Group Title: impact of a physicians' assistant clinic on a rural southern county
Title: The impact of a physicians' assistant clinic on a rural southern county
Full Citation
Permanent Link:
 Material Information
Title: The impact of a physicians' assistant clinic on a rural southern county a descriptive evaluation
Physical Description: x, 144 leaves : ; 28 cm.
Language: English
Creator: Burke, Robert Edumund, 1947-
Publication Date: 1977
Copyright Date: 1977
Subject: Physicians' assistants -- Public opinion   ( lcsh )
Public opinion -- Florida -- Gilchrist County   ( lcsh )
Rural health services -- Case studies   ( lcsh )
Clinics -- Utilization -- Case studies   ( lcsh )
Sociology thesis Ph. D   ( lcsh )
Dissertations, Academic -- Sociology -- UF   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Thesis: Thesis--University of Florida.
Bibliography: Bibliography: leaves 137-143.
General Note: Typescript.
General Note: Vita.
Statement of Responsibility: by Robert Edmund Burke.
 Record Information
Bibliographic ID: UF00098855
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000210044
oclc - 04168425
notis - AAX6863


This item has the following downloads:

PDF ( 6 MBs ) ( PDF )

Full Text


D ,






For my paren's,



LIST OF TABLES . . . . . . . . . . .... . vi

LIST OF FIGURES ... .. . . . . . . . . .. viii

ABSTRACT .. . . . . . . . . . . . ix

I. OBJECTIVES . . . . . . . . . . . .

Introduction . . . . . . . . . . . 1
Research Questions . . . . . . .. 2
Background . . .. . . . ... . . . 3
Societal and Cultural Conditions . ... ..... 4
The Physician's Assistant . . . . . .. 7
Summary . . ....... ......... ...... .
A History of Rural HealTh Care Delivery . . . . 9
Overview of Community . . ... . . . . . 11
Background . . . . . . . . . . 11
History of Health Care Services . . . ... .. .12
County Political Situation . . . . . ... 13
Summary ..... . . . . . . . . 15

S1 REV E'W OF LITERATURE . . . . . . . . . 17

introduction . . . . . . . . ... . . 17
The Physician's Assistant Literature . . . ... 17
The Role of the Physician's Assistant . ... ..... 19
Licensure . . . ...... . . . .. . 20
The RecruiTment of the Physician's Assistant ... 20
The Training of the Physician's Assistant . . .. 21
The Evaluation of the Physician's Assistant ... .22
Health-Services Utilization . . . . . ... 24
Summary . . . ............ ..... 25

1I; CONCEPTUAL FRAMEWORK . . . . . . . . . 27

Introduction . . . ... .. . . . . . 27
Approaches to Health Services Utilization Analysis . 28
The Andersen-Newnan Model . . . . . . 29
Summary . ......... ..... . . . .. 35

IV. RESEARCH METHODOLOGY . . . ... ... . . .... 38

Design . . . . . . . .. . ...... 38
Data Collection . . . . . .. . . 39



IV. (Continued)
The Sample . ... .. . . . . . . . 39
Field Test . . . .. . . . ... . 41
The Data . . . . . . . . ... . . 41
Reliability and Validity .... . . . . . 45
Statistics ...... . . . . . . . 47
Summary . ... . . . . . . . . . 47

V. FINDINGS. ...... . ... . . ... . 49

Overview . . . . . . . . ... .. . 49
Descriptive Survey Results . . . . . . .. 51
Discussion of the Six Specific Questions .. . . 56
Summary . . . . . . . . . . . 8a


APPENDIX A ....... . . . . ..... . . 100

APPENDIX B .... ................. . . 110

BIBLIOGRAPHY..... . . . . . . . . .... 137

BIOGRAPHICAL SKETCH ... . . . . . . ... 144


Table Pace
14. Change in Amount of Worry about Receiving Emergency
Medical Services, Preclinic, Postclinic Opening
Controlled for Clinic Utilization . ... . . . 68

15. Change in Amount of Worry about Having an
UndeTected Disease, Preclinic, Postclinic
Opening Controlled for Clinic Utilizaticn . . . . 69

16. Clinic Utilization by Prospective Utilization ...... 72

17. Perception of Physician's Assistant Clinic as
Compared to Usual Care Received by Prospective
Physician's Assistant Clinic Utilization . . . ... 74

18. Change in Number of Physician Visits, Preclinic,
Postclinic Opening Controlled for Clinic Utilization . 75

19. Change in Amount of Hospitalization, Preclinic,
Postclinic Opening Controlled for Clinic Utilization . 77

20. Change in Utilization of Prescription Drugs,
Preclinic, Pcstclinic Opening ConTrolled for
Clinic Utilization . . . . . . . . 78

21. Change in Having a Pap Test, Preclinic, Postclinic
Opening Controlled for Clinic Utilization . . . ... 79

22. Discriminant Function for Wave I Prediction
Results . . . . .... ... . . . .. . 83

23. Discriminant Functions for Wave II Prediction
Results . . . . . . . . . . .. .. 85

24. Comparison of Wave I and Wave II Variable Means
for Nonutilizers ....... . . . ..... .86

25. Comparison of Wave I and Wave II Variable Means
for Utilizers ..... . . . . . . . 67

26. Wave I Variable ~Means by Utilization .... . . ... . 88

27. Wave II Variable Yeans by Utilization . ... .... . 89


Table Page
1. Timetable for Data Collection for Impact of
Physician's Assistant Clinic Study . . . . . 31

2. General Format Change in Predisposing, Enabling,
and Illness-Level Variables as a Function of
Utilization of the Physician's Assistant Clinic . . .. 50

3. Wave I and Wave II Survey Comparisons of Predisposing
Sample Characteristics . . . . . . .. . . 52

4. Wave I and Wave II Survey Comparisons of Enabling,
Health Behavior Variables . . . . . . . 54

5. Wave I and Wave II Survey Comparisons of Illness-
Level Variables . . . . . . . . . . 55

6. Clinic Utilization by Occupation of Male
Head-of-Household . . . . . . . ... .57

7. Clinic Utilization by Respondent's Age . . . ... 59

S. Clinic Utilization by Family Physician Visits . . . 60

9. Family Physician Visits by Prospective Clinic
Utilization . . . . . . . . ... . . . 62

10. Change in Perception of State of Health, Preclinic,
Postclinic Opening Controlled for Clinic Utilization . 63

11. Change in Perception of Health Interference with
Daily Activities, Preclinic, Postclinic Opening
Controlled for Clinic Utilization . . . . . ... 65

12. Change in Perception of Receiving Enough Care,
Preclinic, Postclinic Opening Controlled
for Clinic Utilization . . . . . . . . . 66

13. Change in Amount of General Worry, Preclinic,
Postclinic Opening Controled for Clinic
Utilization . . . . . . . . . .. . 67


Figure Page
1. Andersen-Newman Model of Determinants of
Health Care Utilization . . . . .. . . . 31

2. Model of Determinants of Health Care
Utilization . . . ... . . . . . . . . 32

3. The Andersen-Newman Model: Determinants of
Health Service Utilization .. ........ . . 34

4. The Modified Andersen-Newman Model:
Determinants of Health Service Utilization . . . . 36

Abr.rcl of Di sseriat;i' Fresen-td io the Graduuie Council
of I t [Ini 'or y of fo id in r Fuyfi n of f o d i i he Requ i -rti nr n I.
for [he [a gree of Doctor of Phil iosophy



Robert Edmund Burke

Augui st, 1977

Ch...i 4ai.,(: e.or-je W'arheii, iPh.D.
j,. ep r n Soc iolov'

i.e p'hyicin's assistant became a primary heal ih-care provider

:, ~.~'th. ii;: tha- me ti physicic a" s aissisFant has beer. -t sWb-

j':'" o" ;, il eba e aid cOFie! itry. Research has just 'beun

.-i d!y.' g 'ih physiciaki' assisanl as a he l. -ri proviJer. The focus

1' i/. di: ;' :.-r- iio0n i, to descr ip -iely ov. i :'L:. t:e community accept-

* c : cl i : on ar '?eid by ph y ician's assi atn s in a rural

. ,-, cu., k, wree no othE- p riary he.o 1h sr. vice ws ava i ldble.

S- ni' .' L e rchi que-.i o n1 : :s measurJ d by Ilinic uti H za-

" ,- d1 c n ily :'crpi 'hC p;y/sic.oan's assi.-is anr c! inic? This

i .: ',: odl'(.'; by 'n'lysi of iwo u ve'ys. Th'' i'rst survey

',' -'!; ,": ; ;-rl ye:,ir bei or. th0 h ci inc opened, lrhe second one yoCr.

Sc;- ;, -. cli.i, h F eo. oE ~ H on.

S,.' r. ';-: ',; s no p, i,'A ry ; a,3 ih care available in the coLunty,

S' ....- [ -'i ,Fi '[i:-. r co 'J; s' I o-, of free SI rvi. i s and ihat

'F 'e c !inc would hO lp 1.1-,; e an eix rcme ned.

The analysis employed a ha th-sorvice u ti ization frm.'work. Variables

which af fecid clini c uiil i action were examined. Th'j results of f ris

anialysos are important to -he growing body of research Iiterature and

to rolhli planners who have the esponsibi lity for developing and moni-

tori'r nerw primary heal h-carc providers in the expand ing heal th syslom.

The si vy dact did no support tlia -the coun ty residents were experi-

encing extronm health-care needs. The daia demonstrate that the

resia-nts had become accustomed to traveling thirty miles to tic

medical centers of e nearby city.

The findings s'ugest the clinic 'was utilized by county residents

for ie ih care cnd thl thcrp,. i wo,z positive change in certain

iIi ness- lve v wr' i los dlocr, the uti l izrs of the clinic.



This dissertation presents data from a research project designed

to investigate the establishment and community acceptance of a physi-

cian's assistant clinic in a rural southern community.

It was assumed by health planners at a nearby university that a

county without recognized, established, and professional medical-care-

personnel living and practicing within its boundaries would probably be

in a state of extreme need. The physician's assistant clinic was seen

by some of these planners as a potentially viable method of alleviating,

at l-east somewhat, the acute shortage of primary-care deliverers in

rural areas in Florida and elsewhere.

!n order to assess the effectiveness of the physician's assistant

clinic in meeting health care needs, a series of health surveys 'as

planned. These surveys were designed to provide information about the

service utilization patterns of residents of the county before and

efter the physician's assistant clinic became operational. After an

ini-iai health survey of the target county was completed, a second one

was designed to answer two basic questions: 1) what was the impact of

The physician's assistant clinic on the health practices of those in

!hs county; and 2) were the services provided a viable alternr.tive -o

traditional primary medical care? The method of examining the impact

of r!:h new professionals upon the county was the measurement of the

utilization of health services rates of a purposive sample of county


residents. These data allow a comparison between the users of the

physician's assistant clinic and the users of traditional health care


The community under investigation, as will be shown in the next

section, is typical of many rural counties in the United States. The

county is sparsely populated, relatively homogeneous, and dependent upon

other neighboring counties for the majority of necessary services. The

impact of the physician's assistant clinic on the health Dractices of

this rural county was seen as having far reaching implications in that

if this clinic was successful, it would become a model for planners of

rural health care delivery at a regional and national level. The

research reported in this dissertation is significant for two reasons.

First, from a sociological perspective, it describes the acceptance

and utilization of a physician's assistant clinic and second, as

aeolied basic research, iT can be a guide for future planning of health

services in rural communities.

Research Cuestions

The research focused on one general question: will the residents

of a rural community accept and utilize a physician's assistant clinic?

This question is comprehensively addressed by these six specific sub-


1. Who in the community will utilize the clinic?

2. For what illnesses or injuries will they utilize the clinic?

3. Will They continue to utilize The clinic over time?

W. Will the clinic be compared favorably to other health-care

services received?

