on

DSM-IV, 2000). Severe psychiatric dysfunction is indicated by a GAF score of 50 or

less, indicating severe symptoms or severe difficulty in social, occupational, or school

functioning (Ruggeri et al., 2000). Although many authors differ as to their definition of

severe and persistent mental illness, particularly with regard to the first component, non-

organic psychosis, most agree on the latter two, duration and severe dysfunction (Ruggeri

et al., 2000; Schinnar, Rothbard, Kanter, & Jung, 1990). In this research, individuals

with severe and persistent mental illness have, according to their records, been involved

with the mental health system for at least an accumulated two year period of time and

have a current GAF score of 50 or less (American Psychiatric Association & American

Psychiatric Association Task Force on DSM-IV, 2000).

A guardian is a person who has been appointed by the court to act on behalf of a

person (Adult Protective Services Act, 2002).

A voluntary admission to a long-term care psychiatric facility consists

of any person, age 18 years and older, who makes an application for admission by

express and informed verbal or written consent, is found to show evidence of mental

illness, is competent to provide express and informed consent, and is determined to be

suitable for treatment (Florida Mental Health Act, 2002).

An involuntary admission to a long-term care psychiatric facility usually occurs if

the person meets criteria for involuntary examination (Florida Mental Health Act, 2002).

Criteria for involuntary examination can be summarized as follows: a person who is








mentally ill and who does not receive treatment is likely to pose a real and present danger

to her/himself and/or others (Florida Mental Health Act, 2002).

Risk manager is a professional who is responsible for the process of making and

carrying out decisions that will promote quality care, maintain a safe environment, and

preserve human and financial resources in healthcare organizations (American Society of

Healthcare Risk Management [ASHRM], 1997-2000).

Assumptions

This research is based on several assumptions. First, it is assumed that there is

variability in the types of illness, the severity levels, and the duration of illness and level

of functioning among residents living in a long-term state psychiatric treatment facility.

Second, it is assumed that persons who reside at any given time in a long-term state

psychiatric treatment facility may experience various types of illness and levels of

severity of mental disorders, physical illnesses, affective states, and cognitive

functioning. It is also assumed that the impact of recent life events, as well as variations

in the type and amount of psychotropic medications that are prescribed influence their

overall level of functioning (Atkinson, Zibin, & Chuang, 1997). Third, based on an

extensive review of fall-related literature, it has been determined that there are both

internal (delusions, personal stimuli such as auditory or visual hallucinations, poor

eyesight, seizures, urinary tract infections) and external (slippery floors, dimly lit

hallways, improper footwear) factors that may affect the level of functioning and, at any

given time, the potential for psychiatric residents to fall (Morse, 1997a, 1997b;

Tideiksaar, 1996, 1998).








Limitations

Identification of residents who have fallen is dependent upon accurate reporting

by all staff witnessing the event. At Wellspring, when a resident falls, an incident report

is promptly completed and forwarded within 24 hours to the risk manager's office as they

occur. The report also includes information regarding antecedent events and related

outcomes. In turn, the risk manager completes a Risk Incident Review Data Form. Due

to subjective interpretation of what constitutes a fall, actual fall incidents may be

underreported or described without the benefit of predetermined parameters or guidelines

(Morse, 1997a, 1997b; Tideiksaar, 1996, 1998).

Exclusion criteria of this study prohibit the involvement of subjects who are non-

English speakers as well as those who are unable to participate because of evidence of a

developmental disability, a cognitive disability such as dementia, or manifestation of

florid psychotic symptoms. Although these individuals are excluded from participating

in this study, they are perhaps the most vulnerable of the psychiatric residential treatment

facility's population. For this reason, the researcher will keep track of all persons

meeting exclusion criteria so that they are referred to Wellspring's professional and direct

care staff for clinical risk assessment and treatment planning.

This research occurred in one facility located in the Southeast United States,

limiting the generalizability of our findings because of regional variations in population

served, approaches to treatment, and specific outcome measures that are emphasized in

the numerous settings. Therefore, this study serves to provide the springboard for more

research in an area that is hampered by a lack of intellectual inquiry. Findings from this

research could stimulate the design of longitudinal studies for fall rate analysis or





17


research that determines whether the level of caregiver burden and stress results in an

increased number of falls. Perhaps the first step would be to develop context-specific

interventions as part of a fall prevention program that would be appropriate to meet the

specialized needs of individuals who reside in a long-term inpatient psychiatric setting.














CHAPTER 2
REVIEW OF THE LITERATURE

The purpose of this chapter is to present pertinent research and theoretical

knowledge about persons who have fallen in long-term psychiatric residential facilities.

Four areas that will be the focus of this discussion are presented to galvanize the eclectic

conceptual framework. Included are a conceptualization of falls, including the nature and

severity of falls, risk factors associated with falls, the consequences of falls, and the

caregivers' perceived level of stress when caring for individuals who have fallen in a

long-term psychiatric residential facility. Caregivers' perceptions regarding residents'

met and unmet needs and risk factors are integrated throughout this review. These

domains will be reviewed and related to the conceptual framework.

Conceptualization of Falls

Definitions for fall events, persons who are at risk for falling, and persons who

fall (or do not fall) abound. For example, a fall can be conceptualized as the sudden

unintentional change in position causing one to land on a lower level. This definition

does not include near falls or incidents due to an overwhelming external force (Baker,

1997a, 1997b). Maki and Femie (1996) define a fall as any occasion on which the body

drops unintentionally to the floor or ground or to some other lower level. Morse (1997b)

defines a fall to be a sudden, uncontrolled, unintentional, nonpurposeful downward

displacement of the body to the floor or ground and/or hitting another object like a chair

or stair. This definition of a fall excludes those incidents resulting from (1) major








intrinsic events, such as stroke, heart attack, seizures, or psychotropic medication-induced

postural hypotension and/or gait disturbances; (2) major extrinsic events, such as being

pushed or knocked down by someone or something; or (3) a controlled or intentional

movement to bed or floor resulting in an interruption of the fall, known as a "near fall."

An example of a "near fall"is when a person is caught by staff person before hitting

the floor/ground (Gryfe, Amies, & Ashley, 1977; Mayo, Korner-Bitensky, Becker, & P.,

1989; Mion et al., 1989; Nevitt, Cummings, Kidd, & Black, 1989; Tinetti & Speechley,

1989; Tinetti et al., 1986).

Types of Falls

Falls are classified according to circumstances surrounding their occurrences. For

example, cluster falls are a series of falls occurring within days or weeks preceding death

that are usually nonpreventable. According to Gryfe, Amies, and Ashley (1977), 64.7%

of patients who had six or more falls died within a year; of these, nearly half had a

clustering of falls immediately preceding death (Gryfe et al., 1977; Lamb, Miller, &

Hernandez, 1987). Premonitory falls or pathological falls are due to acute illness such as

myocardial infarction, cerebral vascular accident, gastrointestinal bleeding, or infection.

Typically, they cannot be halted by standard fall prevention interventions, but rather by

medical management of underlying causes (Barclay, 1988; Rubenstein, Robbins,

Josephson, Schulman, & Osterwell, 1990; Tideiksaar & Kay, 1986). Although current

research does not provide accurate data about the number of falls that are thought to be

attributable to acute disease processes (Barclay, 1988; Rubenstein et al., 1990; Tideiksaar

& Kay, 1986), Vassallo and Sharma (1998) assert that approximately 10% of falls

unrelated to syncope are related to acute illness, such as pneumonia, stroke, anemia, and

dehydration.








Accidental falls are falls caused by extrinsic risk factors within the environment

and arise from circumstances which could cause a fit person to fall (Campbell, Reinken,

Allan, & Martinez, 1981; Mitchell, 1984; Morse, 1997a). These falls represent 24.5% of

falls in hospitals (Morse, Black, Oberle, & Donahue, 1989). Most at risk for injury are

the young elderly, ages 55 to 65, who are able to ambulate on their own, and whose falls

are caused by environmental hazards and who are susceptible to osteoporosis.

Pattern falls, also known as weak spells or drop attacks, are disorders of balance

or postural instability. More of a functional disability, these falls, which are generally

due to health-related causes, affect the older elderly with temporary loss of control with

or without loss of consciousness (Campbell et al., 1981; Morse et al., 1989; Whedon &

Shedd, 1989). This classification of falls represents approximately 13.6% of all falls in

hospitals (Morse et al., 1989).

Anticipated physiological falls are related to age and functional ability.

Deficiencies in current mental status, problems with ambulation, and sensory deficits

increase the risk for falls in this grouping. These falls represent approximately 61.9% of

falls in hospitals. Anticipated falls often can be prevented. Prevention strategies include

administration of reliable fall risk assessment instruments that provide information about

physiological and cognitive deficits associated with the resident at risk for falls and

implementing interventions that alleviate or reduce these deficits (Morse, 1997a, 1997b).

Risk Factors Associated With Falls

Common risk factors leading to falls include environmental conditions, alterations

in sensory perceptions, changes in body control (balance and gait), and medication use.

Medications implicated in increased falls include most psychotropic and cardiovascular

medications. Among some of the psychotropic medications that might be implicated in








falls include (1) diazepam (Valium@), (2) alprazolam (Xanaxc), (3) haldoperidol

(Haldol), (4) lorazepam (Ativan), (5) doxepin (Sinequanc), (6) amitriptyline (Elavil'),

(7) imipramine (Tofranil), and (8) desipramine (Norpramin) (Cameron, 1997a, 1997b;

Ray, Griffin, & Downey, 1989; Ray, Griffin, & Malcolm, 1991; Tack, Ulrich, & Kehr,

1987; Tinetti, 1994a, 1994b). The cardiovascular medications that might be implicated in

falls include (1) diltiazem (Lasix0), (2) methyldopa (Aldomet), (3) prazosin

(Minipress), (4) propanolol (Inderal), (5) atenolol (Tenormin), and (6) glyceryl

trinitrate (Nitroglycerin) (Cameron, 1997a, 1997b; Ray et al., 1989; Ray et al., 1991;

Tack et al., 1987; Tinetti, 1994a, 1994b).

Individuals with severe and persistent mental illness who are commonly

prescribed a variety ofpsychotropic medications to control florid psychotic symptoms

may experience rapid changes in blood pressure, sensory perceptions, and body control.

Similarly, these same individuals may also receive a number of medications to control

cardiovascular disorders such as hypertension, cardiac insufficiency, coronary artery

disease, and a myriad of other cardiovascular conditions. Among these residents are

documented incidences of endocrine disorders such as diabetes and hypothyroidism

(Campbell, 1991; Cumming, 1998; Leipzig, Cumming, & Tinetti, 1999a, 1999b). These

medications and the conditions they treat require close monitoring to prevent serious

alterations in the resident's physiologic and, sometimes, mental status. Therefore, it has

been surmised that individuals with severe and persistent mental illness are at increased

risk for falls in long-term care settings because of their complex mental and physical

status (Campbell, 1991; Cumming, 1998; Leipzig et al., 1999a, 1999b).








Consequences of Falls

Falls can have physical, psychosocial, economic, and legal consequences. These

four domains will be briefly discussed in this next section.

Physical Consequences

Falls can result in physical injuries, disability, and death. Consequences from

falls range from (1) no injury, (2) minor injury (requiring first aid or minor treatment), or

(3) serious injury (fall resulting in fracture, emergency room monitoring for more than 24

hours, admission to an acute-care hospital, bedrest for more than 48 hours, or restricted

activity for more than 72 hours) (Tinetti, 1987). Approximately 30 to 55% of people who

fall suffer minor injuries, 4 to 6% sustain fractures, 2 to 20% have injuries severe enough

to require hospitalization, and 2.2% die as a result of a fall-related injury (Nevitt,

Cummings, & Hudes, 1991; O'Loughlin, Robitaille, Boivin, & Suissa, 1993; Sattin, 1992;

Tinetti & Williams, 1997). The risk of major injury is greatest with falls associated with

loss of consciousness, as compared with nonsyncopal falls, where the individual does not

experience a loss of consciousness or dizziness (Nevitt et al., 1991).

Hip fractures constitute one of the most serious injuries resulting from falls.

Annually, more than 233,000 hip fractures occur in older people, and more than 13,000

deaths occur as a result of these fractures (Sattin, 1992). The incidence of hip fracture

increases exponentially by age, doubling every five years after the age of 50 years

(Grisso & Kaplan, 1994).

Often, falls are markers for an underlying acute disease or diseases (Tideiksaar,

1998). In addition, a clustering of falls often heralds a general physical decline that is

attributable to chronic disease, such as dementia, neuropathy, musculoskeletal, and visual








disorders (Tideiksaar, 1998). Indeed, the risk of long-term institutionalization and death

is high for older adults who experience multiple falls or are hospitalized for injurious falls

(Alexander, Revara, & Wolf, 1992; Dunn, Rudberg, Furner, & Cassel, 1992; Kiel,

O'Sullivan, Teno, & Mor, 1991; Wolinsky, Johnson, & Fitzgerald, 1992). The rate of

death due to falls rises rapidly with increasing age (National Safety Council, 1995).

Residents' perceptions about falls in this domain are not yet clearly understood (Roberts

& Wykle, 1993). Not unlike the resident's perceptions, caregivers' perceptions about

physiologically based falls are also undocumented (Mafullul & Morriss, 2000).

Psychosocial Consequences

Psychosocial reactions to falls may be even more debilitating than some physical

injuries. Falls occurring in the general population can result in fear of falling, depression,

anxiety, loss of confidence, social withdrawal, dependency, and institutionalization

(Arfken, Lach, Birge, & Miller, 1994; Brummel-Smith, 1989; Downton, 1992; Nevitt et

al., 1991). Fear of falling occurs in people who fall as well as those who have never

fallen. Studies indicate that 40 to 73% of people who have fallen and 20 to 60% of those

who have never fallen, fear falling (Maki, Holliday, & Topper, 1991; Nevitt et al., 1989;

Tinetti, 1988). Fear of falling also increases with age and is greater in people with gait

and balance disorders (Arfken et al., 1994). The term, fallophobia, is defined as

psychological damage due to the loss of self esteem and fear of falling again.

Fallophobia can be severely debilitating and can create risks of falling and self-protective

immobility (Tideiksaar & Kay, 1986). This fear commonly coincides with a decrease of

physical activity and muscle weakness that can also lead to falling (Cameron, 1997a,

1997b). Although inconclusive, fallophobia may be related to the post-fall syndrome.








The post-fall syndrome is a transient deterioration in balance and gait after a fall that is

not explained by either neurological or musculoskeletal deficits (Tinetti, 1987). Its origin

has not yet been determined (Cameron, 1997a, 1997b; Tinetti, 1987). The psychosocial

consequences of falls among individuals residing in long-term care facilities could be

anticipated to be greater than in the general population, however no conclusive evidence

or relevant studies currently exist (Tay et al., 2000).

Economic Consequences

High health service use occurs with falls (Kiel et al., 1991; Wolinsky et al., 1992).

Estimated costs of acute care for fall related fractures is $10 billion annually (Sattin,

1992). It has also been estimated that the cost of a fall requiring medical care is $11,800

per hospitalization for individuals between the ages of 65 and 74 (Covington, Maxwell,

& Clancy, 1993). It is important to note that costs depicted by a majority of studies are

conservative figures (Tideiksaar, 1996). Accurate reporting of costs of falls requires

quantitative and qualitative measures to include direct medical and nursing intervention

hours, time spent in reporting and completing forms, resident education, extra health care

expenditures resulting from the fall, and social and psychological impacts on both the

residents and caregivers (Bakarich, McMillan, & Prosser, 1997; Sutton, Standen, &

Wallace, 1994; Tideiksaar, 1996). It is known, however, that caregivers in long-term

psychiatric facilities experience stress and frustration in their efforts to care for

individuals who have fallen in psychiatric institutions (Bakarich et al., 1997; Sutton et al.,

1994). The cost factor is one major reason for caregiver stress (Astrom, Nilsson, Norberg,

& Winblad, 1990; Dunn, Rout, Carson, & Ritter, 1994; Mobily, Maas, Buckwalter, &








Kelley, 1992; Resnick & Baumann, 1988; Romeis, 1989; Weisensee & Kjervik, 1989).

Moreover, litigation is a potential outcome (Hendrich, 1996).

Legal Consequences

Resident falls are responsible for more claims of negligence against hospitals,

long-term care facilities, and nurses than any other type of injury. In general, health care

facilities, long-term psychiatric care facilities, and nurses are not automatically liable for

injuries resulting from resident falls. Rather, liability involves the appropriate standard of

care that nurses should provide to residents in these facilities. That is, falls and fall

injuries are directly associated with a variety of care-related behaviors, frequently

referred to as "failures": failure to monitor residents; failure to use proper

treatment/procedure/performance; failure to ensure resident safety; failure to effectively

respond to the resident; failure to appropriately supervise treatment; and failure to

properly administer medication (Bakarich et al., 1997; Hendrich, 1996; Sutton et al.,

1994).

In an effort to prevent falls, health care professionals have tried to protect

residents by limiting their mobility, often resorting to the use of mechanical or chemical

restraints. However, mobility restrictions and restraint use have proven to be ineffective

(Tideiksaar, 1998). Furthermore, federal regulations governing long-term care in

psychiatric and nursing home facilities actively discourage the use of restraints,

particularly when used for "behavior control" (Berland, Wachtel, Kiel, O'Sullivan, &

Phillips, 1990; Strumpfet al., 1989; United States Senate, 1990; USDHHS, 1989). A

component of the Omnibus Budget Reconciliation Act (OBRA) of 1987, which took

effect in October 1990, declared that nursing home residents have the right to be free








from physical or chemical restraint that is not required to treat specific medical symptoms

(Oswald, Redmond, & Catanzaro, 2001; USDHHS, 1989).

Researchers hypothesize that once fall risk factors have been identified, efforts

can be directed to modify these factors (Morse, 1997a, 1997b; Tideiksaar, 1998;

Tideiksaar & Kay, 1986). One important step in altering risk factors incorporates

"consciousness-raising" of staff and residents, which is achieved through institutionally-

based educational programs for staff and residents. Other suggested strategies to alter the

risk factors of falls are the development of standardized multidisciplinary care plans that

address specific interventions, such as utilization of bed alarms, nonskid slippers, non-

glare lighting, etc. (Gillespie, Gillespie, Cumming, Lamb, & Rowe, 2000; Tinetti, 1994b;

Tinetti et al., 1994; W.K.Kellogg, 1987). However, specific intervention programs that

are not context-based may not be effective for a given population. Factors such as the

perceptions of individuals who have fallen and their willingness to be helped, or the

perceptions of staff regarding the fall risk of residents should be addressed to aide in the

design of context-specific interventions to prevent falls.

The Political Economy Perspective

Many leading theories that attempt to explain problems associated with severe

and persistent mental illness are reductionistic; that is, the onus of the responsibility for

the disability or disease is shifted to the afflicted individual. Similar reductionist

approaches have been used to explain the plights of other disenfranchised individuals,

such as the homeless, persons diagnosed with AIDS, and the elderly. Even individuals

who have fallen are required to assume the responsibility for falling (Tideiksaar, 1998).

Hence, reductionistic thinking focuses blame on individuals for their problems rather than








explaining problems in terms of inequities in resource distribution and access (Estes et

al., 1996), or societal and health professionals' attitudes about these individuals with

specialized needs.

In contrast to the micro theory approach, macro theory, such as Political Economy

Theory, considers relationships among social structure, social processes, and social

psychological states for enhancing the understanding of a particular phenomenon (Estes

et al., 1996; Tickner, 1992). Estes and colleagues' (1996) theoretical framework, the

"Political Economy of Aging," postulates that aging is embedded in the culture and

society in which it occurs, and it cannot be defined and analyzed as a separate, isolated

entity. Therefore, socially constructed perceptions of aging and the aged influence the

development of social policy. Several theorists contend that the political economy of

aging is interwoven in an intricate relationship with the government structure, fiscal

resources and the labor market (Myles, 1984; Myles & Quadagno, 1991). This notion

has significant consequences when considering the social policies that create a division of

labor and inequitable compensation based on gender, ethnicity, disability, and other

forms of social stratification (Tickner, 1992). Therefore, gender, ethnicity, social status,

and other forms of social stratification have a direct effect on the allocation of resources.

This rationing and allocating scarce resources is applied to other marginalized and

disenfranchised populations such as the severe and persistent mentally ill. Hence, the

political economy of aging has practical application when examining problems associated

with the severe and persistent mentally ill.








Caregiver Stress

The concept stress, most frequently identified with Selye (1976), generally refers

to the physiologic response of an organism that results from a particular stressor. Human

responses to stress are usually mediated by the individual's personality, coping skills,

support systems, and level of insight (Astrom et al., 1990; Dunn et al., 1994; Heine,

1986; Mobily et al., 1992; Resnick & Baumann, 1988; Romeis, 1989; Weisensee &

Kjervik, 1989).

Caregiver stress, often used synonymously with the term burnout, has been

identified as the chronic emotional strain of dealing extensively with humans who are

troubled or are having problems (Maslach & Leiter, 1997). Indeed, persons with severe

and persistent mental illness or organic brain disorder are often characterized as being

uncooperative and aggressive, as exhibiting frequent mood swings and insomnia, and as

being unable to adequately communicate their needs (Astrom et al., 1990; Dunn et al.,

1994; Heine, 1986; Mobily et al., 1992; Resnick & Baumann, 1988; Romeis, 1989;

Weisensee & Kjervik, 1989). In addition, persons with severe and persistent mental

illness are often perceived as being incapable of independent living, thereby,

necessitating frequent hospitalizations or the procurement of supervised living facilities

(Astrom et al., 1990; Dunn et al., 1994; Heine, 1986; Mobily et al., 1992; Resnick &

Baumann, 1988; Romeis, 1989; Weisensee & Kjervik, 1989).

The manner in which caregivers respond to the stressors of the occupation

depends on their own personality traits, what they want out of the job, their past

experiences, and the quality of their lives outside of the workplace (Astrom et al., 1990;

Dunn et al., 1994; Heine, 1986; Mobily et al., 1992; Moore & Cooper, 1996; Romeis,








1989; Weisensee & Kjervik, 1989). When they are experiencing job-related stress, or

burnout, patient care may be jeopardized. Indeed, theorists suggest that caregivers'

behaviors always have an impact on the individuals who are receiving care (Peplau,

1952, 1989). Others found that caregiver stress and burnout contribute to higher rates of

maladaptive incidents on inpatient units, such as acts of aggression, elopements, and

exacerbation of psychiatric symptoms, which, in turn, create even more stress in the

caregiver (Goodykoontz & Herrick, 1990).

In an effort to explain determinants of behavior under certain conditions, social

learning theorists have developed the concepts locus of control and self-efficacy

(Bandura, 1977; Rotter, 1954; Sherer & Adams, 1983; Sherer et al., 1982). Whereas

locus of control focuses on the extent to which one believes one's behavior controls

outcomes, self-efficacy is concerned with the confidence one has in her/his ability to

perform certain behaviors (Bandura, 1986). Although many researchers have

demonstrated a positive correlation between internal locus of control and healthy lifestyle

behaviors, little evidence supports the notion that locus of control is associated with

caregiver stress/burnout (Gueritault-Chalvin, Kalichman, Demi, & Peterson, 2000).

Bandura (1977) hypothesized that self-efficacy affects choice of activities, effort

and persistence. This hypothesis has been widely supported in the literature (Hackett &

Betz, 1995; Schunk, 1995). In addition, (O'Leary & Brown, 1995) found that self-

efficacy does mediate the stress response, at least on a physiological level.

Unfortunately, most research focusing on self-efficacy and its relationship to stress in

caregivers deals primarily with family members rather than caregivers who are employed

in long term care settings (Schmall, 1995).














CHAPTER 3
METHODOLOGY

The purpose of this chapter is to describe the research methods that are used to

address the research hypotheses. Specifically, the design, data collection procedures, the

instruments, the statistical analyses, and human subjects concerns are discussed.

Research Design

A descriptive correlational design was used in this study to describe two groups of

individuals residing in a state-owned long-term mental health treatment facility: those

who had experienced at least one fall within the past 12 months, and those who had no

documented falls during the past 12 months. The individuals in this study had

documented diagnoses of mental illness as determined by criteria in the Diagnostic and

Statistical Manual IV-TR (American Psychiatric Association & American Psychiatric

Association Task Force on DSM-IV, 2000). Variables such as demographic data,

psychiatric level of functioning, as well as the number and types of psychotropic

medications prescribed were delineated. In addition, they were asked to provide

information regarding their perceived risk factors. Institutional caregivers were also

approached to provide information regarding their assessment of the residents' risk

factors and their experienced caregiving-related stress. Finally, relationships between

and among the variables were examined.

All data were collected from individuals who had fallen and those who had no

documented history of having fallen while residing in the state-owned long term care








mental health treatment facility, Wellspring. In addition, data were collected from direct

caregivers who were responsible for the care of the residents. Hence, three populations,

residents who had fallen, residents who had no documented history of having fallen, and

their caregivers, were requested to participate in this study.

The Research Setting

Farmer County

Farmer2 County, with a population of approximately 20,000, has an area of about

585 square miles with only about ten percent of the county being developed.

Approximately 33% of Farmer County is either national forest or has been designated as

wetlands. The remaining 67% is used primarily for growing timber or agricultural

products (Chamber of Commerce, 1995).

The county seat, Tippecanoe,3 is located in close proximity to the licensed state

psychiatric treatment facility, the site of this research. Tippecanoe, with a population of

about 5,000, contributes significantly to the labor force of the state psychiatric treatment

facility. Indeed, opening of the facility in August of 1959 provided employment

opportunities for many of Tippecanoe's residents (Chamber of Commerce, 1995) and

was, according to local townspeople, a significant factor in the growth and prosperity of

Tippecanoe and surrounding communities (Chamber of Commerce, 1995).

The 593-bed, state mental health treatment facility, the largest public mental

health hospital in Florida, employs 1,250 full-time staff. Wellspring's continued

emphasis on treating, supporting, and rehabilitating persistent mentally ill individuals in


2 Farmer County is a fictitious name, however, all other facts associated with it are accurate.
3 Tippecanoe is a fictitious name, however, all other facts associated with it are accurate.








as efficient a manner as possible has facilitated its staff's efforts to serve an estimated

800 people annually. One of Wellspring's major goals is to ensure a speedy return to the

community. Of the majority of residents served, approximately 97% are involuntarily

committed under Florida Statutes Chapter 394, the Florida Mental Health Act (2002), or

Chapter 916, Mentally Deficient and Mentally Ill Defendants (2002). The Florida Mental

Health Act (2002) provides the means by which individuals who are determined to be

mentally ill and pose a danger to themselves or others are committed. Chapter 916,

Mentally Deficient and Mentally Ill Defendants (2002), provides the means by which

individuals who are found incompetent to proceed or are not guilty by reason of insanity

through the judicial process are committed. Within the hospital, a wide variety of

professional and paraprofessional staff provide core services among six residential units,

eight clinical departments and 16 support service departments. Core services include

psychiatric treatment and rehabilitation, health care, vocational programming, behavior

analysis, and community reintegration. Services are delivered on the basis of

recommendations made by the service team, which is comprised of the resident, family

representative, psychiatrist, psychologist, qualified mental health professional (QMHP)

or treatment team coordinator (TTC), nurse, primary care physician, social worker,

dietician, rehabilitation therapist, and resident advocate. Additional discipline

representatives may be consulted for services based upon the recommendations made by

the service team members. For example, the primary care physician may refer the

resident to the physical therapist, occupational therapist, and/or speech pathologist, based

upon the assessment results obtained by team members.








As a component of their treatment, residents who are able to do so are encouraged

to participate in the decision making process as related to their treatment activities and

goals. The resident's strengths and needs are reflected in the treatment goals. At

Wellspring, the residents do actively participate in planning, implementing, and

evaluating their overall care.