5. Will utilization of the clinic affect the health seeking

and utilization behaviors of [hose in the community?

6. Will the clinic be utilized by all segments of the


These questions are addressed by analysis of the two health surveys.

The questions were the issues that were important to the developers and

sponsors of the clinic.

The establishment of the physician's assistant clinic was not a

spontaneous event. The growth of advocacy for health care services

nationally and the conditions specific to the target community contrib-

uted to the explanation of the development of the physician's assistant

clinic. Both of these conditions are discussed in the next section.

Back'o round

The availability of medical care has become a concern to many varied

segments cf society. Those in medical organizations, politicians,

sociologists, and the consumer public have their own perspectives re-

garding the need for medical care. Several events are, perhaps, casually

related to this surge of demands for health care. These events include

national health insurance programs which have been successful in other

western nations, third party payment agencies which have made health

care servicess available to individuals not previously covered, and the

federal! programs of Medicare and Medicaid which have withstood Their

shaky beginnings and, in fact, have recently become broader in scope

anid medical care coverage.

The availability of medical services due to third party payment

bh h-ih the private and public sectors has had a great impact on the

h:lith care delivery systems in the United States. More i pdividuals

have coverage for health care services than at any other time in history.

As third party payment became more prevalent, the maximum or ceiling

charges permitted by ihird party payers became the floor or minimum

charges for providers. This fact, together with increasing inflation,

has resulted in a spiral effect in health care costs. Additionally,

labor unions have applied pressure to management for more health care

coverage. The additional coverage has affected both the corporate

profits and the costs of their products to the consumer. This has had

a circular inflationary effect on the economy. It also has the effect of

increasing the demands by the public for health care services. In

addition to increased demands, strict professional control has kept

providers in scarce supply.

On analysis, the mass media reflect the value of health and

health-related products in contemporary society. The media commercials

suggest that health is an individual's right and that good health is

the most important product not only for the individual but also for the

society to obtain and maintain. In the last decade, health spas and

health food stores sprang up over night to meet the demands for health

products and services. "Natural ingredients," "organic" foods and

cosmetics, as well as "health" clothes such as the jogging suit have all

just recently become household words and items. In light of these

formal and informal pressures for more and more services a basic question

persists: how can the American medical system meet these increasing

demands and at the same time remain in control of their distribution,

quality, and the fees charged for them?

Societal and Cultural Conditions

As noted above, many societal factors are making increased demands

up.n the organized medical system for additional health-cnre deliverers.

According to Health United States 1975 (1976), in 1973 there was one

physician for every 562 individuals in the United States. Physicians,

however, are clustered in metropolitan areas where there was one

non-federal physician providing patient care for every 500 residents.

The raTic for small nonmetropolitan areas was about one physician for

every 2,000 to 2,500. The populaTion-physician ratio was about four

or five times greater in the nonmetropolitan areas than in metropolitan

ones. FurThermore, while there has been a decline in the number of

general practitioners in the past ten years, there have been increases

in the number of other medical specialties that are considered to be

primary care deliverers. These specialties are internal medicine,

pediatrics, and obstetric-gynecological medicine. Health manpower does

not seem to suffer from shortage as much as from distribution problems

and this maldistribution is undoubtedly related to many present demands.

The attraction of the physicians to rural areas had been unsuccess-

ful. Placing physician's assistants in rural areas where a metropolitan

based physician would be able to monitor them by Telephone and periodic

visits would aid in The physician, rural population location incongruence.

Utilizing physician's assistants in this way, a physician would be able

to serve a rural population without having to leave the metropolitan

area on a permanent basis.

A simplistic method of conceptualizing this situation is to view

the consumer and the medical system as two sections of an equilibrium

model. In order to exist and be maintained, all crucial demands for

services must be met. Thus, today's consumer demands for increased

medical services should produce or result in an increase in the medical

resources available. On the other hand, a technique which would

convince the consumer to decrease his demands for services would

allow equilibrium fo exist without having lo change the medical system.

There are four methods whereby the consumer and the medical system may

approach equilibrium.

One method of meeting the demand for more primary health care ser-

vices would be to redistribute existing primary health professionals.

Any such proposed redistribution would meet with strong professional

resistance on the part of the medical community: it is not a very

viable consideration.

Another method of meeting rural health needs would be to produce

more physicians pledged to practice in rural areas. There are two

problems with this possible solution. First, the number of physicians

is largely controlled by the medical profession and a plan to increase

the number of physicians would not meet with ready acceptance from this

croup. Second, federal loan programs which are forgiven if physicians

agree to practice in rural communities have not been successful. Less

than five percent of those who received these loans have requested

forgiveness. They, perhaps, have been socialized through their educa-

Sional experience to believe rural areas are not the best location

for their practice.

A third method would be for communities to actively recruit foreign

medical school graduates but foreign medical school graduates have not been

attracted to rural communities; they require colleagueal relationships

to facilitate American liconsure. The reasons these three methods

o' attracting physicians to rural areas have failed are both economic

ind social. Rural areas typically do not have the availability of

laboratories, pharmacies, hospital Leds, and other support networks

required for physicians to prosper professionally and personally.

A fourth method proposed to extend the resources of the medical

system has been efforts to develop a new paraprofessional. The primary

function of such a provider is the provision of primary health care

services to rural communities. This method has been endorsed by the

AMA because this new paraprofessional would permit the medical system

to expand under the direct control of The existing medical associations.

This new paraprofessional is only licensed to give primary health care

services under The direct supervision of a licensed physician. This

method has been actively supported for the last ten years by both the

established medical system and the government. Furthermore, the accep-

tance of this new position would necessitate societal adaptation.

Potential consumers would have to be made aware of the new position,

know the services performed and become convinced of their quality.

The physician's assistant can be considered AMA's response to the

societal demands for more care. This new provider calls for adaptation

from the societal members in terms of their attitudes, expectations,

and behavio-s.

The Physician's Assistant

As noted, one of the primary ways the medical system responded

to tre increased demands for more providers was by the creation of a

new heatlh care status, the physician's assistant. This new health

care provider was sanctioned by the American Medical Association, when in

1970 they drafted "Guiaelines for the Development of New Health

Occupations" and in 1971 the AMA House of Delegates approved "Essentials

of an Approved Education Program for the Assistant to the Primary Care

Physician" (Johnson, 1975). The latter document clearly defined the

essential types of patient services a physician's assistant may


1) The initial approach to a patient of any age group
in any setting to elicit a detailed and accurate
history, perform an appropriate examination, and record
and present pertinent data in a manner meaningful to the

2) Performance and/or assistance in performance of routine
laboratory and related studies as appropriate for a
specific practice setting, such as the drawing of blood
samples, performance of urinalyses, and the taking of
electrocardiographic tracings;

3) Performance of such routine therapeutic procedures as
injections, immunizations, and the suturing and care of

4) Instruction and counseling of patients regarding physical
and mental health on matters such as diets, disease,
therapy, and normal growth and development;

5) Assisting the physician in the hospital setting by making
patient rounds, recording patient progress notes, accurate-
ly and appropriately transcribing and/or executing standing
orders and other specific orders at the direction of the
supervising physician, and compiling and recording detailed
narrative case summaries;

6) Providing assistance in the delivery of service to patients
requiring continuing care (home, nursing home, extended
care facilities, etc.) including the review and monitoring
of treatment and therapy plans;

7) Facilitation of the physician's referral of appropriate
patients by maintenance of an awareness of the communi-
ty's various health facilities, agencies, and resources.
(Johnson, 1975, p. 23)


This section presented a discussion of the factors influencing

the increased demand for health services and how the creation of a new

medical provider meets these demands. The physician's assistant was

viewed as the AMA's creation, an ameliorative agent whose function is to

provide primary care. This new paraprofessional is accepted and

controlled by the medical profession and legitimate by state statutes.

Hence, they will be in existence for soma time to come and are worthy of

being studied.

A History of Rural Health Care Delivery

"The notion that rural life nad its health handicaps, in spite of

fresh air and sunshine, was expressed as early as 1862 by the Commis-

sioner of Agriculture to President Abraham Lincoln" (Roemer, 1976, p.3).

Several sporadic attempts to alleviate the rural health problems were

initiated late in the nineteenth century. It was not until the enact-

ment of -he Sheppard-Towner Act in 1921, however, that a systematic

attempt was made to strengthen rural county health departments by making

available granT funds to support maternal and child health centers. The

depression of 1929 was a setback to this rural health initiative as the

Sheppard-Towner Act was terminated. In order to build up preventive

health programs, Titles V and VI of the Social Security Act of 1935 had

the effect of reinstating the Sheppard-Towner Act. By 1942, some 1,800

counties had public health coverage. The remaining 1,250 counties with-

out coverage were typically rural with small population bases. At the

close of Worla War II, an effort was made to consolidate counties in

order to provide public health coverage. In 1945, Dr. Haven Emerson

launched the "basic six" program of APHA. The duties of public health

deDartments were to be concerned with six tasks: 1) communicable

disease control, 2) environmental sanitation, 3) maternal and child

health preventive services, 4) health education, 5) vital statistics,

and 6) laboratory services. These services, although needed,

do not include direct primary patient care. This component of rural

health-care delivery was left to private physicians. Some states had

proposed plans to recruit physicians to thei- own rural areas. New

Hampshire in 1923, Virginia in 1942, and Tennessee in 1953 had each

enacted legislation to actively recruit and financially support

physicians for their rural areas. In 1967, the National Advisory

Commission on Health Manpower requested there be not only a growth in

numbers of physicians but also an increase of manpower in the delivery

system with new categories of health professionals.

Other countries have had a history of medical assistants as primary

health-care deliverers. Russia had developed the "feldsher", a type of

medical or physician assistant in the 1870s, to function as the primary

deliverer of health care in The rural regions. Other countries, e.g.

Iran and Mexico, also have similar requirements for newly trained

physicians. Following these models, in the late 1960s, federal grant

moneys were made available to academic institutions to develop training

programs for medical assistants; nurse practitioners, physician's

assistants, and midwives were being trained under the auspices of both

medicine and the federal governmenT. Furthermore, in 1970, the

Emergency Health Personnel Act established the National Health Service

Corps. Under this act, physicians, nurses, and dentists were sent to

needy communities where they practiced in lieu of fulfilling their

active military obligation. These programs were designed to increase

the medical manpower in rural areas.

To augment the increase in health manpower, several federal acts

were passed to increase the supply of hospital beds in rural areas. The

most famous of these was the National Hospital Survey and Construction

Act of 1946. This legislation, better known as the Hill-Burton Act,

provided grant funds to states for the construction of hospitals. The

maximum aid was earmarked for rural communities. Today, there is no

longer a disparity of bed supply between urban and rural areas. In fact,

the trend in the past thirty years is that hospital utilization by

rural-living individuals is upward. With Transportation improvements,


rural residents seem to be bypassing the rural community hospital for

the more sophisticated ones in urban areas.

This history demonstrates that the delivery of primary health

care to rural communities has been an issue in this country for over a

hundred years. It has not been, however, until the last decade that

a systematic attempt has been made to alleviate the problems of rural

primary health care.

Overview of Community

The goal of this section is to describe three important factors

in the county which have a bearing on the acceptance of the physician's

assistant clinic. First, the historical background of the community

is discussed. Second, the history of medical services available in

this community is outlined, and third, the local political situation

which gave rise to the initiation of the physician's assistant clinic

is presented.


The county which served as the research site is situated in North

Central Florida. Its north and west boundary is formed primarily by

the Suwannee River as it makes its winding path to the Gulf of Mexico.

The adjacent county due east is a standard metropolitan statistical area,

which has a state university and is a major agricultural trade cross-

roads. The research county is primarily agricultural. One of its

largest employers is the forestry division of a national lumber

conglomerate. The county is not self-sufficient. It does not have

a hospital, college, railroad, or major shopping area. Some of the

churches are served by itinerant or missionary ministers who drive in

each Sunday. The major churches are Baptist and Church of Christ.

The county seat has one restaurant, a high school, court facilities,

several hardware and farm implement dealers, a library, and a bank.

A bus serves the county.

The central section of the county is swamp land. The major

inhabited area forms an "L" shaped figure along the western and

southern edge of the county. There is a state highway going east-

west along the southern edge of the county. According to the 1970 census,

the county has 3,551 residents, of whom 3,259 are white and 287 are black.

The toal male population is 1,797; the total female population is 1,754.

history of Health Care Services

The health services in the county have been limited. At one time,

there was a general practitioner living in the county seat; he did not

remain in practice long as age and ill health caused him to retire.

His son took over the practice but left the county for personal

reasons. On his departure, the county was without its own physician.

Residents in need of medical services-had no alternative but to go


There were naturopaths functioning in and around the county; the

river doctor, for example, was a naturopath who, for a fee of five dollars

in cash, -reated clients and distributed various herbs and teas for cures.

The author learned about the river doctor from a clinic patient. The

river doctor saw clients in his mobile home located outside the county

seat on the bank of a river, hence the name. According to this

informant, the river doctor's cures had been effective and his business

brisk. The informant explained the only reason she came to the

pnys;cian's assistant clinic was that the clinic billed patients at

the innd of the month. By the time the bill arrived, she would have

the ability to pay her fee; the river doctor wanted cash and extended

no credit.

Residents of the county in need of hospitalization were sent to

the county hospital in a neighboring county. In the mid-sixties, two

other hospitals opened in that county, a university teaching hospital

and a Veterans Administration Hospital. These were followed by a

private hospital in 1972. This county grew to become a major

medical center. The residents of the research county have always been

dependent on this neighboring county for hospital services. With the

health-care delivery in the research county having a history of being

scarce, its residents developed a routine of leaving the county to

seek medical care. With the road to the neighboring county wide, flat,

and accessible, it may not have been considered a burden by county

residents to drive to a major medical center, thirty miles to the east.

County Political Situtation

The author also learned that a deceased state legislator had been

an influential political resource in and for the county. The represen-

tative was considered to be the individual responsible for establishing

an adolescent detention center built at the far southwest corner of the

county. The residents and professional staff did not socialize with the

county residents but the center did provide jobs. The representative

and the mayor of the county seat were both responsible for creating the

County Health Board.

This board received state funds for creation and maintenance of

a medical facility. Some of these funds subsidize the physician's

assistant clinic. The County Health Board approached the Department

of Community Health and Family Medicine at the medical college of the


university and requested a rural health clinic, similar to one operating

in another rural county. The response from the medical college was not

to create a clinic similar to their existing one but rather to use the

opportunity to establish a clinic where the physician's assistant was

the principal deliverer of primary care.

Three factors, a new County Health Board eager to begin a primary

health-care-delivery system, an aggressive Department of Community

Health and Family Medicine, and the existence of an acute shortage

of health care services all came together. The result of the confluence

of these three factors was the proposal of an innovative health care

system, the physician's assistant clinic.

The Department of Community Health and Family Medicine and the

Community Health Board came to an agreement whereby the clinic opened

with two physician's assistants who were under the direct supervision

of a physician. In addition, two other physicians from the medical

college also served the clinic. During the negotiation and planning

stage, the state representative who was a vital voice in this process,

died. His death, however, did not hamper the establishing and opening

of the clinic.

This overview has presented a community profile and a history

of health services and has sketched the political factors which influenced

the establishment of the physician's assistant clinic. This information

gives perspective to the issue of acceptance and utilization of an

innovative health care service. The county is not industrial, is not

used to rapid changes, is tied closely to the land, and is dependent

upon its neighboring communities for services. The fact that no

heal-h care services were available in this community is consistent

with the community history. The political factors described are the

important reasons for the establishment of the physician's assistant



The history of health-care delivery shows that only recently has

there been an attempt in this country to systematically address the

issue of providing primary health care to rural areas. The country

had far fewer communities without physicians in 1900 than today. Two

reasons were given to explain this fact. First, the choice location

to practice medicine is the metropolitan areas. Secondly, rural

family practice physicians since World War II have been moving to

metropolitan areas. These locations have medical centers, laboratories,

pharmacies, and other profession-oriented services which facilitate

Ihe practice of medicine.

This research describes the establishment and acceptance of a

physician's assistant clinic in a rural community. It addresses

several important questions concerning the utilization of a new type

of health service for rural communities. The community overview

demonstrated that this community is typical of many rural American

communities. The results of the research are relevant and important

to the study community as well as to other similar rural communities.

Fho next chapter presents a literature review concerning physician's

assistants ane health services utilization research. The rich

literature which focuses primarily on physician utilization lends

itself to an analytical framework for the research.


Chapter III presents a conceptual and analytical framework.

Chapter IV presents the methodology of the research design, the data

collection procedures, and description of key variables. Chapter V

presents the conclusions and summary of the research.



The goal of this chapter is to review the literature on physician's

assistants in the context of the health-services delivery field. The

review is designed to provide a framework for the presentation of the


The PIysician's Assistant Literature

The physician's assistant as a health provider came into existence

in 1968, hence, the literature is limited from that year to the present.

The literature can be placed into five categories:

1) The role of the physician's assistant,

2) The legality of the physician's assistant,

5) The recruitment of the physician's assistant,

4) The setting up of the physician's assistant training

programs, and

5) The evaluation of the physician's assistant.

Ford (i976) attempted to address the sociological implications of the

physician's assistant. Her research suggests that sociological inquiry

into issues of authority, power, status consistence, quality of care,

etc., have not been empirically dealt with. Instead, research that is

based in other professions, primarily nursing, have looked at the

physician's assistant from outside in order to achieve their own

professional goals and legitimacy. Ford, herself, falls into the same

reasoning process as she directly compares the development of the

physician's assistant with the development of the nurse practitioner.

This is problematic in tnat both of these emerging occupational

roles have different origins and power bases. The nurse practitioner

is primarily a development of entrepreneurial nursing; the physician's

assistant as currently developed is directly subservient to a local

practicing physician.

The sociological aspects of the physician's assistant have not

been addressed directly in the research literature. Thus, this review

limits itself to the five categories of available research concerning

the physician's assistant.

The literature has been written chiefly by physicians, nurses,

and medical administrators. Literature writTen by the physician's

assisTants themselves has, as yet, not emerged.

Sadler, Sadler, and Bliss (1972), who are physicians and attorneys,

were the first to prepare a comprehensive treatise dealing with the

various components, issues, and questions surrounding the physician's

assistant. Their research presented each of the above listed research

topics and focused primarily on an analysis of the physician's assistant

in terms of recruitment and his role in the medical system. They also

addressed themselves to the medical and legal ramifications of the

position in the total system. Their research, however, is not based

on empirical data but is a descriptive general overview of the areas

listed acove. Nevertheless, their monograph was the first attempt

to view most of the issues concerning the physician's assistant in

a systematic manner.

The Role of the Physician's Assistant

The greatest body of physician's assistant literature is concerned

with defining the role. These articles, interestingly enough, are

primarily written by nurses who view the development of the physician's

assistant as 'the symptom" of the inability, both of medicine and nursing,

to define their individual roles, to respect each other's competencies,

or to deliver an acceptable level of care (Rothberg, 1973).

Andreoli (1972) raises a different aspect of role behavior. From

her observation, the physician's assistant's relationship with a palient

usually begins before the patient is hospitalized and ends, not with

patient discharge, but rather is continued for posthospitalization

checkups. Andreoli states the nurse is educated, whereas the physician's

assistant is only trained in certain techniques. Lanhertsen (1972, p. 32),

also a registered nurse, considers the role of the physician's assistant

an immediate extension of the physician and argues: "Territorial dis-

putes (between physician's assislanis and nurses) have no place in the

decisions thai must be made. The objectives for health manpower must

focus on the ultimate potential of every person." The nurses are raising

the issue of where in the health manpower network the physician's

assistants are to be placed. Are the physician's assistants going to

take away some of the newly carved professional territory of the nurse?

The physicians, on the other hand, see the role of the physician's

assistant as an immediate extension of themselves. Estes (1973, p. 197)

describes the role of the physician's assistant: "...the assistant must

be able to assisi in and function in any of the sites in which the M.D.

is active--clinic, hospital, or patient's home--and must be available at

night as well as days. Because of this requirement for mobility and

the fact thal nurses were in short supply, it was decided to utilize

men rather than women." The definition of the role is one factor. The

actual role taking and role behavior of Ihe physician's assistant has yet

to be systematically described or verified in the I terature.


The second most commonly reported aspect of the physician's

assistant concerns legitimation. By whom will the physician's assistant

be licensed, and, if licensed, to whom and for what is he really respon-

sible? Can a nurse take professional orders from a physician's assistant?

Young (!972) reports that the AMA had placed a moratorium on the approval

of new licenses for all new health paraprofessionals. Before these para-

professionals are given approval for licensure by the AiA, Young

suggests two issues need to be resolved. The first is to standardize the

physician's assistant in regard to training and practice. The second

is to establish consensus of role definitions of not only the physician's

assistant but the M.D. as well. In other words, if an M.D. musl super-

vise his physician's assistant, of whale does this supervision consist?

Some states such as Florida grant licenses to physician's assistants.

However, this law clearly states physician's assistants are responsible

to the physician under whose supervision he works. Furtnermore, the

physician's assistant is to work only under a physician and a physician

is allowed only two physician's assistants in his employ. (Florida

Statutes, Chapter 458.135, 1975)

The Recru-iient of the Physician's Assistant

Another issue of the literature is that of recruitment. The first

recruitment of physician's assistants (Es-es, 1971) was aimed at military

corpsmen who already had some medical experience. Now that the Viet N-n

conflict has passed and military service has become voluntary, the

availability of corpsmen has declined. Horsley and Aschenbrener (1973),

as well as Onion and Schulten (1973), agree since the services of a

physician's assistant are most needed in the rural areas of this country,

an attempt must be made to recruit high-school graduates who have a

strong desire to serve in these areas. They suggest those desiring

to work in rural areas be given preferential treatment. The literature,

again, is ideologically biased and does not take into account the impact

that the training as a socializaTion agency may have on changing the

physician's assistant's mind as to where he will practice.

The Training of the Physician's Assistant

The training of a physician's assistant and the establishment of

training programs have been dealt with in the literature, with consider-

able overlap. Again, however, little or no empirical research is found.

STone and Bassett (!972) conclude that the greatest benefit of the Medex

training program is that the physician's assistant and the M.D. have

a one-To-one relationship, thereby allowing the physician's assistant

to learn first-hand end by experience. Adamson (1971) suggests

physician's assistants be trained along with M.D.'s and R.N.'s primarily

to learn where they fit into the nealth-care system. Adamson, therefore,

is suggesting -hat the training emphasize socialization. Bergen and

Clapp (1972) clarify that the training program is terminal; that is,

there is limited mobility. They feel the physician's assisTanT program

must not become a manner in which a student can later be admitted to

medical school to become a doctor. Johnson (1976) concerned his

research primarily with the anticipatory socialization process and the

training stages of becoming a physician's assistant. He finds that

physician's assistants are products of both the medical system and

the societal system. More importantly, ne states (p. 185),

"The A.M.A. established guidelines for these programs and
for the use of physician assistant in each state. In this
manner, the medical profession has essentially controlled
the introduction of the physician assistant. Furthermore,
the physician's assistant from training stage to practice
stage, unlike physicians themselves, remains in a depen-
dent relationship to a physician."

The Johnson research and this research are complementary. The former

focuses on socialization and training; this research analyzes community

impact and acceptance once the physician's assistant is in practice.