Hospital administrators and health care providers of Florida's state mental health

facilities have begun to assimilate information from a variety of sources, both local and

national, that are used to provide "like-comparisons" or benchmarking for a number of

quality indicators. One particular quality indicator that is being addressed is the

occurrence of falls in health-care settings. Research data from numerous sources

provided a backdrop for national comparisons for incidents of falls with a like population

(Aisen, Iverson, Schwalbe, Weaver, & Aisen, 1994; Kilpack, Boehm, Smith, & Mudge,

1991; Nyberg & Gustafson, 1995; Poster et al., 1991; Schmid, 1990).

Individuals Who Have Fallen and Individuals Who Have No Reported History of
Falling and are Residing in a Long-Term State Mental Health Treatment Facility

This research involved individuals who have fallen and individuals who have no

reported history of falling and were, at the time of the research, residing in a long-term

state psychiatric treatment facility located in Northeast Florida, referred to as Wellspring.

This facility is licensed to receive individuals who are 18 years of age and older. At no

time does it admit children and adolescents. The approximate average daily census of

this facility is 558 residents. This calculation is based on an average of all daily censuses

accumulated during a 12-month period.

In 12 months, that is, from January 31, 1998 through December 31, 1998, there

were a total of 543 reported fall incidents. During the month of January, 1999, the Risk








Manager within the facility recorded a total of 31 fall incidents on the Risk Incident

Review Data Form. Hence, for a 13-month period, from January 1, 1998 through

February 1, 1999, there were 574 recorded fall incidents. Within the 13-month period,

there were 94 individuals who had a single incident, whereas 107 residents were

identified as individuals who had fallen repeatedly. That is to say, a total of 201 residents

had fallen during the 13-month period, representing approximately 36% of the total

population at Wellspring. Fall rates for the 13-month period were also computed.

According to Morse (1997a, 1997b) and others (Tideiksaar, 1996), fall rates provide

clinicians with more precise data regarding the success of a particular fall prevention

program, by controlling for fluctuations in the census over a particular period of time.

Therefore, for the purpose of this study, fall rates were computed for three-month

intervals and are expressed in quarterly increments. Quarterly fall rates for the period

January 1, 1998 to March 31, 1999 are presented in Table 1.

Table 1
Quarterly Fall Rates Among Residents at Wellspring from January 1998 to March 1999


QUARTER FALL RATES (per 1,000 bed days)


1st Quarter 1998 2.4

2nd Quarter 1998 2.9

3rd Quarter 1998 3.2

4th Quarter 1998 2.4

1st Quarter 1999 2.8

Source: Quality Management, Wellspring, 1999.








Data presented in this table demonstrates that fall rates at Wellspring have been

consistent over the fifteen-month period of time. In addition, Wellspring's fall rate data

provide evidence that falls are a persistent source of concern that should be addressed by

health professionals.

Published fall rates from a variety of health care settings are presented to provide

comparative data and are delineated in Table 2. These data have been used in this study

for benchmarking a method for comparing Wellspring's fall rates with nationally

published parameters.

Table 2

Research Reported Fall Rates Among Residents in Selected Healthcare Specialty Units
That Were Used as Benchmarking Data at Wellspring


TYPE OF SPECIALTY UNIT FALL RATES (per 1,000 Bed Days)


Neurological Rehabilitation Unit 4.1

Medical-Surgical Unit 4.4 4.7

Psychiatric Unit 2.3 -11.3

Stroke Rehabilitation Unit 15.9

Acute Care Unit 3.8 4.2

Source: Aisen, Iverson, Schwalbe,Weaver, & Aisen, 1994; Kilpack, Boehm, Smith, &
Mudge, 1991; Poster, Pelletier, & Kay, 1991; Nyberg & Gustafson, 1995; and Schmid,
1990.

Data presented in Table 2 demonstrate that stroke rehabilitation units are the settings

where individuals are most likely to fall. These falls are related to movements disorders

associated with ambulation, balance, etc. that are the essential elements in any stroke

rehabilitation program (Aisen et al., 1994; Nyberg & Gustafson, 1995; Poster et al.,








1991). In addition to the stroke rehabilitation unit, psychiatric units potentially have a

higher rate of falls. The wide range of fall rates that are presented in Table 2 can be

explained by the variability in residents' disease states and symptomatology at any given

time. In addition, persons with psychiatric illnesses are seen throughout their lifespan;

that is, there is a wide range of ages during which time an individual may be hospitalized

for a psychiatric illness (Poster et al., 1991).

Rohde, Myers, and Vlahov (1990) concur with these findings. They found that

age specific rates for falls were consistently higher for psychiatry and neuroscience

services (Rohde, Myers, & Vlahov, 1990). One would surmise that individuals

diagnosed with psychiatric or neurological disorders frequently have perceptual and gait

impairments, placing them at greater risk for falls than those persons manifesting other

type of physical impairments (Tay et al., 2000). Similarly, Poster and colleagues (1991)

reported that fall rates increase as age advances in a psychiatric setting. See Table 3.

Table 3

Fall Rates in Psychiatric Settings As Reported By Age

Age Group Fall Rate (per 1,000 Patient Bed Days)


All Ages 4.1

Under 20 2.3

21-59 3.1

60-69 7.7

70 and over 11.3

Source: Poster, Pelletier, & Kay (1991).








Population
Participant Selection

The average daily census of Wellspring is 558 persons. These individuals

represent the potential resident population for this study. Individual residents within this

facility who met the inclusion criteria and were at least 18 years of age were invited to

participate. The inclusion criteria for participating individuals were residents

1. who were willing to participate in the study and sign the informed consent

form;

2. whose Guardian was willing to sign the informed consent form; and

3. who spoke and understood the English language.

The exclusion criteria included individuals who were

1. unable to participate because of evidence of a developmental disability, a

cognitive disability such as dementia, or manifestation of florid psychotic symptoms that

prohibited their involvement in this research;

2. unwilling to sign an informed consent form;

3. were under court order in accordance with Chapter 916, Mentally

Deficient and Mentally Ill Defendants (2002) and whose participation would require legal

approval; and

4. not speakers of the English language, and who might not comprehend

what was being asked of them as participants in this research.

Caregivers who were responsible for the day-to-day health and well-being of the

residents were also asked to participate. The inclusion criteria for participating

caregivers were








1. those who were willing to participate in the study; and

2. those who signed the informed consent form.

Sample Size

According to Rosner (1995, 2000), a sample size of 30 individuals who have

fallen and 30 individuals who have no recorded history of falls, for a total of 60 residents,

is sufficient to address the hypotheses. This is based on a formulation of 80% power, a

critical average effect size of 0.30, at least 10 variables, and a significance level of 0.05

for a two-tailed test. Using a stratified random sample, the total number of residents who

participated in the study was 58. Of these, 28 residents had a recorded history of falls

within the past 12 months and 30 had no recorded history of falls. The total number of

direct care staff who agreed to participate in the study was 20.

Residents' Consent to Participate in Study

After permission to conduct the study was obtained from the facility's

administrator, the researcher requested from the administrator a list of residents whose

names appeared on the Fall Risk Incident Review Form over the past 12 months. The

Risk Incident Review Data Form, an institution-wide mechanism of tracking that

provides an accounting of all incidents occurring in a specific period of time, is compiled

and published at the end of every month. Information contained in this document

includes the name and living area of the resident, the date and time of the incident, the

nature of the incident, and the outcome and/or treatment, as recorded by direct care staff.

Therefore, the Fall Risk Incident Review Form was a vital component necessary for

identifying potential subjects for the study.








Fall incidents that result in particular types of outcomes due to the severity of

injury and/or behavior, or the types of interventions required, are identified by the Risk

Manager. These "Significant Reportable Events" are compiled in a report that is sent to

the Secretary of the Department of Children and Families, State of Florida. Copies are

also maintained in the Administrator's office.

Similarly, the researcher requested from the administrator a list of residents who

have resided in the facility during the past year. Individuals who had no recorded falls

within the past year were selected from this roster, known as the "Daily Census Roster",

an institution-wide mechanism of listing each resident who is currently residing in the

facility. The "Daily Census Roster" lists information about each resident, such as date of

birth, date of admission, county of residence, living area, etc. The "Daily Census Roster"

enabled the researcher to develop a list of individuals who have no recorded history of

falling. These individuals were selected to closely approximate the individuals who have

a recorded history of falls for age and living area. Therefore, the "Daily Census Roster"

was a vital component necessary for identifying potential subjects for the study.

The researcher clearly explained all aspects of the study to the potential

participants. Careful explanation, in simple terms, included the purpose of the study, the

time that it would take for the resident to participate in the interviews, the location of the

interviews, and how the data collected would be used in this study. After the residents

had indicated that they understood what they were being asked to do, they were asked to

sign the informed consent. The Phase I interview took approximately 15 to 20 minutes,

and the Phase II interview averaged about 30 minutes in duration. Both Phase I and








Phase II interviews took place in a private, comfortable, pre-designated space of the

residents' living area at Wellspring.

A similar procedure was used for those residents who had been appointed a

Guardian by the courts. Under Florida Statutes 394 and 415, residents who have been

appointed a Guardian by the Courts are considered to lack capacity to consent. That is, a

vulnerable adult does not have sufficient understanding to make and communicate

responsible decisions regarding their person or property (Adult Protective Services Act,

2002; Mental Health Act, 2002). However, these residents also receive quality care

through the use of a Guardian who speaks on their behalf and makes clinical decisions in

the best interest of these residents. One Guardian was contacted and asked for their

permission on behalf of the residents' participation in this study. The researcher mailed a

package of information, to include a self-addressed stamped envelope, to the Guardian.

After the Guardian had been given sufficient opportunity to review these materials, the

researcher telephoned the Guardian to set up an appointment for the purpose of

discussing this research. The researcher carefully explained, in simple terms, the purpose

of the study, the time that it would take for the resident to participate in the interviews,

the location of the interviews, and how the data collected would be used in this study.

The Guardian was also advised that the resident would be involved in the research only if

the Guardian agreed with all aspects of the research and demonstrated this agreement by

signing the Informed Consent, which was included in the packet of information. Since

the Guardian did not return the signed informed consent to the researcher, the resident

was not involved in the study.








Data Collection Procedures

Phase I Records Review and Data Retrieval

This phase of the research was comprised of two components, (1) retrieval of data

from the residents' clinical records; and (2) a face-to-face ten-minute interview.

From the residents' records, the researcher retrieved information that was used to

complete the (1) Demographic Data Form, the (2) Medication Profile Data Form, and the

(3) Structured Clinical Interview for Positive and Negative Symptoms Scale (SCI-

PANSS) (Kay, Fiszbein, Lindenmayer, & Opler, 1986; Kay, Fiszbein, & Opler, 1986,

1987; Kay, Opler, & Fiszbein, 1986; Kay et al., 1991). The Demographic Data Form

lists specific information about each resident, such as age, ethnicity, education, and

diagnoses. Diagnoses are determined by means of a multiaxial assessment. The

multiaxial assessment, a comprehensive and systematic format for organizing and

communicating relevant clinical information, is used to plan treatment and predict

outcomes (American Psychiatric Association & American Psychiatric Association Task

Force on DSM-IV, 2000). In addition to the identification of specific mental disorders,

psychosocial and environmental stressors, and level of functioning, general medical

conditions are enumerated for the purpose of ensuring a thorough evaluation of physical

health concerns and holistic treatment planning. The Medication Profile Data Form listed

all medications that the resident was taking, their dosages, reported side effects, and

whether the resident was willing to take the medication. There was a second component

of the Medication Profile Data Form that required the researcher to have a face-to-face

interview with the resident (see Phase I Face-to-Face Interview, "Personal Experiences

With Medications"). The SCI-PANSS, which is also located on the residents' records, is








routinely administered by members of Wellspring's Psychology Department at six month

intervals to all residents who are 64 years of age and younger. The researcher, who has

received extensive training in the administration of the SCI-PANSS instrument and

interpretation of the raw data generated from the SCI-PANSS, retrieved these scores from

the study participants' records. This tool was used to identify those potential participants

who, due to florid, or severe, psychotic symptoms were excluded from this study. Scores

of 25 for any of the factors, F-1 Negative Factor, F-2 Agitation Factor, F-3 Cognitive

Factor, F-4 Positive Factor, or F-5 Depressive/Anxiety Factor, might indicate severe

psychopathology and/or psychotic symptomatology that could deleteriously affect an

individual's ability to participate in the research (H. Reiff, personal communication,

October 25, 2000).

Phase I Face-to-Face Interviews 1 and 2

Interview 1. After recorded demographic and clinical data were extracted from

the residents' treatment records, Phase I face-to-face interviews with these residents

began. The first part of the face-to-face interview was comprised of two instruments,

"SF-36v2 Health Survey" (Stewart, Hays, & Ware, 1988; Stewart, Ware, & Brook,

1977; Stewart, Chen, & Stach, 1998) and "Personal Experiences with Medications."

"SF-36v2 Health Survey" (Stewart et al., 1988; Stewart et al., 1977; Stewart et al.,

1998) is an eleven-item questionnaire the researcher utilized to acquire information about

residents' perceived physical health concerns. In addition to the 11 items on the SF-

36v2 Health Survey, the researcher asked subjects to recall specific events surrounding

their recent fall, such as why they thought the fall occurred and how they felt about

experiencing a fall. Those subjects who had no recorded history of having fallen were








asked whether they remembered an instance when they might have tripped, lost their

balance, etc. If the resident described an incident in which there was a possibility that a

fall may have occurred, the resident was then queried as to their perception about how the

incident occurred. They were also asked to describe how they felt when the incident

occurred. The 11-item SF-36v2 Health Survey and the additional open-ended questions

about the subjects' experiences with falling were administered during the face-to-face

interview and took approximately 10 to 15 minutes to complete. "Personal Experiences

With Medications," involved the researcher asking the residents four basic questions

regarding their medications: the names of the medications the residents were taking, the

dosages, reported side effects, and whether the residents stated that they were willing to

take the medications. This face-to-face interview lasted approximately ten minutes, and

was necessary to determine the individual's basic knowledge about her/his medications.