The Evaluation of the Physician's Assistant

The last section of the literature involves the evaluation of the

physician's assistant in the medical environment. Four studies have

been found that use data as a basis for evaluation. Coe and Fichtenbaum

(1972) evaluate the physician's assistants employed in a small community

hospital. The evaluation is based on interviews with other members of

the hospital staff. They analyzed the perception of the physician's

assistant role through others in the role set and found no consensus

on the role definition of the physician's assistant. The physician's

assistant were seen by the nurses as a vehicle for their own

upward mobility. The number of physician's assistants aT this

hospital is not mentioned nor is the method of interpretation of their

interviews. Strunk (1973) examines patienT attitudes toward the physi-

cian's assistant. He develops a two-factor index based on a sample of

300. He finds that acceptance of the physician's assistant for minor

medical care is greatest among nonmarried middle-class respondents who

have had some exposure to college. The acceptance drops, however, for

major medical care. Ford (1972) examines the case loads of physicians

who have been utilizing a physician's assistant and found there is

an increase in case load of 20 percent. Demaria, Cherry and Truesdell

(1971) report on the peer review of physician's assistants in the

Marines. The peers were not other physician's assistants but M.D.'s.

This program evaluation shows the benefiTs of their own training program.

They find their physician's assistants perform at a better-than-acceptable

level. Their research demonstrates that, by utilizing physician peer

review, their training program goals have been met.

Evaluation may be of several types. Evaluation can be of the program

(Demaria, Cherry and Truesdell, 1971); the process (Coe and Fichetenbaum,

1972; Johnson 1976); the outcome (Ford, 1972; Strunk, 1973) and the impact

uoon the community. No literature could be found dealing with the impact

of the physician's assistant on the community.

The review of liTerature demonstrates that the physician's

assistant, as a new health provider, has seen described with

regard to his real and ideal roles in the health care system. The

legal, training, recruitment, and program issues are reported primarily

from an editorial or ideological perspective. These reports, however,

contribute to this research since they clearly demonstrate the need

for analysis of not only the physician's assistant as a new phenomenon

but also the acceptance of this new health provider by the community

at large. As presented in the following chapters, this research is

similar to that of Strunk (1973). But, unlike Strunk's, it describes

the response of a community to physician's assistants where they are

the only primary health ca-e providers available. The evaluation of

Strunk focuses on current physician's assistant utilizers in an acute

care setting. This research represents a logical progression in the

systematic analysis of a new health-care provider.

Health-Services Uli izalion

Concepts such as primary health care delivery and health services

utilization are necessary and meaningful for analysis and intcrpretaticl

of the results of the study. Many studies have been undertaken whose

purpose is to explain why individuals utilize a health care facility

in order to build a predictive model for health care facilities and

programs planning.

Bodenheimier (1970), Freeborn and Greenlich (1973), and Rogers (1973)

determined that availability of services or access to services were

Tne primary issues of health care utilization. Chen (1973) developed

a utilization index that was composed of length of time to etc an

appointment, length of travel time, and length of waiting room time.

These research efforts were primarily monocausal. Donabedian (1973),

Mechanic (1972), Shortell (1973), Beck (1973), Andersen et al. (1971),

Anderson and Bartkus (1973), and Andersen and Newman (1973) have

all forsaken a simplistic unicausal explanation for health services

utilization in favor of multifactor explanations. Dcnabedian's (1973)

research documents that socio-organizationral variables such as geographic

or distance variables explain health service utilization. mechanic (1972)

further confirms the importance of these variables bul more importantly

states that the individual must also be willing to seek health care

and perceive lie is in need of these health services. Mechanic further

argues those components of health-care seeking behavior are learned and

are a function of cultural or subcultural values. Shortell's (1973)

research suggests although there are many factors which influence health

service utilization, the individual must also want this service. Beck

(1973) argues the primary reasons for nonutilization are the cost of the

service and its general availability.

These studies typically use the use-of-service as the primary

unit of measurement, that is, individuals who either have utilized

and/or are utilizing the service. These rates-under-treatment studies

do not take into account individuals who may need the service and are

no+ users. Anderson and Bartkus (1973), for example, have developed

a behavioral model for the choice of health care. They key variable

is the patients' ability to recognize symptoms as an indicator of need

and health seeking behavior. Since the sample was composed of patients

at a student health clinic, it may not be representative of the pop-

ulation. Unfortunately, Anderson and Bartkus do not build their model

with respect to nonusers who may be in need of health care.

Andersen (1968) and Andersen and Newman (1973) have developed a

general model of health care utilization ihat divides the variables

that are influential in determining health care utilization into three

major categories: predisposing, enabling, and illness level. This

health services utilization model is used in this study as a framework

for presentation of survey data and explanation of the acceptance of

the physician's assistant clinic by the community. This model is

described in depth in the next chapter.


This chapter has presented a review of literature of research

to date concerning physician's assistants and health care services

utilization. It was documented there is a need for research which


focuses on the facTors which influence the utilization of physician's

assistants. In addition, many models or explanations of health

services utilization have appeared in the last decade. Some have

simplistic unicausal explanations; others have developed multifactor

indices and scales for purposes of explanation. The Andersen-Newman

model of health service utilization provides a framework for the

interpretation of data such as those collected in the research county

health surveys. It is the major framework for data interpretation.

The next chapter will further define the Andersen-Newman model and

explain its applicability to this research.



"Existing research concerning the physician's assistant can be

characterized as fragmented" (Ford, 1975, p. 16). The literature

review contained in ChapTer II demonstrates the fragmentation. Ford

further states these fragments will only become of value to the under-

standing of the physician's assistant when they are drawn together.

Research concerning the physician's assistant is shown to typically

involve separate aspects of the physician's assistant such as his

training, his legaliTy, his tasks and functions, and his evaluation.

The research is also a fragment of a holistic analysis of the physi-

cian's assistant. The goal of this research is to describe the process

of testing whether a physician's assistant clinic is a viable alternative

to traditional primary health care. The acceptance of the physician's

assistant is one of the missing research fragments and the one to which

this research addresses itself. It is necessary to utilize existing

theoretical models to give meaning to the analysis of the sociological

problem. The physician's assistant as a self-contained role has not

been the topic of any research. The physician's assistant has, in

its brief history, been the subject of topical research endeavors.

To strengthen the interpretation of the survey data, which are

the basis of this research, a model of health service utilization was

employed. It was assumed -hat the factors which have been snown to


influence the utilization of transitional health services would also

be the factors which would influence the utilization of a new health

service. A description of the variables is conTained in This chapter.

Approaches to Health Services Utilization Analysis

The evaluation of the utilizaton of a new medical care service has

many stages and approaches. Andersen (1973) has classified these

approaches into the following:

1) Sociocultural Approach. This approach incorporates the

perspective that health service organization and utiliza-

tion are related to complex societal and cultural norms

and values.

2) Sociodemographic Approach. This approach emphasizes pop-

ulation based characteristics for explaining variations

in utilization.

3) Social-Psychological Approach. This approach seeks to

explain processes by which individuals realize that

they are ill and decide on a source of medical care.

4) Organization Approach. This approach tends to examine

differences in utilization as a result of the structure

of the delivery system.

5) Social Systems Aporoach. This approach atTempts To

examine utilization as a function of interactions of

interrelated components with one another and the popula-

tion at large.

Eacn approach tends to examine a set of variables and their relationship

to utilization of health care services. It is not the intention of

this research to describe these approaches in detail, listing the

advantages and disadvantages of each. A discussion, however, of the

social systems approach and its applicability to this study is presented.

It is necessary 1hat the approach used in examining health service

utilization be clearly relnled to the study objective. Though each

approach may be considered distinct, the variables incorporated in a

specific model may not be unique to that model alone, as is the case

of the social systems model. Instead, variables should be selected for

analysis because they are consistent with the approach and that in

turn will satisfy the research objectives.

The evaluation of health care is a new and complex area of inves-

tigation. Historically, the evaluation of health carol has been limited

to professional peer review. Hollingshoad (1973, p. 540) states,.

"Evaluation of health care by a process more objective than peer review

is indicated but, while the voice of the consumer is being raised about

cost and quality of health care, sociologists have given little atten-

tion to the rmurmurings of the populace." Through the utilization of

both survey research and observation methodologies, this dissertation

measures the impact of the new, emerging health-care provider, the

physician's assistant, on a rural community by measuring and inter-

preting the murmurings of the populace. Although sociological methods

are in a constant state of growth and development, this dissertation

is the first analysis and discussion of the impact of the physician's

assistant based on these methods. It uses the social systems model

of health service utilization as its theoretical foundation.

The Andersen-Newman Model

Ande-'.en and Newman (1973) have developed a theoretical framework

concerning societal and individual determinants of medical care

utilization in the United States. Figure 1 schematically presents

their framework.

The Andersen-Newman model is based upon the analysis of the

main components of the framework and their relationships To each other.

The major premises of the model are: 1) societal determinants affect

individual determinants directly and also through the service system,

and 2) individual determinants influence the health services consumed

by the individual. The second premise of the model, individual

determinants of health service utilization, has the same variables for

analysis as those collected in this survey. As such, this model groups

the survey data into theoretically interpretable categories that infiu-

ence health service utilization.

Individual characteristics of the population affect the health care

They receive. The model suggests that health service use is dependent

on: 1) the predisposition of the individual to use services (pre-

disDosing), 2) his ability to secure services (enabling), and 3) his

illness level.

Because of the limited scope of the survey daTa, the societal

determinant component of the model cannot be validated and analyzed.

Thus, this research focuses on the health service system, the individual

determinants, and the health service utilization components of the model.

Figure 2, "Aodel of Health Care Utilization", depicts the segment of

the Andersen-Newman model used in the analysis.

The predisposing variables illustrate that individual character-

istics prior to tne onset of illness are related to use. For example,

people in different age groups have various types and amounts of illness

and, therefore, different patterns of utilization. The enabling



Technology Resources
Norms Orcanizations


IIlness Level


Unit of Analysis


Andersen-Newman Model

of Determinants of Health-Care UTilization


Change in Crganizaticn


II ness Level


Physician's Assistant Clinic


Moael of Determinants of Health-Care Utilization

variables relate to the means available to use services. Health

insurance coverage is a primary enabling variable since it permits

an individual to act on a value or need concerning use. Illness-level

variables represent the most direct case of utilization. Given that

predisposing and enabling variables are favorable, there must be the

perception by the individual of illness or the possibility of its

occurrenceto motivate use. The nature and extent of the care, once

illness is perceived, is partly determined by the service providers.

Andersen and Newman have developed some tentative generalizations

based upon review of the literature and their own experience concerning

the relative importance of each seT of variables in predicting utiliza-

tion. l iness level is seen to be the major determinant in predicting

utilization, followed by demographic, family resource, and social

structure variables. Figure 3 presents these variables schematically.

The Andersen-Newman nmoel has oeen tested on survey data from the

National Opinion Research Center. Their research findings suggest,

from a theoretical perspective, this model has taken into consideration

the primary determinants of health services utilization to the degree

that the need for explanatory variables is minimal (Andersen and

Newman, 1973, p. 108).

Tne Andersen-Newman model, however, does not directly take into

account the reality of no medical services directly available in the

community. They do consider the ratio of health personnel to the

pooulatior as a community variable, bur, if the ratio is zero as it

was in the research county before the opening of the physician's

assistant ci iric, then perhaps the effect of no available heal th-care

services is understated.



Marital status
Past i illness

Social Structure

Family size
Rel igion*
mobil ity



Fami ly

Ratios of health
personnel and
facilities to
Price of health
Region of country*



Disabi i-ty
General state



Values concerning
health and
I lness *'
Attitudes toward
health services
Knowledge about

* Not available directly from survey, but due to the relative homogeneity
of the community they can be assumed as constants.

** Not available from survey.


The Andersen-Newman Model:

Determinants of Health Service Utilization

The model used in this research added the existence-nonexistence

of medical care services as a distinct variable in the enabling cate-

gory. The result modified theAnderserrNewman model, but this modifi-

cation did not dramatically change the model. The model was further

validated by contrasting changes in these variables by actual family

utilization of the new health-care service. Figure 4 presents the

modified Andersen-Newman model.


The goal of this chapter was to develop a conceptual model for

the analysis of the physician's assistant clinic. The purpose of this

study is the description of the evaluation, acceptance, and utilization

of the physician's assistant clinic by the community. The socioanthro-

pologic overview presented in the previous chapter suggested that even

though the community appears geographically isolated, it has a history

of dependence on its neighboring communities for social, economic,

education, religious, and health needs.

The outcome of this research should describe the determinants,

predisposing, enabling, or illness-level variables, which have been

affected by the use of the physician's assistant clinic. The outcome

should also describe which factors demonstrate the ability of the

community to adapt its existing pattern of health service utilization

in a manner which suggests the acceptance of the physician's assistant

clinic as a viable mode of health-care delivery. Furthermore, the

outcome should describe the level of acceptance of the clinic by

demonstrating what Factors made some individuals utilizers of the

physician's assistant clinic while others remained nonutilizers or

selective utilizers.



Social Structure

Bel iefs

Values concerning
health & illness
AtTituces Toward
health services


Family Perceived

Health insurance Disability
Type of regular Symptoms
source General state
Access to regular



RaTios of health
personnel and
facilities to
Price of health
Region of country*

New medical service


Snot available directly from survey, but due to The
of the community they can be assumed as constants.


The Modified Andersen-Newman Mcdel:

Determinants of Health Service Utilization

relative homogeneity


Given these questions which guided this research, the available

survey data, and the descriptive nature of this sTudy, this chapter

has outlined a model of health service utilization which has optimal

concepTual features for later data analysis. The next chapter discusses

in deTail The research design, the research questions, the methodology,

and the variables and statistics used in the analysis.



As noted previously, the purpose of this research is to address

two issues: 1) is the physician's assistant a viable alternative

to the traditional health-care provider; and 2) is the physician's

assistant accepted as legitimate by the consumers being served? The

level of acceptance is measured by the utilization of the clinic by the

community. In order to accomplish these research objectives, a two-stage

survey obtained data along three dimensions: health perception, health

attitudes, and health care service utilization. The data collection

instruments used in the first stage were developed by health planners

at a state university medical center; the second were developed by the

author in collaboration with sociologists and health planners. Data

from these sources include measures of self-perception of health,

attitudes about health and health care, and, finally, health care

utilization paTTerns.

The data have certain limitations. Neither of the survey instru-

ments contained previously tested scales or indices of health care

utilization and further, the data did not contain any health care

outcome measures usually found in health services studies. The

first data obtained were to be used as descriptions of behavior and

attitudinal factors That affected the utilization of the pnysician's

assistant clinic. The original health planners did not view this research



in a theoretical sense; their perspective was immediate and practical.

After the first health survey was completed, a broader, more theoretical

orientation emerged; it focused on the concerns and questions dealing

with more generic issues, e.g., the legitimacy and acceptance of the

new health care provider and his clinic. It was at this point that

the author become involved.

Data Collection

The data were collected from two waves of survey interviews and

from a six-month observation period at the offices which housed the

newly opened physician's assistant clinic. The first interviews

were administered two months before the clinic opened; the second

were administered after fourteen months of operation. The observation

was undertaken about one year after the clinic had commenced services.

Table 1 shows the data-collection timetable.


Timetable for Data Collection for Impact
of Physician's Assistant Clinic Study

Spring Summer Fall Winter Spring Summer
1971 1971 1971 1972 1972 1972

Wave I Survey

Clinic Operating

Wave II Survey

The Sample

The health planners decided that the research questions could be an-

swered most effectively by the adult female in each household who was typi-

cally responsible for the health care of her family. Usually, in a nuclear

family, this person is the mother. A purposive sample of one-third

of the households in the county was chosen. The sample size was based

on the following logic. Since the population of the county was 3,551

and the average American family has approximately four members, it was

estimated there were about 900 households in the county. Thus, a sample

of one-third of the households, or 300 families, was the sampling objec-

tive for the first survey. Since over 85 per cent of the adults in the

county were registered to vote, the sample of households was drawn

randomly from voter registration lists.

Immediately after the sample was collected, it was compared to the

census characterisTics of the county. It was determined that blacks

were underrepresented by 3 percent. A quota sample of 3 percent was

added to proportionately represent the blacks. A total of 227 family

households was reached. Once located the adult female most responsible

for the household was interviewed and asked to report on her own health

perceptions and behaviors as well as those of her spouse, children, and

other adults living in the same household. Data were not available which

described the reasons that the sampling objective of 300 was noT reached.

During The second survey wave, attempts were made to reinterview

each of these 227 family households; 85 percent (193) were located and

reinterviewed. Thirty-four first-year respondents were not able to be

interviewed, five had died during The year, twenty-one had moved and the

remaining eight were never home or refused an interview.

In order to determine if the administration of the first-wave

interview had sensitized and biased the participants witn regard to the

physician's assistant clinic and influenced The utilization of the same,

a comparison group of family households was added during the second

wave. This group was selected using the same sampling Technique.

Eighty-six households were interviewed. One family interview was

incomplete and deleted. Comparisons were made between this group and

the reinterviewed group. No differences were found with respect to

clinic utilization and the sociodemographic variables. This finding

meant that the first survey did not bias county residents with res-

pect to clinic utilization. Further comparative analysis between these

two groups was not undertaken. These 85 additional Wave II interviews

were added to the sample when description of events that occurred within

the past year are presented. Whenever changes in predisposing,

enabling, and illness-level variables are presented only the responses

of those females who participated in both interviews are used.

Field Test

After the second interview schedule was prepared, a field test was

conducted, 6 interviews were completed. After the field test was analyzed,

the schedule was critically reviewed. As a result, some minor revisions

were made. No items were deleted or added.

The Data

The general question was: will the residents of a rural community

accept and utilize the physician's assistant clinic? Responses to items

collected during the interviews provide the independent, dependent, and

control variables. Complete interview schedules are contained in

Appendices A and B. There are six specific questions used to comprehen-

sively address this general question. The first specific question was:

who in the community will utilize the clinic? Items used to address this

question were:

1) Age of respondent, classified into six intervals ten

years apart beginning at age 15;

2) Marital status, categorized as single, married, widowed,

and separated/divorced;

3) Occupation of the male head-of-the-household (an indicator

of socio-economic status). The U. S. Census categories

were collapsed to reflect a rural community. The first

category contains all professional and white collar workers;

the second, craftsmen and foreman; the third, operatives;

the fourth, laborers; and the fifth, farmers;

4) Family composition, categorized as nuclear complete,

living with relatives, and living alone;

5) Distance of the household to the clinic, measured in

miles, but collapsed into three categories: within the

county seat, within ten miles of the county seat, and


6) Welfare recipient during the year, categorized as yes

or no;

7) Health insurance status, insured or not; and,

8) Length of residence in the county.

Although data were collected about respondent's education, race,

and residential mobility, they were not used in the analyses due to their

lack of discriminatory ability. The county was assumed to be a prior

homogeneous with respect to ethnicity and religion. Data concerning

ethnicity and religion were not collected. These characteristics of

the research county were descriptively clarified in Chapter i.

The second specific question was: for what illnesses or injuries

will the county residents utilize the clinic? The items that address

this question are the above items plus questions concerning the

health-seeking behaviors for other service modes, namely physicians,

hospitals, and other health services.

The third question was: will the community residents continue

to utilize the clinic over time? The items that address this question

are clinic utilization and health behavior chance variables. Clinic

utilization was defined as use of the clinic during the year by the

respondent or a member of the respondent's nuclear family. New variables

were created to show change from Wave I to Wave II. These change

variables compare the responses between the two survey waves. For

example, if the perception of the level of worrying was lower during the

second wave than it was during the first, that was considered a positive

change. If the worry was higher in the second wave than in the first,

that was considered a negative change. If no change was detected

between the two surveys, that variable was scored as stable. This type

of variable creation permitted testing of the significance of change by

means of a sign test. Since the data were ordinal level, a variable

that quantified the magnitude of the change uould be erroneous. A change

variable was created for each item that measured health behavior or

health attitude:

1) Perception of state of health;

2) Perception of health interference witn daily acTivities;

3) Perception of receiving enough medical care;

4) General worry;

5) '.Worry about receiving emergency medical services;

5) Worry about having an undeTected disease.


The fourth question was: will the clinic be compared favorably to

other health-care services received? Four questions comparing the

physician's assistant clinic to other health-care agencies were asked.

The comparison questions were in regard to:

1) Cost;

2) Perceived quality of care;

3) Ease in making appointments; and,

4) Ease in talking to clinic staff;

and are shown in the Wave II preceded interview schedule in Appendix B.

The possible responses to these questions were "the clinic was better

or easier than the other health-care service," "it was about the same as

the other health care service," and "it was worse or harder than the

other health care service."

Nonutilizers of the clinic typically would not respond to the

comparison questions and those who did generally indicated they were

relying on hearsay information. Thus, their responses were not analyzed.

The fifth question was: will utilization of the clinic affect the

health-seeking and utilization behaviors of those in the community? Ihe

items that address this question are changes in health-related behaviors

and attitudes as measured by changes in enabling and illness-level

variables. These include changes in physician visits, clinic visits,

and medicine intake. The items show health-related behavior were:

1) Health interference with daily activities;

2) Previous medical-care utilization;

3) Present health-care utilization; and,

4) Stated prospective utilization of the physician's assistant


The items that show health attitudes were:

1) Self-report of perception of state of health;

2) 'orry as measured by response to the question, "How often

are you worried or nervous?";

3) Worry about emergency health care determined by response

to the question, "How often do you worry about being able

to receive health-care services for an emergency?";

4) Worry about personal health determined by response to the

question, "How often do you worry about having an illness

or disease that has not been detected?";

5) Sel'-report of perception of receiving enough medical care;

6) The respondent's and the community's liking of the clinic.

The sixth question: will the clinic be utilized by all segments of

the community?; was addressed by the development of discriminant functions

or utilizers and nonutilizers. These functions contain The variables

which had demonstrated significance.

Reliability and Validity

Reliability and validity must be addressed in all surveys. A

brief review of reliability in this research will be discussed. Reliability

refers to the consistency of the measures over Time. Three types of

reliability were applicable to the survey data: observer reliability,

internal reliability, and, for longitudinal and multistage studies, test-

retest reliability. Experiment observer reliability was demonstrated

by means of a conTrolled simulation. In this experiment, both inter-

viewers recorded answers to survey questionnaires while observing a

simulated interview. The answers of the interviewers were compared.

Tne average agreement was calculated by dividing the number of correct

responses by the total number of questions and dividing by two, the

number of interviews. The average agreement was 92 percent. Since there

were only two interviewers in Wave II, a further statistical procedure

was unnecessary. No observer reliability experiment was performed for

Wave 1.

Other tests of reliability that are frequently used in survey

research are tests to measure the internal consistency and tests to

measure estimate of test-retest reliability. Reliability tests for

internal consistency and test-retest reliability were not performed

on the crucial health related variables. Tests for these types of

reliability were performed for the predisposing, sociodemographic

variables. Because there was nc logical inconsistencies overtime for

these variables, reliability was assumed.

Validity refers to the ability of an item to measure what it

is intended to measure. Content validity was addressed in this research

by having the survey instruments reviewed by other researchers and then

field tested. The results determined if questions were ambiguous and

potentially low in content validity. It was also believed that the

health interview used in this research would not be perceived by the

respondents as personally damaging or threatening to the degree that

they would give inaccurate responses. The interview was assumed to

have face validity.