Interview 2. A second face-to-face interview occurred for individuals in the

sample who were 65 years of age and older. The researcher invited her/him to participate

in the second face-to-face interview. The researcher developed a profile of all residents

who were 65 years of age and older and administered the Clock Drawing Test (Heinik,

Lahav, Drummer, Vainer-Benaiah, & Lin, 2000; Tuokko, Hadjistavropoulos, Miller, &

Beattie, 1992; Tuokko, Kristjansson, & Miller, 1995) and the Folstein Mini-Mental State

Exam (MMSE) (Folstein, Folstein, & McHugh, 1975). These two assessment tools are

standardized clinical and research instruments that are used to determine the degree and

magnitude of cognitive and/or psychiatric impairment (Folstein et al., 1975; Heinik et al.,

2000; Tuokko et al., 1992; Tuokko et al., 1995). Specifically, in this research study, the

assessment scores were used to make clinical decisions about individuals' mental








capacity to participate in this study. Those individuals who had a Clock Drawing Test

score of 6 or less were debriefed, and thanked and were not accepted into the study.

Similarly, those individuals who had MMSE scores of 23 or less were debriefed and

thanked. The remaining subjects were asked to participate in Phase II.

Phase II Face-to-Face Data Collection

Face-to-Face Interview Resident. Those residents whose clinical assessment

data indicated an overall mental capacity to participate in the research were invited to join

Phase II at which time the 30-minute face-to-face interview occurred. The researcher,

who received the signed consent form from the voluntary subjects, made an appointment

with the residents to meet with each of them in a place that was a quiet, comfortable,

predesignated space in their respective living areas. The face-to-face interview consisted

of administering two subscales that comprised the Cardinal Needs Schedule (Marshall,

Hogg, Gath, & Lockwood, 1995) (see Instruments Section for a detailed description).

The face-to-face interviews, Phase I and Phase II, lasted approximately 45 minutes to one

hour. When needed, the resident was given rest periods and the face-to-face interviews

resumed only after the resident had indicated that she/he was willing to continue.

Confidentiality of the subjects was carefully protected.

Face-to-Face Interview Caregiver. After data had been collected from the

residents, the direct caregivers were asked to participate in the study. Direct caregivers

who had resident study participants assigned to their care were approached by the

researcher to provide pertinent information about those particular residents, such as an

assessment of risk factors. In addition, caregivers were asked to provide information

about their experienced caregiving-related stress as well as perceived self-efficacy. It is








important to note that direct caregivers who are assigned the care of specific residents are

knowledgeable about those particular residents in terms of relevant treatment and

rehabilitative issues that can deleteriously affect resident outcomes. It is also important

to note that the same direct caregiver was called upon to provide information on more

than one resident study participant who was assigned to their care at the time of this

study.

Much like the face-to-face interviews with residents, the researcher, who

discussed the research with the caregiver also received a signed consent form from each

of the voluntary direct caregiver subjects. The researcher then made an appointment with

each of the direct caregivers to meet with them on an individual basis in a place that was

quiet, comfortable and private. The face-to-face interview consisted of administering

three data collection instruments, the REHAB scale, The Direct Caregiver Stress

Interview and the Self-Efficacy Scale. The REHAB scale and the Direct Caregiver Stress

Interview are subscales of the Cardinal Needs Schedule. Recall that the same direct

caregiver was asked to complete more than one REHAB and Direct Caregiver Stress

Interview, depending upon the number of resident study participants who were assigned

to their care. Each caregiver was asked to complete one Self-Efficacy Scale.

From previously published data, it has been suggested that these two particular

subscales can be administered in less than 40 minutes. The third instrument, the Self-

Efficacy Scale, is comprised of 30 Likert-type items; two additional open-ended

questions had been added to inquire about how direct care staff prevent falls, and what

type of training they had received regarding the prevention of falls. It was anticipated

that the Self-Efficacy Scale with additional questions would take approximately 10








minutes to complete. In actuality, the two subscales and the Self-Efficacy scale took

direct caregivers approximately a total of 30 minutes to complete. This is especially

important since the researcher did not intend to introduce time constraints that would

impinge on the ability of direct care staff to carry out their required duties and

responsibilities within the unit. Confidentiality of the subjects was carefully protected.

Phase III Records Review Using One Subscale of the Cardinal Needs Schedule

After all resident and direct caregiver interviews had been completed, the

researcher performed a final records review using one subscale of the Cardinal Needs

Schedule, the Additional Information Questionnaire. The Additional Information

Questionnaire was completed following the Resident Opinion and Caregiver Stress

Interviews. Debriefing of residents and direct caregivers occurred following completion

of Phase III. Residents and direct caregivers who participated in this study were given

thank you cards by the researcher as a token of gratitude for their participation.

Instruments Section

Instruments Associated With Phase I: Retrieval of Recorded Demographic and
Clinical Data Extracted From Clinical Records

From the residents' clinical records, the researcher retrieved information that was

used to complete the (1) Demographic Data Form, the (2) Medication Profile Data Form,

and the (3) Positive and Negative Symptoms Scale (SCI-PANSS) (See Appendix A for

Instruments).

The Demographic Data Form. The eleven-item Demographic Data Form

identifies specific data about the residents, such as age, education, and diagnoses. The

researcher acquired this information from the resident's clinical record. (Identified as #1

on the Methodology Diagram, Residents of Wellspring, Appendix B.)








The Medication Profile Data Form. The Medication Profile Data Form

contains six items, and is based on clinical data that are contained within the resident's

treatment record. This form lists the various medications that the resident is currently

taking. It also includes dosages, documented side effects, and whether the resident is

willing to take the medication. (Identified as #2 on the Methodology Diagram, Residents

of Wellspring, Appendix B.)

The Positive and Negative Symptoms Scale (SCI-PANSS). The purpose of

the Structured Clinical Interview for Positive and Negative Symptoms Scale (SCI-

PANSS) was to ascertain the frequency, intensity, and severity of florid psychotic

symptoms that might be evidenced among a population of residents at a facility such as

Wellspring. According to Opler and colleagues (Opler, Caton, Shrout, Dominguez, &

Kass, 1994) and others (Stuart & Sundeen, 1998), a complete assessment of symptom

presentation is crucial when evaluating the level of function and quality of care of

individuals with severe and persistent mental illness. The Structured Clinical Interview

for the Positive and Negative Symptoms Scale (SCI-PANSS), a 30-item interview

format, addresses symptom presentation, such as thought content and form, mood and

affect, the presence of internal stimuli, and somatic concerns. Numerous research studies

have helped to establish the scale's reliability, stability, and validity (Kay, Fiszbein,

Lindenmayer et al., 1986; Kay, Fiszbein, & Opler, 1986; Kay et al., 1987; Kay, Opler et

al., 1986; Kay et al., 1991). The inter-rater reliability is 0.95. At Wellspring, members of

the psychology department routinely administer the SCI-PANSS to residents who are

under the age of 65. Although the SCI-PANSS has repeatedly demonstrated strength as a

psychometric instrument when primarily measuring distinct syndromes in schizophrenia








(Kay, Opler et al., 1986), other researchers have shown it to be effective when evaluating

positive, negative, and cognitive symptoms/factors for mood disorders as well

(Daneluzzo et al., 2002; Purnine, Carey, Maisto, & Carey, 2000). Therefore, the

members of the Psychology Department interview residents to obtain PANSS factor

scores, which are documented on the PANSS Psychological Summary. An analysis of

these five factors, F-l Negative Factor, F-2 Agitation Factor, F-3 Cognitive Factor, F-4

Positive Factor, and F-5 Depressive/Anxiety Factor, can be used to identify the severity

of residents' psychiatric symptoms at a particular point in time. The researcher retrieved

the most current of these scores from the resident's record. In addition to ascertaining the

severity of psychiatric illnesses in resident participants, these data, the scores, were used

to identify those potential participants who, due to florid psychosis, were excluded from

this study. (Identified as #3 on the Methodology Diagram, Residents of Wellspring,

Appendix B.)

Instruments Associated With Phase I: Face-to-Face Interview

SF-36v2 Health Survey_ The SF-36v2 Health Survey, a widely used

instrument that measures self-reported physical health, has been administered

successfully in general population surveys in the United States as well as other countries.

This measure has been demonstrated to be reliable across diverse patient groups (Stewart

et al., 1977; Ware, 2001). Reliability coefficients ranged from 0.65 to 0.94 (McHorney,

Ware, Lu, & Sherbourne, 1994). The content validity of the SF-36v2 has been

compared to that of other widely used generic health surveys such as the Nottingham

Health Profile, the Duke Health Profile, and the Functional Status Questionnaire

(McHorney & Tarlov, 1995). Validity has also been demonstrated when used to compare








the health of 73 subjects with asymptomatic HIV infection and 44 with early AIDS-

related complex (ARC) (Wu et al., 1991). (Identified as #4 on the Methodology

Diagram, Residents of Wellspring, Appendix B.)

Personal Experiences With Medication. Personal Experiences With

Medication incorporated a face-to-face interview with residents who had the opportunity

to discuss with the researcher the names of the medications they are currently taking,

dosages, as well as the experienced side effects, and how they feel about taking their

medications. It is important to note that clinical information obtained from the

Medication Profile Data Form is consistent with routine activities that occur in a clinical

environment and, in this instance, could enhance the care that the residents receive.

(Identified as #5 on the Methodology Diagram, Residents of Wellspring, Appendix B.)

Instruments used to help determine the eligibility of the residents who were 65

years of age or older to participate in the study are the Clock Drawing Test and the

Folstein Mini-Mental State Exam (MMSE). A brief description of the instruments used

in the face-to-face interviews is presented. These instruments were selected because of

several factors. First, they describe and measure the variables of interest in this study.

Second, these instruments have reliability and validity data published in scientific

literature.

The Clock Drawing Test. In contrast to most dementia screening tools that

focus on verbal content, clock drawing relies on visuospatial, constructional, as well as

higher-order cognitive abilities (Heinik et al., 2000; Tuokko et al., 1992). The test

requires merely a sheet of paper and a pencil and can be administered in approximately

five minutes. Test-retest reliability for clock drawing after 12 weeks was 0.78 for








Alzheimer's disease patients (Mendez, Ala, & Underwood, 1992). Tuokko and

colleagues (1995) found similar types of values when Alzheimer's disease patients were

retested after four days. Inter-rater reliability for drawings by elderly normal subjects and

Alzheimer's disease patients was 0.97 and did not differ between clinicians and

nonclinicians (Kozora & Cullum, 1994; Mendez et al., 1992; Rouleau, Salmon, Butters,

Kennedy, & McGuire, 1992; Sunderland et al., 1989; Tuokko et al., 1995). Construct

validity has been demonstrated by correlations with the Mini-Mental State Exam, ranging

from 0.41 to 0.58 and the Global Impression of Neuropsychological Impairment scale

from 0.49 to 0.60 (Kozora & Cullum, 1994; Mendez et al., 1992).

Discriminant validity was demonstrated by Wolf-Klein and associates (1989)

when they differentiated groups of normal elderly subjects, and patient groups with

Alzheimer's disease, multi-infarct dementia, and depression (Wolf-Klein, Silverstone,

Levy, & Brod, 1989). Correct classification in normal subjects was 97%, for

patients with Alzheimers Disease 87%, for multi-infarct dementia 62%, and for

individuals with depression 97% (Wolf-Klein et al., 1989).

Current normative data suggest that scores between seven (7) and ten (10) should

be considered normal, a score of six (6) is borderline, and scores of five (5) or less are

indicative of cognitive impairment (Sunderland et al., 1989; Wolf-Klein et al., 1989).

Hence, residents who achieved a score of six (6) or greater were invited to participate in

the study. Residents with scores of five (5) or less did not meet inclusion criteria;

however, recall that they were referred to nursing and medical supervisory personnel and

identified as individuals with impending "special care" risk factors/needs. (Identified as

#6 on the Methodology Diagram, Residents of Wellspring, Appendix B.)








Folstein Mini-Mental State Exam (MMSE). The MMSE is a major component

of the Cardinal Needs Schedule (CNS). A more detailed discussion of the Mini-Mental

State Examination is included in the Additional Information Questionnaire, a subscale of

the Cardinal Needs Schedule (CNS). (Identified as #7 on the Methodology Diagram,

Residents of Wellspring, Appendix B.)

Instruments Associated with Phase II: Face-To-Face Interviews

The Cardinal Needs Schedule (CNS). The Cardinal Needs Schedule (CNS) is a

comprehensive set of instruments and psychiatric interventions designed to measure

psychiatric and social levels of function (Marshall et al., 1995). It incorporates five

integrated inventories/subscales that generate scores used to identify problems, cardinal

problems, and needs. Once a problem is identified, explicit criteria are provided to

decide whether a need is present: if a need is determined to be present, specific

interventions can be selected. Sixteen identified domains of functioning are integrated

within the five subscales of the Cardinal Needs Schedule: (1) psychosis, (2) side effects

from psychotropic medications, (3) anxiety or depression, (4) self harm or violence, (5)

organic disorder, (6) health, (7) socially embarrassing behavior, (8) drugs and alcohol, (9)

domestic skills, (10) finance and welfare, (11) transport and amenities, (12) literacy, (13)

work, (14) social life, (15) hygiene and dressing, and (16) accommodation. All of these

domains of functioning are of importance; transportation, however, is of less concern

among these study subjects.

Three Levels of Need in the Cardinal Needs Schedule (CNS). In addition to

delineating the 16 domains of functioning, the Cardinal Needs Schedule also assesses

individual need at three stages: (1) identification of problems, (2) identification of








cardinal problems and, (3) identification of needs. These three levels of need have been

associated with the theoretical underpinnings of the five subscales (Marshall et al., 1995).

A brief description of the three stages of need follows.