There are other issues concerning the validity of the data. The

history of health surveys, however, shows that if the limits are recog-

nized a priori, health surveys, such as the one utilized in this

research, are meaningful indicators of health-service utilization. There


was no way to test the validity of the responses. Theoretically, it

would be possible to calculate the validity of some of the survey items

by comparing health utilization responses, with medical records or

health insurance claims. These comparisons, however, were not available

for this research.


The descriptive analysis was performed by utilizing contingency

table statistics. The most prevalently used were Chi-square, sign

tesT, and Gamma. As an attempt to reinforce the descriptive findings

of the impact of the clinic on the county and for future program

planning, clinic utilization was also analyzed by means of discriminant

function analysis. The discussion of this analysis is presented in

the following chapter.


This chapter has outlined the design and methodology used in this

research. The design was a two-stage health survey of adult females

presumed to be most knowledgeable of their family's health needs and

behaviors. A purposive sample of 193 respondenTs representing apprcx-

imately one third of the households in the county was selected. The

sample was asked questions concerning their own health perceptions and

utilization behaviors. They were also asked to be informanTs for the

health perceptions and utilizaTion behaviors of other members of their

families. The items and statistics used in the analyses were described

and a discussion of the issues of reliability and validity was presented.

Chapters I, II and III have described the conceptual issues of the

research and have grouped these issues with relevant research literature

and the research questions. These facTors have led to a discussion of


an analytical framework with which the issues of the research can be

addressed. This chapter has presented the methodological tools used

within the framework. The following chapter presents the culmination

of these four chapters by presenting the research findings.



Chapter I presented a number of questions that had guided this

research. These questions are not formal but are used for the purpose

of interpreting the findings. The findings are reported as responses to

these questions. The major research question was: will the residents of

the county utilize and accept the physician's assistant clinic? To ad-

dress this question, six specific questions are needed.

1) Who in the community will utilize the clinic?

2) For what illnesses or injuries will they utilize the


3) Will they continue to utilize the clinic over time?

4) Will the clinic be compared favorably to other health-

care services received?

5) Will utilization of the clinic affect the health seeking

and utilization behaviors of those in the community?

6) Will the clinic be utilized by all segments of the


Before these questions are addressed in detail, a review of the

technique for measuring changes-over-time and a discussion of the survey

results are presented.

It was noted in tne previous section that only some of the illness-

level items changed between the survey waves. It would be simplistic to


state that the changes in illness-level variables were due solely to

the opening of the physician's assistant clinic. Even in a rural, stable

community iike the research county, other factors not measured in this

research could have caused the changes to occur. Variables that cap-

tured these changes were needed. As described in Chapter IV, by summa-

rizing and creating new variables from the survey data, the needed change

variables were developed.

These variables were then crosstabulated by whether or not an

individual within the family had utilized the physician's assistant

clinic. Thus, the two comparison groups are utilizers and nonutilizers

of the clinic. The data were then compared to determine if, according to

the specific illness-level variable, the utilizers and nonutilizers of

the clinic were independent. Sign tesTs were calculated for the changes

in both groups. Since each group, the utilizers and nonutilizers, had

equal opportunity for change, the sign change would determine which group

changed more significantly. Table 2 presents the general format for the

presentation of These data.


General Format
Change in Predisposing, Enabling, and Illness-
Level Variables as a Function of Utilization of the
Physician's Assistant Clinic

Family Utilized Family Not Utilized
Clinic Clinic

Change in Positive change Positive change
Variables No change No change
Negative change Negat;ve change

In the preceding chapter, the sample design was discussed; a brief

review is needed. When change variables are presented, the maximum N

is 193. These are respondents who participated in both surveys. When

data are presented that do not involve change, the maximum N is 278,

the 193 who responded to both surveys plus the additional group of 85.

Table N's may fall below these totals because of missing responses.

Descriptive Survey Results

The results of each survey are compared and discussed within the

Andersen-Newnan Framework.

Predisposinq variables. Table 3 depicts the predisposing characteristics

of the sample for both survey waves. It can be quickly ascertained that

the respondents who are missing from the Wave II survey do not affect the

sample characteristics. With respect to the predisposing variables,

changes were nonexistent from Wave I to Wave II. This table shows that

the sample were primarily white, married, high school graduates vho were

living in nuclear, complete families within the county for more than twenty

years. The respondents have reported themselves to be rural, stable and

family oriented. More importantly the sample remained the same with

respect to these variables over the year's time.

Approximately, 9 percent reported they or their immediate families

had received some type of welfare during the past year, 90 percent

reported they own at least one car and 60 percent reported ownership

of two or more cars and/or trucks. Data on income were collected but

determined that farmers, some small business owners, and other self-

employed individuals were reporting only net income, the amount remaining

after business costs and depreciations were deducted. Others reported


Wave I and Wave II Survey Comparisons of
Predisposing Sample Characteristics

Wave I Wave I1
Survey Survey

< 35 16 16
35 to 64 62 62
S65 22 22

White 92 94
Black 8 6

Marital Status
Married 79 77
Single 4 2
Widowed 14 18
Separated or Divorced 3 3

Living Arrangement
Nuclear Complete 70 70
Nuclear Incomplete 23 21
Alone 7 9

Length of Time Residing
in Gilchrist County
More than 20 Years 67 70
5 to 20 Years 21 19
Less than 5 Years 12 II

Less than High School Graduate 63 62
High School Graduate 33 37
College Graduate 4 I

N 227


gross incomes. No clear, reliable method of making the amounts comparable

could be ascertained. Income, as a key variable, was deleted from further


Enabling variables. Table 4 presents a listing of these variables. These

health behaviors have remained constant between the two survey waves.

There are, however, two interesting findings. The first is that both

before and after the clinic opened, 80 percent of the sample had seen a

physician. This clearly suggests that the county residents were already

high users of health-care services, and that the opening of a new health-

care service did not radically alter this high health-care service rate.

The second interesting finding was that the use of patent medicines

increased between the survey waves. When this finding was controlled for

clinic utilization, no significant difference was found. One explanation

is that the Wave II interviewers probably explained more precisely what

was meant by patent medicines than did the interviewers in the Wave I


Illness-level variables. Table 5 presents the illness-level variables.

The respondents were asked several questions concerning their attitudes

and perceptions of their state of health as well as their worry about

receiving care. No significant differences were observed between the

two survey waves. Tnere are two interesting findings. In Wave II, the

perception of receiving enough medical care increased by 13 percent.

Secondly, worry about receiving health-care services in an emergency

declined in the "often" and "all the time" categories 17 Dercent, from 28

percent in Wave I to 11 percent in Wave II. These findings suggest that

there has been some impact of the physician's assistant clinic. The

impact is presented more ccmrletely in the next section.


Wave I and Wave II Survey Comparisons of
Enabling, Health Behavior Variables

Wave I Wave I I

Saw A Physician
Yes 80 80
No 20 20

Mean Number of
Physician Visits 4 4.5

Hospital ized
Yes 10.5 9.5
No 89.5 90.5

Took Prescription Druc
During Year
Yes 43 50
No 57 50

Used Patent Medicine
During Year
Yes 19 47
No 81 53

Used a "Nerve Fill"
During Year
Yes 16 25
No 84 75

Women had PaD Test
Yes 44 40
No 56 60

Yes 5 5
No 95 95


Wave I and Wave !I Survey Comparisons of
Illness-Level Variables

Wave I lave I I

Perceived Siate of Health
Good & excellent 65 71
Fair 21 is
Poor and very poor 16 II

Perceived Health Interference
with Daily Activity
Never 50 56
Seldom 12 16
Sometimes 17 6
Often II 13
All the time 10 9

Perceived Receiving Enough
medical Care
Yes 71 34
No 29 16

Never 10 10
Seldom 21 19
Somet i es 30 42
Often 39 30

Receiving Health Care
Services in an Emergency
Never 47 58
Occasionally 25 31
Oftan 25 7
All the Time 3 4

Have Undetected Disease
Never 67 66
Occasionally 23 29
Often and All the Time 10 5

N 227 193

Tables 3 and 4 presented a comparison by survey wave of predis-

posing and enabling variables. These figures show the predisposing

variables and the enabling variables show no dramatic change. Table 5

presented some changes that occurred after the physician's assistant

clinic opened for illness-level variables. Fewer persons perceived

themselves to be in poor health and fewer reported that their health

interfered with their daily activities. A smaller percentage of the

respondents were as worried about receiving medical care in an emer-

gency than during the previous year. Hospitalization and visits to

physicians appeared to be the same. Medicine intake, patent medicine,

prescription drugs, and nerve pills had all increased during the year.

The next section more closely examines these changes by analyzing

the differences in the illness-level variables as a function of family

utilization of the clinic. This analysis provides answers to the

research questions.

Discussion of the Six Specific Questions

The first specific question is: who in the community utilizes

the clinic? Two predisposing variables, age of respondent and occupa-

tion of male head-of-household, are discussed. Tables 5 shows clinic

utilization by occupation of male head-of-household, a predisposing

variable. There is a significant difference between the utilizers and

nonutilizers. The ability to predict utilization by occupation,

however, is minimal, Lambda = .10. Additional analysis showed that of

those who reported they liked the clinic a great deal, 46 percent were

white-collar workers, 23 percent were farmers, 16 percent were craftsmen,

7 percent were operatives, and 8 percent were laborers. Even though the

professional and white-collar workers indicate a greater liking for the


Clinic Utilization by Occupation
of Male Head-Of-Household

Clinic Utilization
Occupation No Yes Total

Professional and
White Collar 8 (12) 36 (40) 44

Foremen 13 (20) 17 (19) 35

Operations 13 (20) 9 (10) 22

Laborers 3 (12) 7 (8) 15

Farmers 20 (29) 20 (23) 40

Total 67 89 156

Chi-square = 15.85 Significant p <.01 d.f. = 4


physician's assistant clinic, Iheir utilization is not significantly

different from other occupational groups. Additional analysis revealed

there was no utilization difference between those who liked the clinic

a great deal and those who liked the clinic somewhat.

Table 7 presents clinic utilization by age. Age was not a signif-

icant factor in clinic utilization. In fact, the percentages of

utilizers and nonutilizers in each age group are almost equal.

As further explication of this utilization question, the idea that

heavy utilizers and needers of health-care services would be more prone

to use the clinic was analyzed. No attempt was made to develop a pro-

file of needs assessment for the research county. Health needs for this

analysis are defined as health-seeking behavior. Since 80 percent of the

sample had sought health care before the clinic had opened, it was

assumed these individuals viewed themselves as having a need for health

services. It was established from the survey that the majority of

respondents perceived they were or had been receiving enough medical

care. Because the health-care needs of the research county were already

perceived as met, the clinic would then have to replace existing pro-

viders or create a new demand for services. In order for the residents

of the county to substitute clinic services for existing services, the

clinic would have to become an accepted, viable, and preferred provider

of health-care services.

Table 8 presents clinic utilization by family physician visits.

The significant Chi-square is interpreted to mean that the utilizers

and nonutilizers are independent groups but, more importantly, the

Gamma of .39 implies there is a moderate association between moro


Clinic Utilization by Respondent's Age

Clinic UtilizaTion
Age of No Yes Total

15-24 4 (3) 4 (3) 8

25-34 18 (13) 18 (13) 36

35-44 29 (21) 32 (23) 61

45-54 32 (23) 29 (20) 61

55-64 26 (19) 25 (18) 51

Over 65 29 (21) 32 (23) 61

Total 138 140 278

Chi-square = 1.054 Not significant d.f. = 5


Clinic Uti ization by
Family Physician Visits

Clinic Utilization
Family Physician Visits Ho Yes Total
% %

None 49 (42) 20 (18) 69

1-5 44 (38) 50 (46) 94

6-10 II (9) 1S (16) 29

More than 10 13 (II) 22 (20) 35

Total 117 110 227

Chi-square = 16.375 Significant p = .001 d.f. = 3

physician visits and clinic utilization. This suggests the more indi-

viduals use health-care services, the more likely they will utilize

the physician's assistant clinic.

Table 9 adds a caution to this interpretation. This table pre-

sents family physician visits by prospective clinic utilization. The

significant Chi-square suggests that prospecTive utilizaTion plans are

independent of family physician visits. The Gamma of -.25 further

suggests that the greater the number of physician visits the lower the

prospective clinic utilization, and, further the data indicate that

the more visits to a physician a family has, the more likely it is they

do not plan to use the clinic. Other survey data verified this impli-

cation in that only eight of the respondents reported that a physician

ever referred them to the physician's assistant clinic. It may be that

the physicians who treat the residents of the research county do not

want to lose patients or perhaps these physicians have not accepted

the physician's assistant clinic as a viable alternative for health-

care delivery.

The second specific question is: for what reasons do the indi-

viduals utilize The clinic? The respondents who had utilized the

clinic answered that physical, checkups, and colds were the reasons

they haa gone to the clinic. These responses are not specific enough

to address the research question. It was decided this question could

be addressed by examining the changes in certain health factors of the

utilizers and nonutilizers. Table 10 shows there was not significant

difference for the utilizers and nonutilizers of the clinic with respect

to change in perceived state 3f health. The sign test, however,

suggests the charge among the utilizers had a significantly greater


Family Physician Visits by
Prospective Clinic Utilization

Family Physician Visits

Prospective More Row
Utilization None 1-5 5-10 than 10 Total
% d

All Services 15 (46) 22 (25) 5 (IS) 3 (9) 45

Some Services 11 (33) 39 (44) 15 (53) 19 (56) 84

Only or Not
at All 7 (21) 28 (31) 8 (29) 12 (35) 55

ToTal 33 89 28 34 184

Chi-square = 13.428
Gamma = -.25

Significant p< .05 d.f.


Change in Perception of State of Health,
Preclinic, Postclinic Opening Controlled
for Clinic Utilization

Clinic Utilization
Perceived State Row
of Health No Yes Total
% %

Setter State
of Health 26 (31) 41 (37) 67

Same State
of Health 39 (47) 48 (44) 87

Worse State
of Health 18 (22) 21 (19) 39

Total 83 i10 193

Chi-square = .74 Not significant d.f. = 2

Sign test
nonutilizers Z = 1.21 accept null hypothesis p = .05
utilizers Z = 2.25 reject null hypothesis p = .05
overall Z = 2.719 reject null hypothesis p = .05


probability to be positive. Table 11 demonstrates that the perception

of health interfering with daily activities was different between

utilzers and nonutilizers. The sign test reports the change for non-

utilizers had a significantly greater probability towards more inier-

ference. Table 12 shows lhat perception of receiving enough medical

care did not differ between utilizers and nonutilizers by family

clinic-participation. The sign tests indicate the probability for

change among both groups was toward a perception of receiving more care

during the second year than during the first year. The difference is

substantially significant.

Table 13 shows there was no difference in general worry by clinic

utilization. The sign test indicates the probability for change among

both clinic utilizers and nonutilizers was toward less worry. Table 1i

shows that there was a significant difference in worry about receiving

emergency medical-care services. The sign test further clarifies the

utilizers had a significantly greater probability to be less worried

about receiving emergency medical-care services. Although there was

some change in worry among the nonutilizers, the probability of the

change being either positive or negative appeared equal and random.

Table 15, more clearly shows that utilization of the clinic may

have affected specific worry. This table presents the change in worry

about having an undetected disease. The sign test indicates the

changes between groups is inverse. The nonutilizers had a significantly

greater probability toward more worry than the utilizers.


Change in Perception of Health Interference
with Daily Activities, Preclinic, Postclinic
Opening Controlled for Clinic Jtilization

Clinic Utilization

Perception of Row
Health Interference No Yes Total
So S

Less Interference 15 (18) 29 (26) 50

Same Interference 28 (34) 52 (48) 92

More Interference 40 (48) 29 (26) 85

Total 117 110 227

Chi-square = 9.35

Sicn test

Significance p <.0! d.f. = 2

= 3.37
= 0
= 8.99

reject null

reject nul



p = .05

p = .05


Change in Perception of Receiving
Enough Care, Preclinic, Postclinic Opening
Controlled for Clinic Utilization

Clinic Utilization

Perception of Row
Receiving Enougn No Yes Total

More Care than
First Year 13 (16) 24 (22) 37

Same Care as
First Year 64 (78) 80 (73) 14r

Less Care than
First Year 5 (6) 5 (5) 10

Total 82 109 191

Chi-square = 1.23 Not significant d.f. = 2

Sign test

= 1.88 reject null
= 3.5 reject null
= 3.9 reject null


= .05
= .05
= .05


Change in Amount of General Worry, Preclinic, Postc!inic
Opening Controlled for Clinic Utilization

Clinic Utilization

General Row
Worry No Yes Total
% %

Less Worry 65 (78) 50 (69) !15

More Worry 18 (22) 22 (31) 40

Total 83 72 155

Chi-square = 1.59 Not significant

d.f. = i

Sign tesT
overall l

4.94 reject null
3.30 reject null
6.83 reject null


= .05
= .05
= .05


Change in Amount of
Medical Services,

Worry about Receiving Emergency
Preclinic, Postclinic Opening
for Clinic Utilization

Clinic Utilization

Worry about Receiving Row
Emergency Medical Services No Yes Total

Less Worry 18 (22) 50 (50) 58

Same Worry 43 (54) 46 (45) 89

More Worry 19 (24) 5 (5) 24

Total 80 i01 181

Chi-square = 21.175 Significant p< 0.001 d.f. = 2

Sign test

Z = -.16
Z = 6.06
Z = .59

accept null hypothesis
reject null hypothesis
reject null hypothesis


Change in Amount of Worry about Having an
Undetected Disease, Preciinic, Postclinic
Opening Controlled for Clinic Utilization

Clinic Utilization

Worry about Having Row
an Undetected Disease No Yes Total

Less Worry 11 (14) 26 (25) 37

Same Worry 47 (60) 63 (61) 110

More Worry 2! (26) !4 (14) 35

Total 79 103 182

Chi-square = 6.761 Significance p <.05 d.f. = 2

Sign test
uti izer

= -1.77 reject null hypothesis
= .89 reject null hypothesis
= .23 accept null hypothesis

= .05
= .05
= .05

Using bivariate contingency-table analysis and an extension of

the binomial test showed that with regard to perceived state of health,

nealth interference with daily activities, perception of receiving

emergency medical care, and worry about having an undetected disease

that there was an overall significantly positive change for the utilizers

of the clinic. The utilization of the clinic demonstrates an inverse

relationship with perception of worry about receiving emergency medical

services. After a family utilized the clinic, there was a reduction of

specific health-related worry.

The major reported reasons for clinic utilization were generic.

The utilization, however, resulted in significanT, positive changes

in health perception and worry. Thus, iT was shown that, in addition

to direct health care, the utilizers of the clinic had more positive

change in health-related perceptions.

The third specific question is: will the county residents continue

to utilize the clinic over time? It addresses the continuance of

prospective utilization of the clinic. The question in the Wave II sur-

vey which addressed prospective utilization is: "Would you go

to the medical center in the county seat for all your health-care needs,

some of your health-care needs, only in the case of an emergency, or

not at all?" The respondents reported as follows: I would go to the

clinic for all health-care needs (22 percent), some (46 percent),

emergency only (22 percent), and not at all (10 percent). These figures

raise additional questions. Is there a difference in prospective

utilization if one has already utilized the clinic? Do differences

in perceived quality of care result in a difference in prospective

utilization? Both utilizers and nonutilizers staTed that their


prospective utilization of the physician's assistant clinic would be

for some of their medical care but not for all of it. These data suggest

that the clinic will be utilized for some of the health-care needs of

the community.

Table 16 presents utilization with respect to intended future

utilization. The clinic utilizers are a significantly different group

than the nonutilizers in regard to prospective utilization. The Chi-

square is significant at p C .05. The ability to relate actual utili-

zation from prospective utilization is moderately low with Gamma = .24.

These data indicate a tendency. Once the clinic is actually utilized,

the user is somewhat more likely to return to the clinic for some

of his(her) health-care needs. Furthermore, those who used the

clinic were quite satisfied with the care rendered. Of those utilizers

who liked the clinic a great deal, 90 percent could not list anything they

disliked. Consistently, those who utilized the clinic perceived both

the cost and quality of care to be the same or better than their usual

place of health-care services. Of those who intended to use the clinic

for some or all of their health-care-service needs, 65 percent perceived

the cost of services at the physician's assistant clinic to be less than

the cost of services at their usual place of health-care services.

These findings suggest that some dissonance of acceptance or non-

acceptance by the community of a nonlicensed primary health provider

was resolved after a member of the family had utilized the clinic. The

reason for the initial utilization could not be obtained from the survey

data. After the utilization of the physician's assistant clinic the

user probably felt more favorable about the clinic in that the user was

more likely To continue to use the clinic for some of his health-care

needs than the nonutilizers.


Clinic Utilization by
Prospective Clinic Utilization

Clinic UtilizaTion

Prospective Row
Utilization No Yes Total

All Services 9 (12) 36 (33) 45

Some Services 23 (31) 61 (56) 84

Emergency Only
or Not at All 42 (57) 13 (ll) 55

Total 74 110 184

Chi-square = 43.295
Gamma = .24

Significant p .05 d.f. = 2


A logical additional inquiry was made to determine if the perception

of quality of care of clinic utilizers was an important factor in their

decision for prospective utilization. It may be inferred from i able 17

that there was a difference of perceived quality of care at the clinic

by planned future use of the clinic. Those who intended to utilize ihe

clinic for some or all of their health-care-service needs appeared to

perceive the clinic to be providing the same if not better quality of

care than was otherwise available. Those who did not plan to use the

clinic appeared to perceive the quality of care to be the same or worse

than the quality of other care. Because so few responded that the care at

the clinic was not as good as other providers and so few responded that

they do not intend to use the clinic at all, it would be misleading to

perform any statistical test on these data.

The idea that the health-care needs of the county were being met

by the physician's assistant clinic can not be statistically supported

with these data. There is an indication that those who have utilized

the clinic will continue to. These data also lead to the speculation

that the residents were perhaps rigidly entrenched in their previous

patterns of health-care-services utilization and they did not perceive

the need of the physician's assistant clinic for the major portion of

their health-care-service needs. Table 18 presents the change in physi-

cian visits by clinic utilization. No differences were shown. It was

thought that if there were more physician visits, then there would be

lower clinic utilization. It was also thought that if clinic utilizers

had an increase in physician visits, then the physician's assistant had

identified patients to be referred to physicians for additional special--

ized treatments, thereby causing these patients' utilization rates of


Perception of a Physician's Assistant
Clinic as Compared to Usual Care Received by
Prospective Physician's Assistant Clinic Utilization

Perception of Care at Clinic
Compared to Usual Care Received

Prospective Row
Utilization Better Same Not Good Total

All Care 8 30 0 38

Some Care 8 54 3 65

Care Only I 5 3 9

Not at All 0 I I 2

Total 17 90 7 114


Change in Number of Physician Visits,
Preclinic, Pcstclinic Opening
Controlled for Clinic Utilization

Clinic Utilization

Physician Row
Visits No Yes Total

More MD Visits 15 (29) i6 (25) 38

Same MD Visits 20 (40) 27 (42) 47

Less MD Visits 16 (31) 22 (33) 31

Total 51 65 116

Chi-sauare = 2.78

Sign test

Not significant d.f. = 2

Z = -.17 accept null hypothesis
Z = -.97 accept nuil hypothesis
Z = .84 accept null hypothesis

p = .05
p = .05
p = .05

physicians io increase. These arguments cannot be verified in the survey

data. The point here is that the residents were already users of health

services. The survey reported that 80 percent of the informants had

utilized some type of health-care service before the clinic had opened.