Stage 1. Identifying problems. Standardized instruments are utilized to

assess the resident participant's performance in the 16 domains of functioning. The

assessment of performance in each domain of functioning is then compared with pre-

established thresholds already incorporated into the CNS. If a resident's performance in

a domain falls below the pre-established threshold, a problem is then identified (Marshall

et al., 1995).

Stage 2. Identifying cardinal problems. The CNS determines whether the

problems identified in Stage 1 should be considered cardinal. Three criteria are used to

identify a cardinal problem, and they are (1) cooperation criteria, (2) caregiver stress

criteria, and (3) severity criteria. The cooperation criteria are based on the resident's

view of the problem and the desire to be helped. The caregiver stress criteria emphasizes

the caregiver's view of the problem. The severity criteria rests on the nature and

enormity of the problem as perceived by the resident (Marshall et al., 1995).

Stage 3. Identifying needs. For each identified cardinal problem a list of

suitable interventions are formulated by the Cardinal Needs Schedule. These three

criteria, when superimposed upon the 16 domains of functioning of the CNS, can be used

to assess quality of care. For example, the cooperation criteria could be applied to

domains where the resident might be offered an opportunity to participate in a specialized

training program, such as vocational rehabilitation. The caregiver stress criteria are

applied in domains where the presenting problems causing stress to the direct caregivers








are identified along with suggested interventions for relieving the stress. The resident's

cooperation need not be present. As a rule, the severity criteria are applied in those

domains where the identified problem can present an imminent or potential risk to the

safety of the resident, caregivers and others (Marshall et al., 1995).

Validity. Researchers who developed the original Cardinal Needs Schedule have

provided evidence of the validity of this approach of measuring patient needs (Marshall

et al., 1995). The hypothesis, "Patients would more likely receive new interventions in

domains of functioning where a need was initially rated, than in domains of functioning

where the patient had no need or no identified cardinal problems" was tested. Marshall

and colleagues (1995) concluded that a subject rated initially as having a need in a

domain of functioning was 3.60 times more likely to receive an intervention in that

domain than a subject not rated as having a need in the same domain. The researchers

reported a 95% confidence interval.

Reliability. There are four studies of the interrater reliability of the CNS

approach to needs assessment. These findings support the initial reports of the

investigators of the CNS. Researchers report an inter-rater reliability of 0.96 or better

(Brewin, Wing, Mangen, Brugha, & MacCarthy, 1987; Marshall et al., 1995).

Subscales of the Cardinal Needs Schedule

The CNS embodies five subscales that are used to provide data about the 16

domains of functioning. They are the Demographic Characteristics Inventory, The

Resident Opinion Interview (ROI), The REHAB Scale, the Direct Caregiver Stress

Interview, and the Additional Information Questionnaire. The Demographic

Characteristics Inventory and the Resident Opinion Interview (ROI) will be addressed in








the next section entitled, "Instruments Administered to Residents." The Direct Caregiver

Stress Interview and the REHAB Scale will be addressed in the next section,

"Instruments Administered to Direct Caregivers." Finally, the Additional Information

Questionnaire will be discussed in the section entitled, "Phase III Records Review Using

the Additional Information Questionnaire." See Appendix A for the CNS subscales.

Instruments Administered to Residents

Demographic Characteristics Inventory. This questionnaire consists of 15

items concerned with medical, psychiatric and social history as reported by the resident

during the interview. (Identified as #8 on the Methodology Diagram, Residents of

Wellspring, Appendix B.)

The Resident Opinion Interview (ROI). This is a semi-structured, 20-item

interview designed to evoke the perceptions of the residents regarding quality of care,

whether they wish to change the living environment and if they are concerned about any

current physical problems. (Identified as #9 on the Methodology Diagram, Residents of

Wellspring, Appendix B.)

Instruments Administered to Direct Caregivers

The REHAB Scale. After all face-to-face interviews with Wellspring residents

were completed, the researcher began interviewing the direct caregivers. The REHAB

Scale, developed by Baker and Hall (1988), was incorporated into the CNS because of its

specificity related to the assessment of resident needs based on the severity of

symptomatology. This 23-item instrument is specifically designed to assess people with

disabling chronic psychiatric or medical illnesses (Baker & Hall, 1988). Seven of the 23

items address deviant behaviors. The remaining 16 items address deficits in social and








community functioning/activities, speech, and self-care. Inter-rater reliability ranges

from 0.61 to 0.92 (Baker & Hall, 1988). (Identified as #10 on the Methodology Diagram,

Caregivers of Wellspring, Appendix B.).

The Direct Caregiver Stress Interview. The Direct Caregiver Stress

Interview is a semi-structured interview that determines whether the direct caregiver

perceives the resident's level of functioning, presence of psychiatric symptoms, or the

amount of time required to provide care as causing considerable stress. (Identified as #11

on the Methodology Diagram, Caregivers of Wellspring, Appendix B.)

Phase III Records Review Using the Additional Information Questionnaire

The Additional Information Questionnaire. This 18-item questionnaire

addresses eight domains of functioning, including psychotic symptoms, self-harm or

violence, organicity, health, social functioning, finances, employment, and leisure

activities. A particular domain that is explored in the Additional Information

Questionnaire is the presence of an organic disorder. The Folstein Mini-Mental State

Examination is utilized to validate the existence of an organic disorder. It serves two

basic purposes in this study: (1) To determine if the resident's current mental status and

her/his overall mental performance will indicate participation in this study; and (2) To

provide essential clinical data that serves as a backdrop for understanding the 16 domains

of functioning (Marshall et al., 1995). The Additional Information Questionnaire is

administered during Phase III of the data collection process, and after completion of the

other four subscales of the CNS.

The Folstein Mini-Mental State Examination (MMSE) is a comprehensive

instrument that deserves special attention. It measures recall, orientation, attention,








calculation, registration, language, and ability to copy a drawn figure (Anthony,

LeResche, Niaz, von Korff, & Folstein, 1982; Ashford, Kolm, Colliver, Bekian, & Hsu,

1989; Brandt, Folstein, & Folstein, 1988; Folstein et al., 1975). It can be administered in

less than 10 minutes; additional time is required if the person is being tested for

pronounced impairments (Ashford et al., 1989; Brandt et al., 1988). Internal consistency

for a mixed group of medical patients was 0.96 (Foreman, 1987; Jorm, Scott, Henderson,

& Kay, 1988). Inter-rater reliability was 0.65 (Folstein et al., 1975; Giordani et al., 1990;

O'Connor et al., 1989a). Test-retest reliability for intervals of less than two months

generally fell between 0.80 and 0.95 (Folstein et al., 1975; Giordani et al., 1990;

O'Connor et al., 1989a). The advantage of this test is that it clearly tests short-term

memory by asking the resident to recall three previously identified words. It also tests

reading and writing ability and the aptitude to copy a design and follow a three-step

command. A score of 23 points or less (out of 30) for a person with more than 8 years of

formal education is indicative of cognitive impairment. Scores increase with educational

level and thus lower scores must take into account the extent of formal schooling (Crum,

Anthony, Bassett, & Folstein, 1993; O'Connor et al., 1989a; O'Connor, Pollitt, Treasure,

Brook, & Reiss, 1989b). (Identified as #12 on the Methodology Diagram, Residents of

Wellspring, Appendix B.)

The Cardinal Needs Schedule incorporates an integrated, multitrait assessment

approach: the CNS assesses the level of functioning and quality of care of individuals

with psychiatric diagnoses by integrating data acquired from five subscales. The five

subscales are:








Demographic Characteristics Inventory

Resident Opinion Interview (ROI)

REHAB Scale

Direct Caregiver Stress Interview

Additional Information Questionnaire

Britain's Medical Research Council, in an effort to improve the level of

functioning and the quality of care of individuals with psychiatric disorders, developed a

Needs for Care Assessment in Oxford, England (Marshall et al., 1995). The Council's

research efforts directed the inception of the Cardinal Needs Schedule, which has been

evaluated as one of the most comprehensive and multi-faceted needs assessment tools

currently available to practitioners and researchers (Brewin et al., 1987; Marshall et al.,

1995).

Health and mental health care planning for persons residing in long-term

psychiatric treatment facilities is dependent upon some appraisal of their need for care

and the quality of care delivered. The Cardinal Needs Schedule (CNS) has been used

primarily in England, France, and Italy in community settings. Information gleaned from

using a research tool in a different country, such as the United States, as well as within

the context of a long-term state mental health hospital, will provide researchers with

substantive information regarding the use of the instrument for the determination of needs

in a variety of (1) cultural settings; (2) circumstances where individuals are at different

stages of illness severity and duration; and (3) inpatient and outpatient settings. Hence, it

will provide numerous opportunities for comparison data to be generated and used in

planning and evaluating care in local and global communities. Significantly, data








generated from this study could substantially inform researchers and clinicians because it

has the potential for strengthening the linkage or connectedness to hospital and

community-based care. More data might help to reduce the recidivism rates among the

severely and persistently mentally ill (Brewin et al., 1987; Marshall et al., 1995). The

researcher has received extensive training in the administration, scoring and analysis of

the CNS by developers of the CNS.

The Self-Efficacy Scale (SES). The Self-Efficacy (SE) Scale was used to

measure direct caregiver perceived self-efficacy (Sherer et al., 1982). The 30-item SE

instrument is composed of two subscales, the General Self-Efficacy (GSE) subscale and

the Social Self-Efficacy (SSE) subscale. The 17-item GSE subscale measures caregivers'

perceptions of general competence; scores range from 17 to 85 (Sherer et al., 1982). The

six-item SSE subscale measures caregivers' perceptions of their ability to deal effectively

with others; scores range from 6 to 30 (Sherer et al., 1982). The higher the scores on

both subscales, the higher the perceived self-efficacy.

Cronbach alpha reliability coefficients of 0.86 and 0.71 have been obtained for the

GSE and SSE subscales, respectively (Sherer et al., 1982). In terms of criterion validity,

the GSE subscale predicted past success in vocational, educational, and military areas

(Sherer & Adams, 1983). The SSE subscale was also predictive of past vocational

success (Sherer & Adams, 1983). Construct validity was demonstrated by confirming

predicted relationships with personality measures on the Internal-External Control Scale,

the Marlowe-Crowne Social Desirability Scale, the Ego Strength Scale, the Interpersonal

Competency Scale, and a Self-Esteem Scale (Sherer et al., 1982). (Identified as #13 on

the Methodology Diagram, Caregivers of Wellspring, Appendix B.)








Data Analyses

All data from the five CNS subscales were entered into the Cardinal Needs

Schedule computer program, Autoneed Version 7. Autoneed is a 32-bit, stand-alone

program, written using Microsoft Visual Basic 4 and running under Microsoft Windows

98. Data were stored in Microsoft Access, an industry standard format and were

therefore accessible to SPSS and SAS for statistical analysis.

Autoneed Version 7 was comprised of two components: the program itself and a

database file containing individual resident data obtained from the interview schedules

and the results of the needs analyses. The database file stored information on four key

elements: (1) the composition of the standardized instruments; (2) an inventory of

domains of function to be assessed; (3) the thresholds used to determine the ratings of

each domain of functioning, and (4) the interventions appropriate for each domain of

functioning.

Descriptive statistics were employed to describe the demographic characteristics

of the sample. T-test statistical analysis was performed to identify differences between

the two groups of resident study participants, those who had a recorded history of falls

within the past year and those who had no recorded history of falls within the past year.

Finally, correlation and regression analyses were utilized to address the research

hypotheses.

Regression analyses were also used to estimate the magnitudes of the total (direct

and indirect) effects of age, BMI, gender, perceived and recorded deficits in physical

health, severity of psychiatric symptoms, length of stay, numbers and types of






60

psychotropic medications, number of years of formal education, and Cardinal Risk Factor

scores on the outcome variable, falls.

Protection of Human Subjects

This research involved face-to-face interviews with residents and direct caregivers

of a long-term state mental health treatment facility. Permission to conduct this research

was granted from the Institutional Review Board of the University of Florida Health

Science Center. See Appendix C.














CHAPTER 4
RESULTS AND DISCUSSION

This chapter includes the results of the data analysis and a discussion of the

research findings. The first section provides an overview of the long-term psychiatric

residential treatment facility.

The second section discusses characteristics of both groups of residents, those

who had a recorded history of falls for the previous 12 months and those who did not

have a recorded history of falls. Characteristics such as demographic variables,

psychiatric treatment history, current severity of psychiatric symptoms, psychotropic

medication treatment, and medical status are addressed.

The third section considers the differences between the two groups of resident

research participants, those who had a history of falls for the previous 12 months and

those who did not have a recorded history of falls. The research hypotheses are

addressed in the fourth section and applicability to the proposed theoretical model, an

application of Andersen's Behavior Model for Vulnerable Populations (Andersen, 1995)

is presented. Finally, the fifth section discusses additional findings relevant to the study.

Overview of the Residential Psychiatric Facility Sample

Once approval from the University of Florida Institutional Review Board was

obtained, the researcher submitted the proposal to Wellspring's research committee. This

standing committee, composed of facility staff appointed by the Administrator to oversee








all research projects conducted at Wellspring, met with the researcher to discuss the

proposal. The researcher addressed issues such as purpose, data collection procedures,

and benefit to the facility.

After the residential psychiatric facility's research committee had an opportunity

to review and approve the research proposal, the proposal was then submitted to the

facility administrator for his approval. After the administrator had read the proposal and

signed the final approval form, the researcher arranged a meeting with the six Unit

Treatment and Rehabilitation Directors (UTRDs) who manage the day-to-day activities

of each of the residential units. The purpose of this meeting was to assure the UTRDs

that the residents' treatment activities would not be interrupted as a result of the

residents' participation in this research. In addition, the researcher stressed that direct

caregivers' day-to-day activities would be minimally affected by their participation in

this research.

Prior to beginning the study, the researcher prepared a list of potential research

participants by identifying those persons who sustained a fall during the previous 12

months, October 2001 to November 2002, as indicated on the Fall Incident Report.

Initially, 60 residents who had a recorded history of falls were selected for the study.

Similarly, the researcher reviewed the "Daily Census Roster" to create a list of

individuals who were most similar in demographic characteristics to the proposed study

participants.

Of those residents selected for research participation, 58 residents agreed to meet

with the investigator and provided written consent for inclusion into the study. Of these

58 residents, 30 residents represented those persons who experienced a fall within the








past year, and 28 residents represented those who did not report a fall during the past

year. The 30 residents who experienced a reported fall during the past year were a

representation of 161 persons who sustained recorded falls between the months of

October 2001 to November 2002, which accounted for a total of 422 falls. This meant

that the sample of residents who fell during a 12-month period was approximately 18.6%

of the total number of persons who were recorded as having fallen.

Those remaining 62 residents who were initially selected to participate in the

study did not meet inclusion criteria (59 residents) either because of severe

symptomatology that would prohibit them from participating in the research, or because

they refused to participate. Specifically, 32 residents refused to participate on initial

contact; 22 residents were unable to participate due to florid psychotic symptoms; two

residents had severe dementia and were unable to understand simple instructions; one

resident was identified as forensic, and therefore met exclusion criteria; one resident was

in restraints on multiple occasions and was unable to be interviewed; and one resident

was discharged. In two instances, residents withdrew their consent after agreeing to

participate. One resident was excluded from participating in the study because the

Guardian who was asked to provide consent for a resident never returned the signed

Informed Consent form to the researcher.

Twenty direct caregivers were approached and invited to participate in the

research study. These direct caregivers had been assigned the care of resident study

participants and were, therefore, knowledgeable about these individuals. They agreed to

provide written consent for study participation.








Resident Sample Characteristics

Demographics

Demographic characteristics of the psychiatric residential treatment facility's

residents in this study included gender, age, ethnicity, marital status, and educational

level as indicated in Table 3. Participants included 32 females (55.2%) and 26 males

(44.8%). The mean age of resident participants was 48 (SD = 9.18) for the residents who

had a recorded history of falls and 50 (SD = 10.68) for those residents who had no

recorded histories of falling. Thirteen residents were between the ages of 18 and 40

(22.4%), 29 residents were between the ages of 41 to 55 (50%), and the remaining 16

residents were between the ages of 56 and 70 (27.6%). Three African American

residents (5%), three Hispanic American residents (5%), and 52 Caucasian residents

(90%) agreed to participate in the study.

Approximately 33 residents (56.9%) were single and had never been married, one

(1.7%) was separated, and 17 (29.3%) were divorced. Three residents (5.2%) were

widowed, whereas four (6.9%) residents were currently married.

The highest level of education completed by one resident (1.8%) was an

Associates degree; 13 (23.2%) had taken college coursework but did not attain a degree;

30 (53.6%) obtained a high school diploma; and 12 resident participants (21.4%) did not

complete high school. Demographic data are summarized in Table 4.

In addition to the demographic variables of the resident sample, data on

psychiatric treatment history, current severity of psychiatric symptoms, psychotropic

treatment and medical status were also collected and analyzed. These resident








characteristics were hypothesized to have an effect upon a resident's level of function and

perceived needs (Lehman, Reed, & Possidente, 1998).

Table 4

Demographic Data of Wellspring Resident Sample (n = 58)


Variable n %


Falls

Recorded History of Falls 30 51.7

No Recorded History of Falls 28 48.3

Gender

Female 32 55.2

Male 26 44.8

Ethnicity

African American 3 5.0

Hispanic American 3 5.0

Caucasian 52 90.0

Age

18 to 40 13 22.4

41 to 55 29 50.0

56 to 70 16 27.6








Table 4. Continued


Variable n %



Marital Status

Single 33 56.9

Separated 1 1.7

Divorced 17 29.3

Widowed 3 5.2

Married 4 6.9

Education

Completed 2 year degree 1 1.8

Some college coursework 13 23.2

High school graduate 30 53.6

Completed up to 11th grade 4 7.1

Completed up to 9th grade 7 12.5

Completed up to 6h grade 1 1.8


Psychiatric Treatment History

Psychiatric treatment histories of residents included (1) principal diagnosis; (2)

total number of previous state psychiatric residential treatment hospitalizations, and (3)

length of stay during current residential inpatient admission.

The principal diagnosis was defined as the condition established after clinical

observation to be chiefly responsible for necessitating the admission to the psychiatric

inpatient residential treatment facility. Although the majority of residents had more than








one Axis I or Axis II diagnosis, the first one listed in the chart was considered by the

researcher to be the principal diagnosis. This was consistent with the Diagnostic and

Statistical Manual IV-TR (American Psychiatric Association & American Psychiatric

Association Task Force on DSM-IV, 2000).

Principal diagnoses were coded according to DSM IV criteria and were (1)

Delirium, Dementia, and Amnestic and Other Cognitive Disorders; (2) Mental Disorders

Due to a General Medical Condition Not Elsewhere Classified; (3) Substance-Related

Disorders; (4) Schizophrenia and Other Psychotic Disorders; (5) Mood Disorders; and (6)

Anxiety Disorders (American Psychiatric Association & American Psychiatric

Association Task Force on DSM-IV, 2000). The principal diagnosis for each of the

resident research participants is identified in Table 5.

Table 5

Principal Psychiatric Diagnosis for Resident Sample (n = 58)


Principal Psychiatric Diagnosis n %


Delirium, Dementia, and Amnestic and Other Cognitive 5 8.6
Disorders

Substance-Related Disorders 0 0

Schizophrenia and Other Psychotic Disorders 43 74.1

Mood Disorders 7 12.1

Anxiety Disorders 0 0


The majority of residents (n = 43, 74.1%) had an Axis I diagnosis of

Schizophrenia. The second most frequent diagnosis was Mood Disorders (n = 7, 12.1%).








Recall that many residents had more than one documented psychiatric diagnosis listed in

their clinical record. For example, residents diagnosed with Schizophrenia were also

often diagnosed with a co-occurring substance abuse problem or a co-existing Axis II

(Personality Disorder, Developmental Disability) diagnosis. The focus of this study was

on the primary diagnosis or the Axis I diagnosis.

Twenty residents reportedly had never been hospitalized in a state residential

psychiatric facility prior to the current admission; whereas, one resident reported 17 prior

long-term psychiatric residential hospitalizations. The mean number of prior long-term

psychiatric hospitalizations for the sample was 1.85 (SD = 3.03).

Total average length of stay in months a resident stayed in Wellspring was 52.6

(SD = 53.9). Length of stay during this current hospitalization ranged from one month to

202 months. Table 6 summarizes the number of prior hospitalizations and length of stay.

Table 6

Summary Measures of Number of Prior Hospitalizations and Length of Stay of
Residential Psychiatric Inpatient Facility Resident Sample (n = 58).

Value


Variable Mean SD Minimum Maximum


Number of hospitalizations 1.85 3.03 0 17

Length of stay at Wellspring (months) 52.62 53.91 1 202


Current Severity of Psychiatric Symptoms

One way in which current severity of psychiatric symptoms was measured was by

using the PANSS instrument. Recall that members of Wellspring's Psychology








Department interviewed residents to obtain PANSS factor scores, which were

documented on the PANSS Psychological Summary. An analysis of these five factors,

F-1 Negative Factor, F-2 Agitation Factor, F-3 Cognitive Factor, F-4 Positive Factor, and

F-5 Depressive/Anxiety Factor were used to identify the severity of residents' psychiatric

symptoms at a particular point in time. An analysis of the data revealed that means for

Cognitive and Positive Factors were higher, which was consistent with a documented

primary diagnosis of Schizophrenia, the primary diagnosis for a majority of resident

research participants. Recall that PANSS scores were not available for persons aged 65

and older; that is, three research participants had not been assessed using the PANSS.

There are other methods for determining severity of psychiatric symptoms, one of

which involves using specific computations derived from the REHAB instrument (Baker

& Hall, 1994), a subscale of the Cardinal Needs Schedule. PANSS scores were used to

determine symptom severity since this was consistent with the standards established at

Wellspring. A summary of the current severity of symptoms is addressed in Table 7.

Psychotropic Medication Treatment

Psychiatric medications were viewed as a crucial treatment modality for persons

with severe and persistent mental illness. All 58 residents in the sample were prescribed

at least one psychiatric medication. As indicated in Table 8, 28 residents (48%) were

prescribed four or more different psychiatric medications. The minimum number of

prescribed psychiatric medications was one medication, and the maximum number was

seven medications. Hence, the average number of psychotropic medications prescribed

to residents participating in this research was 3.4 (SD = 1.34).








Table 7

Current Severity of Psychiatric Symptoms (n = 55)

Value


Variable Mean SD Minimum Maximum


F-1: PANSS Negative Symptoms 13.23 4.23 5 22.1

F-2: PANSS Agitation Factor 14.57 5.19 5.7 30

F-3: PANSS Cognitive Factor 15.86 4.50 6 26

F-4: PANSS Positive Factor 16.20 5.44 5 35

F-4: PANSS Depressive/Anxiety 13.07 3.51 7 22
Factor


Psychotropic medications were of primarily four types, or classes: antipsychotic,

antidepressant, mood stabilizing, and anxiolytic. The mean number of types of

psychotropic medication prescribed to resident research participants was 2.62 (SD =

0.88). Frequency and percentage of different types of psychotropic medications

prescribed to resident sample are summarized in Table 8.

It was noted that nine (16%) resident research participants were prescribed more

than four psychotropic medications. This meant that these individuals received more

than one medication in the same class of psychotropic medications. The 30 residents

with a documented history of falls within the past 12 months were prescribed 38

antipsychotics, 32 mood stabilizers, 15 antidepressants, and 21 anxiolytics. The 28

residents with no documented history of falls were prescribed 37 antipsychotics, 23 mood

stabilizers, 10 antidepressants, and 11 anxiolytics.








Table 8

Frequency and Percentage of Different Types of Psychotropic Medications Prescribed to
the Resident Sample (n = 58)


Number Types of
Psychotropic Medications n %


1 4 6.9

2 25 43.1

3 18 31

4 11 19


Medical Status

The concomitant occurrence of medical illness or other health risk factors such as

obesity among persons with severe and persistent mental illness was an ongoing concern

among mental health care providers. Only two residents reported an absence of medical

illness. Eleven residents (19%) reported one medical illness; 9 (16%) reported two

medical illnesses; 9 (16%) reported three; 12 residents (21%) reported four; 7 (12%)

reported five; and 7 (12%) reported six or more medical illnesses. One resident was

documented as having ten medical illnesses requiring complex medical treatment.

Residents reported a variety of medical and health-related concerns. The majority

of residents reported cardiovascular problems (n = 16, 28%), the most common being

hypertension. Diabetes and hypothyroidism were identified among 12 residents (21%)

and 11 residents (19%) had neuromuscular-related disorders, such as Multiple Sclerosis

or Huntington's Disease. Seven residents (12%) had respiratory problems, such as

Chronic Obstructive Pulmonary Disease.








BMI was used to determine obesity. A BMI of 25 to 29.9 was considered

overweight, whereas a BMI of 30 are more was considered obese (National Safety

Council, 1995). Mean BMI for both groups, persons with a recorded history of falls

within the past 12 months, and those with no recorded history of falls was 28.91 (SD =

5.37).

Finally, the residents were asked to provide their perceived health status by means

of responses on the SF-36v2. Their mean score for the SF-36v2 was 64.3 (SD =

19.03). The minimum score was 21.25 and maximum was 98.61.

Differences Between Two Resident Sample Groups

T-test statistics were computed to determine differences between the two resident

sample groups, residents who had a recorded history of falls during the past year, and

residents who had no recorded history of falls. Variables that were considered during this

analysis were age; BMI; Cardinal Risk Factor score (CRS); length of stay; number of

psychotropic medications; number of types of psychotropic medications; years of formal

education; number of medical diagnoses; perceived health status, as measured by the SF-

36v2; PANSS negative symptoms; PANSS agitation symptoms; PANSS cognitive

symptoms; PANSS positive symptoms; and PANSS depressive/anxiety symptoms. See

Table 9.

T-test statistics revealed no significant differences between the groups for age,

BMI, Cardinal Risk Factor scores, LOS, number of years of formal education, number of

types of psychotropic medications prescribed, medical diagnosis, SF-36v2 scores,

PANSS negative symptom scores, PANSS agitation symptom scores, PANSS cognitive

symptom scores, PANSS positive symptom scores, or PANSS depressive/anxiety








symptom scores. There was, however, a significant difference between the two groups

with regard to the number of psychotropic medications prescribed (p < .05). That is, the

group of resident research participants who reported a history of falls was prescribed

more psychotropic medications than those who reported no history of falls. It is

interesting to note that, although not significant, there were differences in BMIs for the

two groups. The group that reported a history of falls had higher average BMIs (mean =

29.83) than those who reported no history of falls (mean = 27.93).

Table 9

Differences Between Two Resident Sample Groups: Persons Who Have a Recorded
History of Falls During the Previous 12 Months and Persons Who Have No Recorded
History of Falls (n = 58)


Variable Fall Mean SD SE Min Max DF T p-value


No 49.96 10.68 2.02 30 69
Age 56 0.94 0.3490
Yes 47.50 9.19 1.68 27 64

No 27.93 6.25 1.18 19 41
BMI 56 -1.36 0.1794
Yes 29.83 4.31 0.79 23 39

No 2.82 1.89 0.36 0 9
CRS 56 -0.23 0.8198
Yes 2.93 1.84 0.34 0 9

No 58.25 62.23 11.76 1 202
LOS 56 0.77 0.4472
Yes 47.37 45.25 8.26 1 157

Number of No 3.04 1.20 0.23 1 6
Meds 56 -2.04 0.0461
Yes 3.73 1.39 0.25 2 7








Table 9. Continued


Variable Fall Mean SD SE Min Max DF T p-value


Number of No 2.43
Types
of Meds Yes 2.80

No 11.62
Education
Yes 11.27

No 2.86
Medical
Diagnosis Yes 3.50

No 67.51
SF-36v2
Yes 61.30

No 13.89
Negative
PANSS Yes 12.68

No 15.01
Agitation
PANSS Yes 14.19

No 15.84
Cognitive
PANSS Yes 15.87

No 16.60
Positive
PANSS Yes 15.87

No 13.56
Depressed
PANSS Yes 12.67


0.96

0.76

1.70

2.12

1.96

1.59

17.26

20.36

4.72

3.77

5.25

5.19

4.74

4.37

4.03

6.44

3.72

3.34


0.18

0.14

0.33

0.39

0.37

0.29

3.26

3.72

0.94

0.69

1.05

0.95

0.95

0.80

0.81

1.18

0.74

0.61


1 4

2 4

8 14

5 13

0 6

1 6

21.5 90.4

21.3 98.6

5 22.1

5.6 20.7

7.9 30

5.7 27.9

6 24

7 26

8 25


5 35

8 22

7 19


56 -1.64 0.1068


54 0.67



56 -1.38



56 1.25



53 1.05



53 0.58



53 -0.02


50 0.51



53 0.94


0.5039



0.1742



0.2168



0.2976



0.5650



0.9828


0.6091



0.3522








Research Hypotheses

The first research hypothesis examined the relationships between and among age,

Cardinal Risk Factor scores (CRS), and the response variable, the incidence of falls.

Based on the results of the Spearman correlation coefficient tests, there was insufficient

evidence to support significant relationships among the variables age, CRS, and

incidences of falls.

The second research hypothesis explored the relationships between and among the

variables BMI, CRS, and the incidence of falls. Based on results of the Spearman

correlation coefficient tests, there was insufficient evidence to suggest significant

relationships among the variables BMI, CRS, and incidence of falls.

The third research hypothesis addressed relationships between and among the

variables perceived and recorded deficits in residents' physical health, CRS, and the

incidence of falls. Findings based on the results of the Spearman correlation coefficient

tests are presented in Table 10.

Based on the results of the Spearman correlation coefficient tests, there were

significant relationships between perceived deficits in physical health and the incidence

of falls (r = -.322, p = .014) and between recorded medical deficits (the number of

medical diagnoses) and the incidence of falls (r = .293, p = .027) at the 0.05 significance

level.








Table 10

Relationships Between and Among the Variables Perceived and Recorded Deficits in
Residents' Physical Health, Cardinal Risk Factor Scores, and the Incidence of Falls Using
Spearman's Correlation Coefficient (r)


Number of
Falls Reported Medical
in Last Year CRS SF-36v2 Diagnosis


Number of
Falls Reported
in Last Year .222 -.322* .293*

CRS .222 -.243 .153

SF-36v2 -.322* -.243 -.021

Medical .293* .153 -.021
Diagnosis

denotes significant finding.

The fourth research hypothesis examined the relationships between and among

the variables severity of mental illness, length of stay (LOS), CRS, and the incidence of

falls. Based on results of the Spearman correlation coefficient tests, there were no

significant relationships between/among the variables severity of mental illness, LOS,

CRS, and incidence of falls.

The fifth hypothesis explored the relationships between and among the number

and types of psychotropic medications, CRS, and the incidence of falls. Based on the

results of the Spearman correlation coefficient tests, there were no significant

relationships between and among the number and types of psychotropic medications,

CRS, and incidence of falls.








The sixth hypothesis addressed gender, a dichotomous variable, and its

relationships between and among CRS and the incidence of falls. Based on the results of

the Spearman correlation coefficient tests, there were no significant relationships between

and among the variables, gender, CRS and the incidence of falls.

Finally, the seventh hypothesis addressed the relationships between and among

the factors number of years of formal education, CRS and the incidence of falls. Based

on the results of the Spearman correlation coefficient tests, there were no significant

relationships found between and among the variables number of years of formal

education, CRS, and the incidence of falls.

A stepwise multiple regression analysis was used resulting in two variables that

significantly explained the incidence of falls in the study: the resident's perceived health

deficits as measured by the SF-36v2 (F = 6.03, p = .0175); and CRS (F = 5.70, p =

.0210). Data also demonstrated that age significantly explained the number of

psychotropic medications residents were prescribed (F = 4.06, p = .0494). Although not

significant at the 0.05 level, the incidence of falls could be explained by the number of

psychotropic medications residents were prescribed (F = 3.05, p = .0869). Finally, data

revealed that female resident research participants had 1.378 times fewer CRS than did

men. That meant that men who participated in this study were identified as having 1.378

times more needs than women.

Resident Experiences With Falls

Those residents with a recorded history of having fallen within the past 12 months

were asked to respond to open-ended queries about what they remembered about their

fall experiences, how they thought the fall occurred, and how it made them feel.








Similarly, residents with no recorded history of having fallen were asked whether

they thought they might have sustained a fall. Residents with a recorded history of falls

usually denied having ever fallen. In fact, 27 of the 30 resident participants who had a

recorded history of falls denied that the fall(s) ever took place. One resident research

participant, Ms. M, has several documented fall incidents occurring within the past year.

When asked about these fall incidences, she stated that she did not really fall: "Oh, I get

like that sometimes." When asked to explain what she meant by "get like that," she

stated,

I feel sick a lot of the time. I never feel right. I just sorta needed to
lay down. It's OK to lay down there [the resident research participant
pointed to the floor in a comer of the living area where the interview
occurred]. I just sometimes feel like laying down. I don't see anything
wrong with that. The staff don't like me laying on the floor. They tell me I
gotta get up. Why can't they just leave me alone? They don't let us lay in our
beds during the day they tell us we gotta be in the Learning Center. When I
feel sick, they [the staff] make me go to see the doctor. You know, the one who
comes here all the time. He don't listen to me anyway. He just looks at me and
talks really fast. I don't always understand what he says.

Ms. M has been diagnosed with numerous medical illnesses, which were being

treated with a variety of medications. According to her clinical record, at the time of her

participation in the research, she was receiving approximately 16 different medications

per day; of these, six were identified as psychotropics. Many of the medications Ms. M

was prescribed required multiple dosing, meaning that they were administered at various

times throughout the day. Based upon Ms. M's responses on the instrument, "Personal

Experiences With Medication," she was unable to list any of the medications she was

currently taking, nor could she identify any experienced side effects. In fact, when asked

to describe what she knew about her medications, Ms. M stated, "I think I take Haldol....

No that's not right. I don't know. It's hard to keep all this stuff straight."








Two of the three residents who did acknowledge that they experienced a fall

within the past year recalled that they had a "spell" at which time they felt weak and

dizzy. Although the two resident research participants could not recall specific events

leading up to the falls, they did attribute the feelings of weakness and dizziness to their

psychotropic medications. One of the two resident research participants stated that

having fallen made her feel "bad" but she did not elaborate.

The majority of respondents with a recorded history of falls within the past year

echoed the sentiments offered by one of the resident research participants who stated,

"I'm alright. There's nothing wrong with me. I don't know who told you that I fell."

One resident, Mr. B, acknowledged that he was involved in a fall, but stated that

someone else was responsible for the fall. He asserted that the fall occurred while staying

overnight at a medical facility, where he was transferred due to a medical emergency

involving his swallowing of batteries. He stated, "The floor was wet and they weren't

watching what they were supposed to be doing. I'm going to sue them." When asked

about a recent fall incident that he experienced within the past month, he stated that he

never fell at Wellspring, despite the fact that an incident report was generated by direct

caregivers who had witnessed the event.

The majority of residents with no recorded histories of falls (n = 24, 86%) denied

having ever fallen. The remaining four individuals stated that they had fallen sometime

in the past, prior to their admission to Wellspring, and that it had contributed to some of

their current medical problems, such as lower back pain, difficulty in walking, etc. One

resident research participant, Ms. Q, who was admitted to Wellspring one month prior to

her participation in this study stated,








I know the reason I fell is because of that awful medicine they
put me on. They don't seem to understand that all I need is Valium. I've
taken Valium for years. Now they want to change it to something else.
They give me all sorts of crap to take. You know, I tripped
and fell when I was staying at the other hospital. I think that's the reason
my back hurts all the time and my legs. See, I can't walk. They want me
to go to the Learning Center, but I don't know how I'll get there. Look,
I can't even make it to the bathroom at night [resident showed researcher
a wastebasket located near the resident's bed containing approximately
two cups of urine.] No one understands or wants to help me. They are
such shits. Can you contact my son? I need some candy and a phone card.
Can you call him for me?

Although an analysis of emergent themes from these scenarios was not within the

scope of this research, further study using a qualitative approach may yield valuable

information regarding perceptions of residents about their fall experiences.

Direct Caregivers

Twenty direct caregivers agreed to participate in this study. Three instruments

were utilized during the face-to-face interviews. Two of the three instruments were

subscales of the Cardinal Needs Schedule, the REHAB and the Caregiver Stress

Interview. The responses from these particular instruments provided specific information

about the resident research participants that they cared for on a day-to-day basis. This

information was integrated with the Cardinal Needs Schedule to produce the Cardinal

Risk Factor score. The Caregiver Stress Interview was of particular interest in that it

asked direct caregivers about their perception of stress when working with residents.

The Caregiver Stress Interview was an 11-item, semi-structured interview. It

assessed direct caregiver stress in response to affective symptoms as expressed by

residents; violent or aggressive acts committed by residents that either placed themselves

or others at risk; physical health concerns; socially embarrassing behaviors; and








residents' needs for assistance with domestic skills, finances, transportation, social

activities, personal hygiene, and dressing.

As indicated in Table 11, issues, which were involved with resident health

concerns and resident's display of embarrassing behaviors, such as incontinence,

swearing, or sexually inappropriate behavior, produced the most stress in direct caregiver

respondents. It was interesting to note that resident research participants' violent

behaviors, either self-directed or directed at others, did not significantly contribute to the

stress levels of direct caregivers.

Table 11

Frequency and Percentage of Caregiver Stress Ratings When Working With Resident
Research Participants (n = 58)


Variable Often % Sometimes % Rarely % Never %


Overall Stress 13 22 26 45 16 28 3 5

Anxiety 12 21 19 33 14 24 13 22

Depression 7 12.1 13 22.4 10 17.2 28 48.3

Violence 2 3.4 11 19.1 10 17.2 35 60.3

Self-Harm 3 5.2 3 5.2 3 5.2 49 84.4

Health 34 59 16 27 0 0 8 14

Embarrassing 30 52 9 16 0 0 19 32
Behaviors

Domestic 22 38 15 26 0 0 21 36
Skills

Self Care 24 41 8 14 0 0 26 45








As indicated in Table 11, 13 direct caregivers responded that 13 (22%) of the

resident research participants frequently created stressful situations. Resident study

participants who displayed symptoms of anxiety and/or depression reportedly produced

varying degrees of stress. Only five residents who exhibited self-harmful or violent

behaviors reportedly caused high stress in direct caregivers. Interestingly, caregivers

responded that they did not believe themselves to be at risk of harm from violence caused

by residents; that is, caregivers' reported stress levels from being injured by residents

were minimal. Yet, there was an overarching concern that "when residents act up, there

aren't enough men staff available to cover the living areas" (Direct caregiver research

participant, personal communication, November 16, 2002). In fact, a number of female

and male staff stated that more male staff members were needed to "keep residents in

line. Some of them get violent or out of control and having a man there helps to keep

things under control" (Direct caregiver research participant, personal communication,

November 16, 2002).

Residents who displayed socially embarrassing behaviors did contribute to the

stress level of caregivers; caregivers reported experiencing stress often based on the

embarrassing behaviors of 30 (52%) resident research participants. Nine resident (16%)

research participants sometimes created stressful situations for caregivers.

The third instrument direct caregivers were asked to complete was the Self-

Efficacy Scale (SES). The SES was used to determine each caregivers' self-efficacy.

Perceived self-efficacy was thought to influence the direct caregiver's perception of

stress. The SES has a minimum average score of one (1) and a maximum average score

of five (5). The higher the score, the greater the person's self-efficacy. The direct care








research participants had a minimum score of 2.57 and a maximum score of 4.87. The

mean was 3.84 (SD = 0.54).

When examining the SES scores in relation to the direct caregivers' primary work

sites, there were some apparent differences. For example, one direct caregiver who was

employed in the medical unit (Living Area 2) had a mean SES score of 4.39. Those

direct caregivers who were employed in the Geriatric living areas (Living Area 5)

reported a mean SES score of 4.02 (SD = 4.45). See Table 12 for direct caregivers' SES

scores.

Table 12

Direct Caregivers SES Scores Based on Living Area


Living Area n Mean SD


Living Area 1 1 3.84

Living Area 2 1 4.39

Living Area 3 4 3.77 0.82

Living Area 4 2 3.43 0.37

Living Area 5 7 4.02 0.45

Living Area 6 5 3.70 0.54

TOTAL 20 3.84 0.54


Direct caregivers reporting the highest SES scores, that is, those staff employed in

Living Areas 2 and 5, were primarily trained as Certified Nursing Assistants, whereas,

the other Living Areas employed direct caregivers who were identified as "mental health

specialists", indicating minimal on-the-job training. Based on the results of the Spearman








correlation coefficient tests, there were no significant relationships between SES scores

and caregiver stress.

When asked about how they prevented falls, direct caregivers primarily focused

on environmental maintenance; that is, they stated that they cleaned up spills, tried to

prevent residents from "running" along the corridors, and kept obstacles and debris out of

the path of residents. When asked whether they had received inservice/training about

falls within the past two years, six direct caregivers (25%) stated that they had received

training in the form of either classes or informal inservice education. All but one of the

respondents reported the two types of training to be helpful. The remaining direct

caregivers (75%) reported that they had never received training in the prevention of falls.

Summary

In this chapter, the results of findings related to the seven hypotheses were

provided. T-test statistics were initially computed to determine differences with regards

to the variables being studied between the two resident sample groups, residents who had

a recorded history of falls during the past year, and residents who had no recorded history

of falls. The researcher found that there was a significant difference between the two

groups with regard to the number of psychotropic medications prescribed (p < .05).

There was evidence to suggest that resident research participants who reported a history

of falls were prescribed more psychotropic medications than those who reported no

history of falls.

To answer the research hypotheses on the relationships between and among

variables, Spearman's correlational coefficient analyses were performed and the results

reported. There was insufficient evidence to indicate a relationship between age, BMI,








severity of psychiatric symptoms, length of stay, the number and types of psychotropic

medications, the number of years of formal education, CRS and the incidence of falls.

Therefore, when examining the first research hypothesis, which explored the

relationships between and among age, Cardinal Risk Factor scores (CRS), and incidence

of falls, the researcher found that there was insufficient evidence to indicate a significant

relationship between and among the variables. Similarly, when answering the second

research hypothesis, which examined relationships between and among the variables

BMI, Cardinal Risk Factor scores, and the incidence of falls, the researcher found that

there was insufficient evidence to indicate a significant relationship between and among

the variables.

To answer the third research hypothesis, which addressed relationships between

and among the variables perceived and recorded deficits in residents' physical health,

CRS, and the incidence of falls, the researcher found that there was a significant negative

relationship between perceived deficits in physical health and the incidence of falls, and a

significant positive relationship between recorded medical deficits (the number of

medical diagnoses) and the incidence of falls.

To answer the fourth research hypothesis, which examined the relationships

between and among the variables severity of mental illness, length of stay, CRS, and the

incidence of falls, the researcher found that there was insufficient evidence to indicate

significant relationships between/among the variables. The fifth hypothesis, which

explored the relationships between and among the number and types of psychotropic

medications, CRS, and the incidence of falls was not supported by the data; that is, there

was insufficient evidence to indicate any significant relationships between and among the








number and types of psychotropic medications, CRS, and incidence of falls. To answer

the sixth hypothesis, which addressed the relationships between/among gender, CRS and

the incidence of falls, there was insufficient evidence to suggest any significant

relationships between and among the variables. The seventh hypothesis which addressed

the relationships between and among the factors number of years of formal education,

CRS and the incidence of falls was unsupported by the data in that there was insufficient

evidence to indicate any significant relationships exist between and among the variables.



A stepwise multiple regression analysis was used resulting in two variables that

significantly explained the incidence of falls in the study: the resident's perceived health

deficits as measured by the SF-36v2 and the CRS. Data also demonstrated that age

significantly explained the number of psychotropic medications residents were

prescribed.

Although not significant at the 0.05 level, the incidence of falls was explained by

the number of psychotropic medications residents were prescribed. Finally, data revealed

that women who participated in this study were identified as having 1.378 times fewer

needs than men, or, conversely, men were identified as having 1.378 times more needs

than women.














CHAPTER 5
DISCUSSION, CONCLUSIONS, PRACTICE IMPLICATIONS, AND
RECOMMENDATIONS

The purpose of this chapter is to provide a discussion about the research findings

and present conclusions. In addition, implications and recommendations for future

research, mental health practice, and mental health policy are presented.

Discussion of Findings

The Administrator of the long-term psychiatric residential treatment facility,

Wellspring, permitted the investigator to conduct the study within the facility. Two

groups of residents were identified, those who had a history of falls within the past 12

months and those without a history of falls. The total number of residents participating in

this study, 58, represented approximately 12% of the total population of the facility.

Twenty direct caregivers who were assigned to care for residents in the research

population agreed to participate in the research. There were no refusals.

Characteristics of the Residents

Many characteristics were examined to determine if there was sufficient evidence

to indicate a relationship among the various characteristics of two groups of residents

living in the long-term psychiatric residential treatment facility. Demographic

characteristics such as gender, age, and number of years of formal education were

analyzed. Additional variables such as psychiatric treatment history, current severity of

psychiatric symptoms, psychotropic treatment, and medical status were also analyzed.








These characteristics were considered to be influential in the incidence of falls in

persons with severe and persistent mental illness. The conceptual framework that suited

this particular study was an adaptation of Andersen's Behavior Model for

Vulnerable Populations (Andersen, 1995). Vulnerable populations are those for whom

the risk of poor physical, psychological, or social health has or is quite likely to become a

reality (Aday, 1993). The model proposed by Andersen (1995) focused on the theoretical

structure of access to health care integrating the elements of predisposing, enabling, and

need of the individual.

Predisposing domain, the first element in the Behavior Model for Vulnerable

Populations, has been examined in the context of the severe and persistent mentally ill

resident. This element included age, gender and the number of years of formal education.

Gender and age represent a biological imperative suggesting that the residents will have

health needs or deficits (Andersen, Rice, & Kominski, 2001). Social factors, another

component of the predisposing domain, consists of elements such as an individual's

education, marital status, which may influence the social network of a person in the

community, and can ultimately facilitate or impede an individual's perceived health care

needs and access to services (Andersen et al., 2001). Hence, an individual's position in

the social structure can influence one's vulnerability (Aday, 1993).

The second element of the Behavior Model for Vulnerable Populations is the

enabling domain (Andersen, 1995). Factors concerning this element are REHAB

indicators, perceived and documented deficits in physical health, body mass index (BMI),

primary caregiver stress, primary caregiver self-efficacy, resident length of stay, and

number and types of psychotropic medications.








Finally, the Behavior Model for Vulnerable Populations has a need domain,

which focuses on severity of psychiatric illness and mental status, and the resident's

perceived risk factors and needs. The presence of needs was determined by integrating

the responses of residents, direct caregivers, and the clinical judgment of the investigator

as guided by the Cardinal Needs Schedule (Marshall et al., 1995) resulting in

identification of 126 needs among the 58 residents. In recent years, the concept of needs

for care has been proposed as an exemplar in planning mental health service interventions

(Anderson & Lyons, 2001; McCrone & Strathdee, 1994; Murray et al., 1996). Though

this approach is considered a potentially useful model, it has not yet been extensively

implemented nor researched.

Based on the measures and analyses used in this study, very few significant

relationships were found among the various characteristics of residents. No significant

relationships were discovered among the variables, age, gender, and number of years of

formal education, BMI, severity of mental illness, length of stay, the number and types of

psychotropic medications, the response variable incidence of falls, and Cardinal Risk

Factor score.

Perceived Deficits in Physical Health

Spearman correlational coefficient analyses demonstrated a significant negative

relationship between perceived deficits in physical health and incidence of falls (p =

.014). That is, scores on the SF-36v2 were higher for resident research participants who

had a recorded history of falls within the past year than for those resident research

participants with no recorded history of falls.








There are several possible explanations for this finding. First, recall that the

occurrence of falls may have serious consequences for both the resident and the

caregivers who are responsible for their resident's care. Falls are commonly believed to

be attributable to individual failure, carelessness or the natural process of aging

(Tideiksaar, 1998). Too, residents may become fearful that they may sustain another fall

or may find themselves restricted to the living area for "their own protection." Many of

the residents at Wellspring who experience falls are identified as being on "fall

precautions" which may involve placing the individual on "Supervised Grounds Access."

Supervised grounds access limits the resident's ability to leave the living area during pre-

designated times in order to go to the canteen, bank, library, boutique, or other outside

recreational areas. The residents must rely on direct caregivers to escort them wherever

they need to go. In addition to having restricted access to outside activities, residents

who are on fall precautions, and therefore supervised grounds access, are also expected to

eat their meals in the satellite dining rooms, which are located within the confines of the

respective units. Restriction of a resident to her/his living area may be perceived as a

form of physical restraint. Indeed, in the psychogeriatric setting, the prevention of

hazards related to falls is often one of the justifications afforded to the use of physical

restraints. In one study examining reasons for restraint use by staff in nursing homes, it

was found that 40% of caregivers used restraints as an intervention against falls (Cohen,

Neufeld, Dunbar, Pflug, & Breuer, 1996). It is a routine practice to restrain

psychogeriatric patients to prevent falls (Waring, 1991).

Those persons who are restricted to their living areas for reasons of safety or

health deficits may perceive themselves to be viewed as socially unfavorable. Being








viewed favorably, or social desirability, is an important characteristic and has been

reported as leading to response bias in mentally ill research participants who complete

self-report measures regarding perceived health status (Bardwell, Ancoli-Israel, &

Dimsdale, 2001). Therefore, residents who are at risk for falling may try to portray their

perceived health status in a more positive light in order to minimize the perceived

consequences of the fall incident. Recall, when asked an open-ended question about their

falls, while living Wellspring, a majority of those residents who had a reported fall within

the past 12 months denied having ever fallen, or stated that someone else was responsible

for their fall.

Although a negative correlation between perceived health deficits and incidence

of falls was established, stepwise multiple regression analysis identified perceived health

deficits as measured by the SF-36v2 (F = 6.03, p = .0175) as significantly explaining the

incidence of falls. Therefore, this is the first factor in the conceptual model adapted from

Andersen's Behavioral Model for Vulnerable Populations (1995) that proposes an

explanatory model related to the occurrence of falls in a long-term psychiatric residential

treatment facility.

Recorded Medical Deficits

A statistically significant relationship also existed between recorded medical

deficits and the incidence of falls (p = .027). Indeed, persons with severe and persistent

mental illness have many debilitating and chronic illnesses, such as hypertension, chronic

obstructive pulmonary disease, poor oral health and dentition, and diabetes (Lawrence &

Coghlan, 2002; Murphy, Gass-Sternas, & Knight, 1995). Acute or chronic illnesses,

Parkinson's Disease, vestibular problems, stroke, incontinence, gait/balance impairment,








postural hypertension, foot problems and sensory impairment are conditions that can lead

to an increased incidence of falls. In fact, up to 10% of falls unrelated to syncope are

related to acute illness, such as pneumonia, stroke, anemia, or dehydration (Vassallo &

Sharma, 1998).

Cardinal Needs Schedule

A stepwise multiple regression analysis was used resulting in two variables that

significantly explained the incidence of falls in the study: the resident's perceived health

deficits as measured by the SF-36v2 (F = 6.03, p = .0175); and CRS (F = 5.70, p =

.0210). The current discussion will focus on the CRS, or needs as measured by the

Cardinal Needs Schedule. This is the second factor in the conceptual model adapted from

Andersen's Behavioral Model for Vulnerable Populations (1995).

The Cardinal Needs Schedule utilizes a negotiated model of needs assessment, in

which needs are not a fixed concept that can be objectively measured. Rather, needs are

viewed as dynamic, being influenced by a range of contextual factors and on which there

is no single correct perspective (Natten & Beecham, 1993; Slade, 1994; Slade &

Thornicroft, 1995). Hence, the Cardinal Needs Schedule included input from staff,

residents, and the clinical record. Although some researchers suggest that disparites may

arise between staff and residents' perceptions of needs for care (Lesalvia, Ruggeri,

Mazzi, & Dall'Agnola, 2000), the Cardinal Needs Schedule also included data gleaned

from the clinical record to complete the needs assessment. It was not within the scope of

this research to determine if differences existed between staff and residents' perceptions,

and/or whether or not they were significant or predictable.








The areas of functioning that were assessed using the Cardinal Needs Schedule

were (1) psychosis, (2) side effects from psychotropic medications, (3) anxiety or

depression, (4) self harm or violence, (5) organic disorder, (6) health, (7) socially

embarrassing behavior, (8) drugs and alcohol, (9) domestic skills, (10) finance and

welfare, (11) transport and amenities, (12) literacy, (13) work, (14) social life, (15)

hygiene and dressing, and (16) accommodation. Those areas of functioning where needs

were identified were (1) psychosis (27 needs), (2) side effects (7 needs), (3) anxiety or

depression (33 needs), (4) self harm or violence (27 needs), (5) health (1 need), (6)

socially embarrassing behavior (28 needs), (7) domestic skills (1 need), (8) finance and

welfare (1 need), and (9) hygiene and dressing (1 need).

It is interesting to note that the area of functioning, health, only has one identified

unmet need. Since residents are based in a tertiary facility, the assumption could be that

health care needs are being met adequately, and that there are no unmet needs. This

assumption was formally investigated in research by Lasalvia, Ruggeri, Mazzi, and

Dall'Agnola (2000). The research findings of other investigators suggest that if persons

with severe and persistent mental illness see their primary care physicians on a routine

basis, they are more likely to think that they receive the right kind of help for their

physical health needs and be more satisfied with the type of help received (Beecroft et al.,

2001). At Wellspring, the residents see their primary care physician on a daily basis and

during medical rounds.

Although not a significant finding, data revealed that women who participated in

this study were identified as having 1.378 times fewer needs than men, or, conversely,

men were identified as having 1.378 times more needs than women. Data revealed that








men were identified as having more needs in the functional areas of psychosis, side

effects related to psychotropic medications, self harm or violence, health, socially

embarrassing behaviors, domestic skills, finance and welfare, and hygiene and dressing.

Caution needs to be exercised when attempting to explain these findings. Current

research indicates that derivational thinking and gender bias are particularly evident in

the healthcare arena and may influence the way in which healthcare professionals identify

needs of their residents (Hardman, 1996; Leo, 2001). Leo (2001) asserts that gender

biased instructional materials and training in medical schools influence gender role

stereotypes. In fact, previous investigations of gender bias in anatomy or physical

diagnosis texts have noted that illustrations of male models exceed those of female

models (Giacomini, Rozee-Koker, & Pepitone-Arreola-Rockwell, 1986; Mendelsohn,

Nieman, & Isaacs, 1994). Researchers have further suggested that medical students and

other healthcare workers may be unfamiliar with female healthcare issues, and consider

male healthcare issues to be the "norm" (Mendelsohn et al., 1994; Nechas & Foley,

1994). Hence, the female becomes invisible; female residents' needs and/or concerns

may go unnoticed, minimized or ignored by the person conducting the assessment. Russ

(1983) describes this phenomenon as that of 'double standard of content': the notion that

female residents' experiences can only be defined in terms of the existence and

characteristics of male residents. Hence, male resident's needs and concerns become the

focus of attention, particularly when they involve stereotypical attributes, such as violent

or embarrassing behaviors (Hardman, 1996; Russ, 1983).

Based on the finding that only one health need was identified, context-specific

interventions as part of a fall prevention program may not be perceived as a health-related