For a county with no health-care services, this is a high percentage. The

clinic, then, should not be viewed as the only provider of primary care,

it should be viewed as a more convenient alternative.

Table 19 presents change in hospitalization by clinic utilization.

The utilizers and nonutilizers are shown to be independent groups. The

nonutilizers, however, had a significantly greater probability for less

hospitalization in Wave I1 than in Wave I. The change in hospitalization

for utilizers was random. The idea then, that the physician's assistant

clinic was functioning as a referral source, is speculative at best.

Table 20 presents change in utilization of prescription drugs. There

was no difference between the utilizers and nonutilizers of the physician's

assistant clinic nor was there any significant probability of the direc-

tion of change overall among the groups. As an indicator of preventive

medical-care pract-ice, data concerning Pap tests were collected in the

surveys. Table 21 shows that there was no significant difference be-

tween the physician's assistant clinic utilizers in regard to consis-

tently, yearly, having a Pap test. There was no significant probability

of the direction of the change overall.

These findings regarding health-care utilization indicate -here has

been a minimal change which could be attributable to the utilization of

the physician's assistant clinic. The reason the clinic had no dramatic

affect on the health-seeking behaviors of the respondents during this


Change in Amount of Hospitalization,
Freclinic, Postclinic Opening
Controlled for Clinic Utilization

Clinic Utilization

Hospitalization No Yes Total
% %

Less Hospitalization 39 (33) 1i (10) 50

Same Hosoitalization 69 (59) 86 (78) 155

More Hospitalization 9 (S) i3 (12) 22

Total 117 110 227

Chi-square = 18.073 Significant

Sign tesT

p(0.00I d.f. = 2

= 3.04 reject null
= -.04 accept null
= 2.62 reject null


S= .05
p = .05
p = .05


Chance in Utilization of Prescription
Drugs, Preclinic, Postclinic
Opening Controlled for Clinic Utilization

Clinic Utilization

Utilization of Row
Prescription Drugs No Yes Total

Taking More
Prescription Drugs 6 (7) 8 (7) 14

Taking Same
PrescriDtion Drugs 66 (80) 90 (82) 156

Taking Less
Prescription Drugs II (13) 12 (11) 23

Total 83 110 193

Chi-square = .25 Not significant d.f. =

Sign test
nonutiiizer Z = 1.21 accept null
utilizer Z = -.89 reject null
overall Z =-1.08 accept null


= .05
= .05
= .05


Change in Having Fap Test,
Preclinic, Postclinic
Opening Controlled for Clinic Utilization

Clinic Utilizaticn

Having Row
Pap Test No Yes Total
% c'

Had Pap Test
Second Year,
Not First 12 (i5) 9 (8) 21

No Change 60 (72) 85 (77) 145

Had Pap Test
First Year,
Not Second il (13) !6 (15) 27

Total 83 110 193

Chi-souare = 1.93 Not significant d.f. = 2

Sign test

Z = .21 accept null hypothesis p = .05
Z =-1.40 accept null hypothesis p = .05
Z = .86 accept null hypothesis p = .05

research time-frame may be due to the already high health service


The sixth question is: will the clinic be utilized by all segments

of the county? The method for addressing this question is to use the

findings thus far reported to make explicit and test a model of clinic

utilization with each wave of survey data. The rationale for this analysis

was to attempt a multifactor description of acceptance by the county

of the physician's assistant clinic. It was also desired to determine

if the collected descriptive data could be interpreted meaningfully by

a multivariate statistical technique. Because there are two survey

waves, answers TO questions regarding who is more likely to use a new

health-care service before and after it was available could be determined

and the results then compared. If there are significant differences in

the characteristics of these two groups, the results would be of interest

to health-care planners and policy makers.

To this point, The analysis has used bivariate contingency table

statistics. Several variables, however, have been shown to affect

utilization of the physician's assistant clinic. They were perceived

state of health, perceived health interference with daily activity,

worry about receiving health care in an emergency, utilization of

prescription drugs, family utilization of physician services, percep-

tion of receiving enough medical care, occupation of the male head-of-

household, and prospective utilization of the physician's assistant

clinic. The interaction between these variables has not been explored.

A multivariate analysis was undertaken to determine if these variables

together would influence physician's assistant clinic utilization.

The multivariate technique chosen was discriminant analysis.

The purpose of this analysis was to determine if certain items or

factors could predict and classify who would utilize the clinic.

Several utilization models were discussed in Chapter II. These models

focus on current utilizers of medical services and their illness

symptoms. Andersen and Newman state, "The Model should also serve

as a guide in the selection of relevant variables to include in the

analysis" (1973, p. 106).

The objective of discriminant analysis is to develop one or more

functions from a given set of variables which will discriminate

between members of the various groups. Discriminant analysis derives

the function by maximizing the between group differences and minimizing

the within group variances. The two groups used in this analysis are

those who reported themselves or their families as utilizers of the

clinic and those who did not report any utilization.

Discriminant analysis provides two kinds of resulTs. First, after

a group of variables are found to discriminate between utilizers and

nonutilizers, the analysis selects those variables of a group with the

greatest power to discriminate between these utilization categories.

Second, the variables selected by the analysis are checked with the

observed survey data to determine the percentage of the data that the

model predicts.

From the data collected before the clinic opened, the discriminant

analysis determined that the greater differences between utilizers and

ronutilizers was a function containing the items, occupation of male

head-of-household, health interference with daily acTivities, worry about

receiving emergency medical care, and distance of home from the

physician's assistant clinic. This implies that the higher the occupa-

tional status, the greater the worry about receiving emergency

medical care, and the closer one lives to the clinic, then ihe more

likely one was to use the clinic. These findings are representative

of the predisposing, enabling, and illness-level components of the

framework. Table 22 presents the prediction results for Wave I da+a.

The model predicted that, out of the fcrly-four actual nonutilizers,

thirty-seven would not utilize the clinic within the first year. It

also predicted that, out of the sixty utilizers, forty would utilize the

clinic within the first year. This means the nonutilizers have been

accurately identified by these variables 84.1 percent of the time and

utilizers 66.7 percent. The weighted average correctly classified was

74 percent.

The discriminant function derived from Wave II data was different.

The reason the set of variables has changed is that the differences

between variable means for utilizers and nonutilizers in Wave II are

closer to zero than in Wave I. In general, this means that both utilizers

and nonuiilizers experienced change during the year. Furthermore, the

selected set of variables in the Wave I function no longer clearly

discriminates differences. Utilizers and nonutilizers in Wave II cai

no longer be classified as different in terms of perception of health,

health interference with daily activities, and worry about receiving

emergency medical care. Occupation of male head-of-household remained a

classifying item in Wave II, but the difference added less discriminat-

ing power to the function than it did in Wave I.


Discriminant Functions for Wave I
Prediction Results

No. of Predicted Group Membership
Actual Group Cases Nonutilizer Utilizer

Nonutilizers 44 37 7

Utilizers 60 20 40

Percent of "Grouped" Cases Correctly Classified: 74.04 percent

The discriminant function that determined the greatest difference

between utilizers and nonuti I izers in Wave 1I was a function composed

of prospective utilization, distance from home to clinic, and perception

of receiving enough care. The utilizers used the clinic if they intended

to use the clinic for some or all of tneir health-care needs, lived

close to the clinic, and perceived that they were receiving enough

medical care. Table 23 presents the prediction results. These vari-

ables were able to accurately predict nonutilization 81 percent of the

time and utilization 65.6 percent of the time. The segments of the

framework represented were distance from the clinic, an enabling compo-

nent, and perception of enough care, an illness-level component. The

predisposing component was no longer included.

Tables 24 and 25 depict the comparison of the means of the selected

variables for Wave I and Wave II. Tables 26 and 27 depict the comparison

of selected variables by utilize and nonutilizers. These four tables

delineate tne items and their importance in The discriminate functions.

The data have shown the clinic has not gained its share of poten-

tial utilizers in tne counTy. The results of the discriminant analysis

isolated predictive determinants of potential utilizers. The policy

makers of the clinic must decide if this is the segment of the county

that they want to serve.


This chapter presented the results and the interpretations of

The survey findings. Six specific questions were addressed. The com-

position of each survey was presented. The lack of change or impact of

the survey items was demonstrated, as each of the six specific questions

was addressed. The conclusion drawn from these analyses addressed the


Discriminant Functions for Wave II
Prediction Results

No. of Predicted Group Membership
Actual Group Cases Nonutilizer Utilizer

Nonutilizers 42 34 8

Utilizers 64 22 42

Percent of "Grouped" Cases Correctly Classified: 71.70 percent


Comparison of Wave I and Wave II
Variable Means for Nonutilizers

Variables Wave I Wave II

Occupation 3.0682 2.9296

Marital Status 1.0682 1.0000

Perceived State of Health 1.3182 1.3571

Perception of Health 1.8182 1.7857

Worry About Emergency 1.5909 1.6667
Health Care Services

Use of PrescripTion Drugs 1.6364 1.6190

Perception of Enough Care 1.2045 1.1667

Sector 2.4773 2.4524

Family Physician Visits 0.8636 1.6667


Comparison of Wave I and Wave II
Variable Means for Utilizers

Variables Wave I Wave II

Occupation 2.3167 2.3906

Marital Status 1.0333 1.0156

Perceived State of Health 1.3500 1.3125

Perception of Health 1.6167 1.7188

Worry About Emergency 2.1333 1.5000
Health Care Services

Use of Prescription Drugs 1.6000 1.5469

Perception of Enough Care 1.2500 1.0469

Sector 1.8667 1.8750

Family Physician Visits 0.9333 1.5625

Table 26

Wave I Variable Means by Utilization

Variables Nonutilizers Utilizers

Occupation 3.0682 2.3167

Marital Status 1.0682 1.0333

Perceived State of Health 1.3182 1.3500

Perception of Health 1.8182 1.6167

Worry About Emergency 1.5909 2.1333
Health Care Services

Use of Prescription Drugs 1.5364 1.6000

Perception of Enough Care 1.2045 1.2500

Sector 2.4773 1.8667

Family Physician Visits 0.8636 0.9833


Wave II Variable Means by Utilization

Variables Nonutilizers Utilizers

Occupation 2.9286 2.3906

Marital Status 1.0000 1.0156

Perceived State of Health 1.3571 1.3125

Perception of Health 1.7857 1.7188

Worry About Emergency 1.6667 1.5000
Health Care Services

Use of Prescription Drugs 1.6190 1.5459

Perception of Enough Care 1.1667 1.0469

Sector 2.4524 1.8750

Family Physician Visits 1.6667 1.5625

Prosoective Utilization 2.6429 1.8281

the major research question which asked, did the county accept, as

measured oy clinic utilization, the physician's assistant clinic as

a viable source of primary health-care delivery. In addiTion to

oivariate analysis, a nultivariate analysis was undertaken. The latter

analysis is limited by the data. The items used in the discrininant

analysis are primarily ordinal. It would be erroneous to discuss the

magnitude of the items in each function. A discussion of the direction

of each variable, however, is valid. The analysis does lisT the ranking

of the response categories. For example, the Wave I discrininant

function demonstrated that occupation of the male nead-of-househcld

was a key variable in the pattern of utilization. The higher the

occupation score then, the greater the prediction of clinic utilization.

This means that white-collar workers are more prone to use the clinic.

Because of the limitations of the data, it does not mean that white-

collar workers have an exact mathematical probability of using the

clinic tnat is greater tnan the other occupational groups.

The findings and interpretations are assembled and discussed in the

next chapter, Summary and Conclusions.

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs