MSU

LIBRARIES

 

 

RETURNING MATERIALS:
PTace in bodk drop to
remove this checkout from
your record. FINES will
be charged if book is
returned after the date
stamped beIow.

 

 

 

 

 

THE RELATIONSHIP BETWEEN PERCEIVED
FAMILY FUNCTIONING AND PERCEIVED
COMPLIANCE BEHAVIOR OF THE AMBULATORY
ADULT ONSET DIABETIC

BY

Barbara Bye Sipes

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF NURSING

School of Nursing

1981

ABSTRACT

THE RELATIONSHIP BETWEEN PERCEIVED
FAMILY FUNCTIONING AND PERCEIVED
COMPLIANCE BEHAVIOR OF THE AMBULATORY
ADULT ONSET DIABETIC

BY
Barbara Bye Sipes

A descriptive study of 25 adult onset diabetics was
undertaken to determine the relationship between the
patient's perceived family functioning and her/his perceived
compliance with selected aspects of the therapeutic regimen.
The data were gathered by means of a questionnaire with two
parts that was cOmpleted by the patient. The questionnaire
was given one time only to each subject and was administered
by the researcher to the first 25 adult onset diabetic
patients who were between the ages of 30 to 70, lived with
at least one other person, and were being treated by one of
four private physicians. .

The data were analyzed by using Peason
Product-Moment correlation coefficient. The findings
revealed no significant relationship at the .05 level
between perceived family functioning and perceived
cOmpliance behavior of the ambulatory adult onset diabetic.
HOvever, the nurse should continue to investigate the
influence of family functioning on the health state of the

diabetic as well as investigate barriers to compliance.

To My Husband
Gary
and
Children

Dan, Susan, Laura, and Linda

ii

n-

v»:

to

.—.
uh.

ACKNOWLEDGMENTS

I would like to thank Dr. Barbara Given, Ucommittee
chairperson, for her assistance and for the contributions
made by my thesis committee members, Bonnie Elmassian, Carol
Franck, and Mary Horan.

I appreciate the assistance and cheerfulness of
Karen Hollinger, Mrs. Kelly, Colleen Hollenbeck, and Debbie
Hessel, the physicians' nurses who helped me identify
patients for the study. I am grateful to the physicians Dr.
Orestes Iung, Dr. James Kelly, Dr. Eric Gutscher, and Dr.
Marvin Penwell for. their permission to conduct research
using patients from their private practices.

A thank you is extended to Brenda Walworth who typed
for many hours for several days in an effort to meet a dead
line for a preliminary draft of the thesis. I also wish to
thank Kay Ashmore for doing a professional job drawing my
model.

A thank you is also extended to Dr. Isabelle Payne,
Dr. Dugan, and Dr. Elaine Chernay for their support and
encouragement to keep trying.

I will be forever grateful to Dr. Rita Gallin for
Patiently, persistently, and'selflessly devoting many hours
this summer in assisting me to develop an_acceptable writing
Style. I

A special thank you is extended to David Moore, Dean

of Computer Science at Mott Community College, for granting

iii

me permission to type my thesis on the computer and for
spending several hours teaching me how to use the computer.

I am grateful to my children, Dan, Susan, Laura, and
Linda, for assistance in maintaining our family functioning
in spite of the strain completing this thesis placed upon
our family. My parents, Weston and Lillian Bye, are also
thanked and appreciated for their support and concern.

And finally, my greatest appreciation is extended to
my husband, Gary, who assumed many of the household

responsibilities, and encouraged and supported me every step

of the way.

iv

TABLE OF CONTENTS

LIST OF TABLES O O O O O O O O O O O O O O O O 0

LIST OF FIGURES O O O O O O O O O O O O O O O O 0

Chapter

I.

II.

III.

THE PROBLEM . . . . . . . .

Introduction . . . . . .
Purmse O O O O O O 0
Statement of Problem . .
Statement of Hypothesis .
Subresearch Questions . .
Definition of Concepts .
Subresearch Questions
Ambulatory . . . . . .
Therapeutic Regimen .
Perceived Compliance Behavior
Perceived Family Functioning
Descriptive Data . . . . . . .
Confounding Variables .
Limitations . . . . . .

Assumptions . . . .
Overview of Study . .

CONCEPTUAL FRAMEWORK O O O O O O O O O O 0

Introduction . . . . . . . . . . . . .
Compliance in Persons with Chronic Illness
Family Functioning and Family' 3 Influence
on Compliance Behavior . . . . . . . . .
Summary . . . . . . . . . . . . . . . . . .
Nursing Theory . . . . . . . . . . . . . .

REVIEW OF THE LITERATURE . . . . . . . . .

Introduction . ... . . . . .
Compliance Behavior . . . . .
Magnitude of Noncompliance
Variables Reviewed . . . .
Demographic Variables .
Knowledge . . . . . . .
Disease Features . . .
Features of the Regimen

QQO‘QO’IO‘O‘U'N l-‘

50

50
51
51
52
53
55
57
59

IV.

Approaches to Measure Compliance . .
Objective Measures . . . . . . .
Clinical Outcome Measures . . . .
Subjective Measures . . . . . . .

Summary of Compliance Literature Revi

Chronic Illness - Diabetes . . . .
Family Functioning and the Family' 8 In

(D
0000.30

0
coo-emcee...

on Compliance Behavior . . .
Family Functioning . . . . . .
Measurement of Family Functioning
Family's Influence on Compliance .

summary 0 O O O O O O O O O O O O O O

ooooofiomoooo

METHODOLOGY AND PROCEDURE . . . . . . .

Overview . . . . . . . . . .
Operational Definitions of the Variables
Perceived Family Functioning . . . .
Perceived Compliance Behavior.
Clinical Variables . . . . . .
Confounding Variables .
sample 0 O O O O O O O O O
Source of Sample . . . . .
Data Collection Procedure .
Human Rights Protection . .
The Instrument. . . . . . . .
Perceived Family Functioning
Perceived Compliance Behavior
Reliability. . . . . . . . .
Validity . . . . . . . . . .
Research Question, Hypothesis, and S
Questions . . . . . . . . . .
summary 0 O O O O O O O O l O O O O

b e e r h

0.0000000000000000

oomooooooooooooo
oomococooooooooooo

o
I

oofioocooooooooo

DATA PRESENTATION AND ANALYSIS . . . .

Overview . . . . . . . . . . . . . . . .
Hypothesis . . . . . . . . . . . . .
Subresearch . . . .

Descriptive Findings of the Study Sample

Reliability of Instrument . . .

Data Presentation for Hypothesis
The Statistical Technique . .
Hypothesis . . . . . . .
Subresearch Questions .

Clinical Parameters . . .

Confounding Variables . .
Confounding Variable 1
Confounding Variable 2
Confounding Variable 3a
Confounding Variable 3b
Confounding Variable 3c

vi

106
107
108
113
113
116
117
117
120
123
125

126
128

1.30

130
130
131
131
138
140
140
140
142
143
145
145
146
146
146
146

.1.

....

Confounding Variable 3d
Confounding Variable 4
Summary . . . . . . . . .

VI. SUMMARY OF FINDINGS, INTERPRETATION,

AND INTERPRETATION .

Overview . . . . . . . . .

Summary and Interpretation of Findings . . . . .
the Study

Descriptive Findings of
HYPOtheSiSo O O O O O O
Subresearch Questions .
Clinical Parameters . .
Confounding Variables .

Nursing Implications for Practice.
Implication for Nursing Education.
Implications for Future Research .

Summary. . . . . . . . . .

APPENDICES

Chart Audit Work Sheet
Letter to Physician . .

to the Study . . . .
Introductory Letter . .
Consent Form . . . .
Family APGAR Tables . .

curau:m. HquUt3u1w

BIBLIOGRAPHY O O O O I O O O 0

vii

Check List of Patients Admitted

Questions Used for Data Analysis

Confidential Patient Data Sheet .
Questionnaire Family Functioning Section
Questionnaire Compliance Behavior

Population.

Section

147
147
147

149

149
149
149
152
154
161
167
175
184
186
191

192
195
199
203
204

206
207
208
209
213

216

Table

U10§UN

LIST OF TABLES

Sociodemographic Data by Age, Sex, Race.
and Income of Participants . . . . .

Living Arrangements of Participants . .

Number of Chronic Illnesses of Participants .

Page

0 O 132

. . 133
. . 135

Types of Other Chronic Illnesses of Participants . 135

Complexity of Regimen Measured by Number
Medications Other than Medication for
Diabetes 0 O O O O O O O O O I O 0 0

Complexity of Regimen Measured by Number
Times per Day Taking Medication . . .

Complexity of Regimen Measured by Number

Diet Restrictions Other than the
Diabetic Diet 0 O O O O O O O O O O 0

viii

of

O O 136

O O 136

. . 137

:.‘...,
uo1ib

 

LIST OF FIGURES

Figure Page
1 An Integration of the Nursing Process as it
Relates to Family Functioning and Compliance

Behavior in Chronic Illness . . . . . . . . . . . . 49

ix

CHAPTER I

THE PROBLEM

Insuring-.1911

The management of a chronic illness by a patient may
require that he/she will have to acquire new behaviors in
order to live with the illness. This new behavior we call
compliance behavior since the patient must comply with a
suggested health regimen. Reviews of literature that relate
to compliance behavior in all illness regimens reveal that
at least one- third of all patients do not comply with their
therapeutic regimen (Haynes et al., 1979; Blackwell, 1973).
In view of this lack of success, compliance currently may be
one of the greatest therapeutic challenges facing the health
professions (Gillum & Barsky, 1974).

The most recent and fairly complete review of the
literature on compliance behavior, conducted by Haynes and
associates in 1979, suggests that research into the arena of
compliance has focused on five main areas. These include:
(1) demographic features of the patient such as age, sex,
education, socioeconomic status, income, marital status,
race, religion; (2) characteristics of the disease such as
diagnosis, severity, duration, degree of disability: (3)
characteristics of the therapeutic regimen such as

complexity, duration, type of regimen; (4) psychosocial

features of the patient such as health be11efs and knowledge
of the disease and regimen; and (5) doctor-patient
relationships.

The studies that were reviewed have provided some
useful data. For example, we know that characteristics of
,the regimen have an effect on compliance behavior (Haynes et
al., 1979). That is, noncompliance occurs most often when
the regimen is long in duration, complex, and requires
extensive behaviorial changes. Thus, since each of these
characteristics is a feature of the treatment program for
the chronic illness adult onset diabetes, it seems
reasonable to assume that compliance may be a problem for
the adult onset diabetic.

Few studies, however, have focused on the compliance
behavior of the adult onset diabetics. The few studies that
have focused on the compliance of the diabetic have
investigated the effects of such variables as knowledge of
the disease and therapeutic regimen (Bowen et al., 1961;
Tagliacozzo et al., 1974; Watkins et al., 1967; Williams
et al., 1967), doctor-patient communication (Hulka et al.,
1975), and health-beliefs (Cerkoney & Hart, 1980). This
past.research has provided some useful information.
‘However, few studies have investigated the effecrs of the
family variable on the diabetic's compliance behavior, i.e.,
the influence of the adult onset diabetic's family on the

jpatient's adherence to a therapeutic regimen. Neglect of

pu-

IM 4

£9
.0:

 

‘I
On

p.
O.
b

‘11.

‘1

I!)

this subject is unfortunate since chronic illness usually
has an effect on the individual's family as well as the
patient (Rakel, 1977; Barsky, 1976). Moreover, the manner
in which a family responds to the illness can have a marked
effect on the course of that illness as suggested by Rakel
(1977).

Some studies do exist that investigated the family's
influence on compliance behavior in chronic illness. The
findings of such studies show that there is a positive
relationship between family influence and compliance
behavior. However, these studies were not of diabetes but
rather of arthritis (Oakes, 1970; Ferguson & Boles, 1979),
hypertension (Caplan et al., 1976), orthopedic disability
(Litman, 1966), and kidney disease (Steidl et al., 1980).
Moreover, most of these studies focused mainly on family
support related to adherence to a regimen.

It is reasonable to expect that family support may
be influenced by the overall functioning of the family.
Yet, the writer found only one study that examined the
relationship of family functioning to compliance behavior.
This study was conducted by Steidl and associates (1980) who
found that there was no significant relationship between
total family functioning as perceived by family therapists
and compliance behavior in kidney dialysis patients.
However, Steidl and associates' study (1980) did not

investigate the family functioning as perceived by the

individual within the family. Since an individual's
perception may influence her/his behavior the individual's
perception of her/his family functioning may influence
her/his adherence to a therapeutic regimen. There are some
studies (Pless & Satterwhite, 1973; Smilkstein, 1978) that
have examined the question of the evaluation of perceived
family functioning in relation to health care in general but
not specifically in relation to compliance behavior.

In summary, it is apparent that many studies have
been completed on compliance behavior in chronic illness and
a few studies have considered the influence of family
support on compliance behavior. Some studies have been
completed that deal with compliance behavior in adult onset
diabetes. However, there have been no studies located that
have investigated the relationship between perceived family
functioning and perceived compliance behavior in the adult
onset diabetic. The investigator believes that this is a
significant deficit in compliance research since she
believes that the way in which a family interacts may have
an effect on compliance behavior.

For example, if the relationships among family
members are positive, it might be expected that significant
others would provide support that might motivate increased
compliance such as encouragement, validation of self-image,
and/or reminders. By contrast, if such psychological and/or

functional support was lacking, one might expect that a

per

.1!-
:IL

I

UL

'0'
.lll

‘5:

UV;

'II

1‘.

person would be less motivated to comply. Social support
such as the example just given are reflected in the concept
of family functioning. Thus, it would seem beneficial to
investigate the relationship between family functioning and
compliance behavior.

Therefore, because nurses are involved with
teaching, support, and compliance behavior of ambulatory
patients, the nurse should investigate the impact of family
functioning on compliance behavior. This is so because the
goal of nurses in the expanded role is to help patients
reach their maximum potential for daily living and reaching
this maximum potential, when an individual has a chronic
illness, often depends on compliance with a therapeutic
_regimen. An analysis of information gained from a study of
perceived family functioning and perceived compliance
behavior could be useful to nurses in practice, education,

and research.

211132283
The purpose of this study is to investigate the

relationship between the perceived family functioning and
the perceived compliance with selected aspects of the
prescribed therapeutic regimen of a sample of ambulatory

adult onset diabetics.

 

 

”it. n. n :U

'5 I:

 

ans

Wm

Is there a relationship between the perceived
family functioning and the perceived compliance
with selected aspects of the prescribed
therapeutic regimen of ambulatory adult onset
diabetic patients?

3W
There is a relationship between the perceived
family functioning and perceived compliance with
selected aspects of the prescribed therapeutic

regimen of ambulatory adult onset diabetic
patients.

511W

1. How does the ambulatory adult onset diabetic
perceive her/his family functioning?

2. How does the ambulatory adult onset diabetic
perceive her/his compliance with the
therapeutic regimen?

E E' 'l' E C |
In this section the concepts that will be defined

are (1) adult onset diabetes, (2) ambulatory care, (3) the
therapeutic regimen, (4) perceived compliance behavior, and

(S) perceived family functioning.

W

Diabetes is a disorder of carbohydrate, protein, and
fat metabolism and is classically characterized by
hyperglycemia and glycosuria (Skillman & Tzagournis, 1977).
There are two types of diabetes, adult onset diabetes and

juvenile onset diabetes. Ninety percent of those with the

.
.II‘
gush

11..

...,
554':

fly:
...I
c

gm.
iii

‘5-
‘91

a)

.,,
’ I

“v

‘u

disease have adult onset diabetes (Cohen & Etzwiler, 1976).
Since adult onset diabetes is the most prevalent type of
diabetes, subjects with adult onset diabetes, age 31 to 70,
will be selected for this study.

Adult onset diabetics have the following
characteristics: (1) are diagnosed after age 30, (2) have a
gradual onset of the disease, (3) are not usually insulin
dependent, and (4) are usually overweight (Cohen & Etzwiler,
1976). Adult onset diabetics are less prone to develop
acute complications such as acidosis and coma, but do tend
to develop chronic complications. These chronic
complications are due to the fact that diabetics tend to
develop arteriosclerosis at an earlier age than nondiabetics
(Guthrie & Guthrie, 1977). And, arteriosclerosis frequently
results in the development of other chronic vascular disease
such as coronary artery disease and peripheral vascular
disease.

The criteria to be used for the selection of the
subjects who will participate in this study will be

discussed in more detail in Chapter 4.

mm

Ambulatory is defined as not confined to bed, to
hospital, or to home. Adult onset diabetics who are
confined to bed at home or confined to a health care
facility will be excluded from this study. Being confined

to bed or a health care facility usually results in care

being administered by someone else. This study is concerned
with adult onset diabetics who are responsible for their own
care. However, although subjects will not be under the
direct supervision of a health care provider, they will be
receiving follow-up care for their diabetes at a minimum of

two times per year in an ambulatory care setting.

I] l' B .

The therapeutic regimen is the clinical
recommendation prescribed by the health care provider for
diabetes management. Certain aspects of the regimen are
common to many adult' onset diabetics. For example, diet,
medication, and follow-up appointments are usually given as
written prescriptions by the health care provider.
Therefore, these will be the dimensions of the therapeutic
regimen that will be considered for this study.

First of all, most basic to all diabetic regimens is
a calculated diabetic diet. If no attention is paid to the
diabetic patient's food intake it is impossible to keep
blood sugar levels within prescribed limits (Cohen &
Etzwiler, 1976). For this study, the aspect concerning the
diet will include adherence to general principles such as
eating the recommended foods in the recommended amounts at
the appropriate times.

In addition to a prescribed diet, the diabetic may
be on an oral hypoglycemic or a small dose of insulin (less

than 60 units per day). The aspects considered for this

N t

‘I.c'
61155.

e ,
0 Of
a fit

‘1
\u

study concerning medication will be taking insulin or oral
hypoglycemics in the prescribed doses, at the prescribed
times, and watching for side effects.

Finally, most adult onset diabetics need to be seen
by their health care provider at two to three month
intervals for continued management, although the frequency
of follow-up appointments may vary with the patient, the
nature of the disease, and the physician (Cohen & Etzwiler,
1976). Aspects of the follow-up appointments for this study
include keeping well and ill follow-up appointments and
laboratory appointments for the assessment of blood sugar.

Exercise, urine testing, and general hygiene
including special foot care also are important aspects of
diabetic care. These latter dimensions, however, are not
usually given as written prescriptions, but are often based
on verbal instruction of the health care provider. This
information given per verbal instruction is not always
emphasized equally in all settings and to all types .of
diabetics (insulin dependent as well as non-insulin
dependent). Compliance with exercise, urine testing, and
general hygiene may be indicated by the patient as poor when
in fact the information in these areas may not have been
emphasized and the patient may not perceive them as being
important. Therefore exercise, urine testing, and general
hygiene will not be measured as dimensions of compliance in

this study.

E i i C 1' E l .

Compliance is defined as the extent to which a
patient's behavior (in terms of, for example, taking
medication or following a diet) coincides with medical or
health advice (Haynes et al., 1979). Perceived compliance
behavior indicates the patient's image of reality concerning
her/his adherence to the prescribed regimen. This study
will ascertain the patient's image of reality concerning the
extent to which he/she follows her/his diet, takes
medication, and keeps follow-up appointments as prescribed
by her/his physician.

A one time measurement of perceived compliance
behavior will be used. Due to insufficient objective
parameters to measure compliance behavior with the diabetic
regimen from a one time measurement (Bondy & Felig, 1971;
Leslie et al., 1979; Tattersal et al., 1980), compliance
will be based on the patient's report of how the clinical
prescriptions of the health care provider are implemented.
According to Haynes and associates (1979) self-reporting
compliance behavior can be as effective a means of measuring
compliance as more objective means. A study by Green and
colleagues (1979) substantiates the effectiveness of self-
report. By contrast, Ferguson & Boles (1979) found that
patients tend to over-estimate compliance behavior.
Nevertheless, in this study, even if a patient

over-estimates her/his compliance behavior it would not

10

negate the fact that a relationship may exist between
perceived family functioning and perceived compliance
behavior.

Although a one time measurement of objective
clinical parameters, such as weight and blood sugar, are not
sufficient to measure compliance behavior, weight and blood
sugar will be collected to substantiate the effectiveness of
self-report. For example, weight maintenance or weight loss
gives an estimate of compliance with the diabetic diet and
blood sugar levels give an estimate of diabetic control.
However, it is realized that blood sugar levels are
influenced by several things such as stress, exercise,
infection, in addition to the prescribed diet and
medication. Nevertheless, the presence of blood sugar
levels within normal limits will add veracity to the
self-reported compliance behavior concerning diet and
medications. The parameters used to estimate diabetic

control will be discussed in more detail in Chapter 4.

P . i E .1 E I' .
Smilkstein (1978) defines family functioning as that

behavior which promotes emotional and physical growth and
maturation of all members. Perceived family functioning
indicates the patient's image of reality concerning the way
in which her/his family functions. Family is defined by
Smilkstein as a group consisting of the patient and one or

more persons, children or adults, in which there is a

11

commitment for members to nurture each other. This
definition is encompassing enough to include a wide range of
family types including the traditional nuclear family and
families more broadly defined.
Smilkstein (1978) established parameters by which a
family's functioning could be measured by devising five
basic components of family functioning which he called the
Family APGAR. The components are defined as follows:
Adaptation: Adaptation is the utilization of intra
and extra- familial resources for problem
solving when family equilibrium is
stressed during a crisis.

Partnership: Partnership is the sharing of decision

making and nurturing responsibilities by
family members.

Growth: Growth is the physical and emotional
maturation and self-fulfillment that is
achieved by family members through mutual
support and guidance.

Affection: Affection is the caring or loving
relationship that exists among family
members.

Resolve: Resolve is the commitment to devote time
to other members of the family for
physical and emotional nurturing. It

also usually involves a decision to share
wealth and space.

Smilkstein (1978) developed this family function paradigm
believing each of these components has a unique function
while still being interrelated to overall family
functioning.

Family functioning usually is a subjective

measurement and is based on an individual's perception, that

12

is, her/his image of reality (King, 1971). Studies by
Haggerty (1965) and Geismar (1962) have shown that internal
functions within a family are not easily observable by
outsiders and it is difficult to obtain agreement between
observers. A more recent study, by Good and her colleagues
(1979), seems to substantiate this finding. That is, Good
and her colleagues (1979) found only a moderate agreement (r
= .64) between the patient's perception of her/his family
functioning as indicated by the Family APGAR index score and
how the therapists perceived the individual's family
functioning. Therefore, since the way in which the
individual perceives an event or situation is her/his image
of reality and is usually the basis for her/his behavior,
this study is concerned with how the diabetic patient
perceives her/his family functioning rather than how the

individual's family functioning is perceived by others.

D . I' D I

Research has shown that demographic variables are
not predictive of compliance behavior (Haynes et al., 1979;
Marston et al., 1970). Therefore, the demographic variables
in this study will be used only for descriptive purposes.
The demographic data that will be utilized for this study
are: age, sex, race, income, and marital status. Since
this study is concerned with family functioning, the size
and composition of the household of the participants will

also be determined, i.e., the number, age, sex, and

13

relationship of people with whom the subject lives.

C E 3' fl . l]
The confounding variables that will be investigated
in this study are: (l) the number of years the patient has
been a diabetic: (2) the number of chronic illnesses the
patient has in addition to diabetes; (3) the complexity of
the patient's regimen; and (4) whether the patient has been
hospitalized for diabetes within the last six months.
1 According to Haynes and his colleagues' (1979)
review of the literature, there is no association between
the number of years the patient has had a disease and
compliance. By contrast, there is an inverse relationship
between duration of the regimen and compliance behavior.
That is, the longer an individual is on a regimen, the less
likely he/she is to comply. Since the diet is basic to all
diabetic regimens it is assumed that patients would have
been on a diabetic diet for the length of their illness.
This is so because the first treatment prescribed to control
diabetes is diet. Therefore, data concerning the number of
years with diabetes will be used as an indirect measure of
duration of treatment.

Reasearch (Haynes et al., 1979) has shown that the
complexity of the regimen influences compliance behavior,
i.e., more complex regimens are inversely related to
compliance behavior. The presence of chronic illnesses

other than diabetes would serve to make the therapeutic

14

l
295

:‘3

M.

:4.

:J

.A“

.‘n

un-

D‘-
5.4.

Q
I:
5v.

5‘.
in

“h

I),

0’)

regimen more complex. For example, patients who have a
heart condition may have salt and/or cholesterol restriction
in addition to their diabetic restrictions. Furthermore,
patients with other chronic illnesses may need medications
in addition to those required for their diabetes. Moreover,
these different medications may need to be taken at various
times throughout the day. Therefore, the number of other
chronic illnesses as well as the complexity of regimen will
be included as variables that may confound the relationship
between perceived family functioning and perceived
compliance behavior.

Still another confounding variable to be included is
the evidence of hospitalization within the last six months
for diabetes. Hospitalization for diabetes may result in
reinforcement teaching for diabetes. Research has shown
that the patient's knowledge of the disease or regimen is
not consistently associated with compliance behavior (Haynes
et al., 1979). Yet, it seems reasonable to assume that the
recent reinforcement teaching may serve temporarily to
improve perceived compliance behavior.

In summary, the purpose of this study is to
ascertain the relationship between perceived family
functioning and perceived compliance behavior. However,
since research has shown that other variables appear to
influence compliance behavior as well, data will be

collected concerning the following confounding variables:

15

(l) the number of years the patient has had diabetes; (2)
the number of chronic illnesses other than diabetes the
patient may have: (3) the complexity of the patient's
regimen: and (4) whether or not the patient has been
hospitatlized within the last six months for her/his
diabetes. An analysis will be made to determine if these
confounding variables are correlated with perceived
compliance behavior. In addition, the data for the
confounding variables will be used for descriptive purposes

concerning the subject's diabetes.

Limitations
The limitations of the study are;

1. Patients with adult onset diabetes who have various
other chronic illnesses and a more complex regimen may
respond to questions differently than those who only
have adult onset diabetes.

2. This study does not control for other variables also
believed to influence compliance such as the
doctor-patient relationship, and health beliefs, e.g.,
perceived susceptibility and the perceived severity of
the disease, or the costs and benefits of following the
therapeutic regimen.

3. The subjects in this study have not been controlled for
severity of diabetes.

4. The varied lifestyles of the patients involved may

influence how they perceive compliance and family

16

10.

11.

functioning.

The patient's perceived view of family functioning may
be variable and there may be a slight change from day
to day. A crisis may temporarily alter the patient's
perceived family functioning to a great degree.

The subjects who agree to participtate in the study may
be different from those who refuse.

The sample size is small, a convenience sample from
limited geographic areas and limited settings, and
therefore not generalizable to the larger population.
A rare case of juvenile onset diabetes has its onset
after the age of 30 and may inadvertently be included
in this study.

The age range for adult onset diabetes extends across
several developmental stages of the life cycle. A
consideration of way in which developmental stages
influence response to the family functioning section of
the questionnaire is not included in this study.

Both family functioning and compliance behavior are
measured subjectively and an individual who is positive
about one variable (e.g., family functioning) may tend
to be positive about the other (e.g., compliance
behavior).

The instrument was developed by the researcher and has

not been tested for validity or reliability.

17

W

In this study the researcher is making the following

assumptions:

1.

It is assumed that perceived family functioning can be
measured.

It is assumed that measurement of "perceived"
compliance behavior is an accurate and reliable method
of measuring compliance in the diabetic patient.

It is assumed that patients' perceived family
functioning and perceived compliance behavior can be
adequately measured by means of the questionnaire.

It is assumed that the family functioning instrument
and the compliance instrument will be sensitive enough
to indicate level of perceived family functioning and
level of perceived compliance behavior.

It is assumed that the patient's answers to the
questions on the instrument reflect her/his true
perceived compliance behavior and perceived family
functioning.-

It is assumed that the age group selected is typical of
adult onset diabetics.

It is assumed that the family functioning section of
the questionnaire will be applicable to. all the
developmental stages of the life cycle that adult onset

diabetes crosses.

18

(n
r V
I).

‘. ‘.
.16..

All

M?

‘a‘\

f“...
“on.

"I
(I)

A!
A
N.“

W

The study is presented in six chapters. In Chapter
I the introduction to the problem, purpose of the study,
statement of the problem, definition of concepts, statement
of the hypothesis, and limitations and assumptions of the
study are presented. Chapter II contains the conceptual
framework and focuses on family functioning and compliance
behavior in chronic illness as well as how each of these
concepts relates to nursing. Literature that focuses on
compliance, chronic illness, diabetes, family functioning,
and the family's influence on compliance is reviewed in
Chapter III. In Chapter IV the method used to collect data,
the research design, operational definitions, population,
sample, and source of subjects are described. In Chapter V
the data are presented and discussed. The summary,
conclusion, implications for nursing, and the
recommendations for additional research are presented in

Chapter VI.

19

CHAPTER II

CONCEPTUAL FRAMEWORK

Intmduction

Health is a dynamic state in the life cycle .of an
individual. According to King (1971), this dynamism implies
continuous adaptation to stresses in the internal and
external environment through optimum use of one's resources
to achieve maximum potential for daily living. Health may
be viewed as a continuum with wellness and illness as the
end points. At any point on this continuum the individual
has positive attributes of wellness simultaneously with
negative attributes of illness. The balance of these well
and ill attributes determines the individual's state of
health. The highest level of wellness is the maximum level
of functioning of a person at any specified time and place.
The ultimate aim of nursing is to enable individuals to
reach their highest level of wellness.

Illness is the inability of the individual to meet
her/his needs in a way that will allow her/him to achieve
maximum potential for daily living. Illness is a stressor
to the individual's internal and external environment which
requires adaptation. When the illness is acute, only a
temporary adjustment need be made. For example, there may

be a need to alter the diet, to take medication, and/or to

20

limit activities for a brief period of time. In addition,
there may be a temporary strain on financial resources. By
contrast, when the disease is chronic, it may be necessary
to make permanent adjustments. That is, there may be a
permanent need to restrict activities, to change diet, to
take medications for an indefinite period, or to adjust to
an indefinite strain on financial resources. These
adjustments are permanent because in most cases the
chronically ill individual is not expected to recover.
Moreover, to avoid the complications of the disease and to
lead as normal a life as possible, the individual is
expected to comply with the therapeutic regimen throughout
her/his lifetime to maintain her/his maximum potential for
daily living.

The effects of any chronic illness are not limited
to the patient alone, however. Most individuals live within
a family unit. According to Rakel (1977), illness in one
family member affects all other family members and recovery
from illness is likewise hampered or facilitated by the
relationships within the family. In other words, in chronic
illness, the ability of a patient to reach her/his maximum
potential for daily living may be hampered or facilitated by
the interaction of family members, i.e., the functioning of
the family.

In short, in chronic illness, the interaction of the

thembers within the family, i.e., family functioning, appears

21

to contribute to a patient reaching her/his maximum
potential for daily living. Moreover, in chronic illness,
compliance behavior appears to contribute to a patient
reaching her/his maximum potential for daily living.
Therefore, it is reasonable to expect that there may be a
relationship between family functioning and compliance
behavior in chronic illness. Furthermore, since an
individual's image of reality is reflected in her/his
perception (King, 1971), it also is reasonable to expect
that there may be a relationship between perceived family
functioning and perceived compliance behavior. Finally,
since the ultimate aim of nursing is to help the patient
reach her/his maximum potential for daily living, insights
derived from research into the relationship between
perceived family functioning and perceived compliance
behavior in chronic illness can be of value to nursing
practice and education.

Accordingly, the conceptual framework for this study
is based on these two concepts, i.e., perceived family
functioning and perceived compliance behavior in chronic
illness. These concepts are discussed in more detail in
subsequent sections as is their relationship to nursing and
the nursing theory, based on King (1971), which is used in

this study.

22

C 1' E l . i E 'II :1 . 1]]

Compliance is defined as the extent to which a
patient's behavior coincides with medical or health advice
(Haynes et al., 1979). Studies have shown that compliance
with all therapeutic regimens is disturbingly inadequate and
that approximately one-third of all patients do not adhere
to therapeutic regimens (Blackwell, 1973: Marston, 1970:
Davis, 1966). Furthermore, research has shown that
compliance behavior is reduced even more when the regimen is
long in duration, complex, and requires extensive behavioral
changes (Haynes et al., 1979).

Diabetes is an example of a chronic illness in which
compliance behavior presents difficult and challenging
problems for patients. This is so because control of
diabetes involves extensive behavioral change and a complex
regimen throughout the patients lifetime. For example, the
therapeutic regimen for diabetes involves dietary
alteration, taking daily medications, urine testing,
alteration in exercise, special attention to hygiene and
foot care, and frequent follow-up appointments. Moreover,
the regimen for diabetes is complex. For example, the
patient must not only understand and follow the prescribed
diet but he/she also must eat the food in specified amounts
at specified times. Furthermore, the patient must
coordinate the amount of activity he/she engages in with the

food eaten and medication taken. Finally, since many

23

diabetics frequently have other chronic illnesses such as
cardiovascular diseases (Guthrie & Guthrie, 1977), they also
must take medications for these diseases, further increasing
the complexity of their regimen.

Compliance behavior, however, presents difficult and
challenging problems for the health care provider as well.
For example, since most diabetic patients are treated in an
ambulatory care setting, the patient is expected to
participate in diabetic control through compliance with a
therapeutic regimen. But, the health care provider has
little ability to enforce compliance behavior, e.g., he/she
cannot ensure that a patient comply with the diabetic diet
as directed, as would be the case in an institutional
setting. In addition, there is no ideal way for the health
care provider to detect whether or not the patient is
complying. Therefore, it is difficult to know if a lack of
improvement is due to noncompliance, an inadequate treatment
plan or other factors. For example, variables such as
infection, stress, and exercise, can influence blood sugar
levels and affect diabetic control (Guthrie & Guthrie,
1977). Moreover, clinical outcomes, such as blood sugar and
urine glucose levels give only a fragmentary glimpse of
diabetic control (Bondy & Felig, 1971). As a result, a
laboratory result from one point in time may be reflective
of these other factors rather than lack of compliance

behavior. At the same time, reliance oh self-reports of

24

[11

In

compliance frequently used by physicians, also may be an
inadequate measure of patient behavior. This is so because
patients may overestimate their compliance behavior.

Evaluating compliance, therefore, is a problem for
the provider. Evaluation of compliance behavior, however,
also may be a problem for the researcher because it is
difficult to find a reliable means to measure compliance
behavior from a one time measure. Yet, studies of
hypertensive patient have shown a relationship between
self-reported compliance with taking antihypertensive
medication and blood pressure control (Green et al., 1979:
Hershey et al., 1980). Therefore, even if study subjects do
overestimate their compliance behavior, there may be a
relationship between perceived compliance behavior and
actual compliance behavior in the adult onset diabetic
patient.

Perhaps because compliance behavior is such a
problem, much research has been conducted to explore its

sources. Several of the potential variables that have been

investigated extensively are: (l) demographic
characteristics: (2) doctor-patient relationships: (3)
patient characteristics such as health beliefs and

knowledge: (4) disease features: and (5) features of the
regimen.
Demographic variables (e.g., age, sex, education,

socioeconomic status, income, occupation, marital status,

25

(ECG:

392:;

are.

m:

'5

Q

COLI

9C .

M

«hive
F
”w
b

a ‘I

at.“

race, and religion) have been studied most extensively
because they are easy to identify and measure. Although
demographic variables are easy to identify and measure they
are relatively enduring. That is, even if a consistent
relationship could be found between one or more of these
variables and compliance behavior, in most instances little
could be done to alter the situation. However, research has
found that demographic characteristics are not predictive of
compliance when examined apart from other variables (Haynes
et al., 1979: Marston, 1970). Demographic variables,
however, are valuable for descriptive purposes to depict the
population represented in the study.

The doctor-patient relationship also has been
studied and it has been found that indeed the patient's
degree of satisfaction with the health care provider and the
clinic has shown to be correlated with compliance (Becker,
1979). Nevertheless, since the investigation of
doctor-patient relationships is a study in itself, this
variable will not be included in this study in an effort to
narrow its scope. The exclusion of doctor-patient
relationships as a variable will be considered a limitation
to this study.

Still another variable that has been studied
extensively as a potential predictor of compliance is the
health beliefs of patients. Health beliefs include the

patients' (1) perceived susceptibility to the disease and

26

its complications: (2) perceived severity of the disease:
and (3) perceived barriers to following the treatment
regimen such as physical, psychological, and/or financial
costs related to the therapeutic regimen. Studies of
mothers of children with acute illnesses have produced data
that show a positive association between compliance behavior
and health beliefs (Becker et al., 1972; Charney et al.,
1967; Francis et al., 1969). However, conflicting data has
been produced regarding the health beliefs of adults with
chronic illness (Green et al., 1979: Taylor, 1979:
Cerkoney & Hart, 1980). Thus, even though health beliefs
may influence compliance behavior it is apparent that other
factors influence compliance as well. Accordingly, in an
effort to limit the scope of this study, health beliefs will
not be included as a variable. The lack of control for
health beliefs will be considered a limitation to this
study.

Knowledge also frequently has been examined 'to
determine if it influences compliance behavior. The
research conducted, however, has produced conflicting data.
That is, some studies have shown no association between
knowledge and compliance (Becker et al., 1974: Bille, 1977:
Blackburn, 1977; Donabedian & Rosenfeld, 1964; Vincent,
1971) and others have shown a positive relationship between
knowledge and compliance (Heinzelmann, 1962: Tagliacozzo &

Ima, 1979; Watkins et al., 1967). Although knowledge of

27

diabetes or the regimen will not be investigated in this
study, hospitalization for diabetes may be a confounding
variable in this study. This is so because when a patient
is hospitalized due to uncontrolled diabetes it may be
secondary to noncompliance or it may be due to infection,
illness, or severe stress. Nevertheless, regardless of the
cause, patients who are hospitalized for uncontrolled
diabetes usually receive reinforcement teaching. And, this
reinforcement teaching may serve to improve their compliance
temporarily. Therefore, hospitalization for diabetes may be
a confounding variable in this study.

Yet another variable that has been thought to be a
potential predictor of compliance behavior and studied is
features of the disease such as diagnosis, severity of the
disease, and degree of disability. For example, Hulka and
her colleagues (1975), in a study of patients with
congestive heart failure and adult onset diabetes, found no
relationship between disease features and compliance
behavior. Similarly, Diamond (1968), in a study of
orthopedic and neurological problems, found no relationship
between diagnosis or severity and compliance. In contrast,
Donabedian and Rosenfeld (1964) in a study of chronically
ill patients showed that the more severe the disability the
more compliant the patients were. Thus, the relationship of
features of disease to compliance behavior remains in

question.

28

 

rel:
assc
lite

feat

a
(I,
. a-

con:

 

 

 

Finally, features of the treatment regimen
frequently also has been examined to determine if they are
related to patient compliance behavior. And, Haynes and
associates (1979), following an extensive review of the
literature, indicated that a relationship does exist between
features of the regimen and compliance behavior. That is,
compliance with a therapeutic regimen is reduced when the
treatment program involves extensive behavioral changes, is
more complex, and is continued over time (Haynes et al.,
1979). It might be expected, then, that the features of the
treatment regimen for diabetic patients would be related to
their compliance behavior. Thus, since diabetes is the
chronic illness being investigated in this study, features
of the regimen such as duration and complexity may be
confounding variables in the study.

To summarize, approximately one-third of all
patients on therapeutic regimens do not comply.
Furthermore, compliance behavior is more difficult when the
regimen is long in duration, complex, and requires extensive
behavioral changes. The above features of a therapeutic
regimen are characteristic of most chronic illnesses,
including adult onset diabetes. Therefore, it is reasonable
to expect the compliance behavior in adult onset diabetes
maYbe a difficult and challenging problem.

Compliance behavior of the adult onset diabetic is

Challenging in that compliance affects control of diabetes,

29

fi.nt
‘n “5

k . I
wen:

5“":

compliance affects control of diabetes, ability of the
patient to reach her/his maximum potential for daily living.
Compliance for the patient is difficult because in diabetes
the therapeutic regimen is complex, requires extensive
behavioral changes, and is extended throughout the patient's
lifetime. Since it is difficult to detect noncompliance,
compliance behavior of the patient presents a problem for
the health care provider as well. That is, it is difficult
to assess the efficacy of the treatment plan since he/she
cannot know whether a lack of improvement is due to an
inadequate treatment plan or a lack of compliance. However,
it has been reported that simply asking the patient in a
nonthreatening manner is as effective approach as more
objective means to detect noncompliance (Haynes et al.,
1980).

Because compliance presents problems for both the
patient and health care provider, many variables have been
investigated to ascertain their influence on compliance
behavior. Some of these variables have been shown to
influence adherence to a regimen and will be investigated in
this study since they may act as confounding variables. For
example, since duration and complexity of the regimen
influence compliance behavior and since adult onset
diabetics tend to develop other chronic illnesses, further
increasing the complexity of the therapeutic regimen, it is

reasonable to expect that the number of chronic illnesses

30

0th

'4
m
u.)

kno‘
inf.
c‘iai

dial

be¢

doc1

9V: "

other then diabetes and the complexity and duration of the
regimen may be confounding to this study. Similarly, since
knowledge of the disease and therapeutic regimen may
influence compliance behavior, and since hospitalization for
diabetes may lead to reinforcement teaching concerning
diabetes and temporarily influence compliance behavior,
hospitalization for diabetes in the last six months may also
be confounding to this study. Health beliefs and
doctor-patient relationship also appear to influence
compliance behavior. However, in an effort to limit the
scope of this study, these variables will not 'be
investigated, but instead, the lack of inclusion will be
considered a limitation of this study.

Demographicvariables also will be included in this
study even though it has been suggested that they do not
appear to influence compliance behavior. when viewed
separately. Nevertheless, in this study demographic data
will be used only for descriptive purposes since they are
useful to depict the sample studied.

Another variable that appears to inluence compliance
behavior of patient with chronic illness is the provision of
support from family members (Caplan et al., 1976: Haynes et
al., 1979). From research concerning the family's influence
on compliance behavior (Heinzelmann & Bagley, 1970: Litman,
1966: Oakes et al., 1970: West et al., 1977) there is

evidence that the way in which a family functions affects

31

mi
one
def
ml

are

the patient's compliance behavior. In the next section
family functioning and the family's influence on compliance

behavior will be considered.

F . F c . . . . ,
W

Most individuals live within a unit. This family
unit may be a traditional or extended family, or it may be
one of several other modern forms of the family more loosely
defined. That is, to be considered a family, the unit need
only include one or more persons, children or adults, who
are commited to nurture each other (Smilkstein, 1978).
Within such a family, relationships exist among members and
each individual in the family affects all other individuals
within the family (Rakel, 1977). Thus, when an individual
has a chronic illness, the effects of the illness may not be
limited to the patient alone. That is, illness in one
family member may affect all other family members.
Similarly, when a person has a chronic illness, her/his
adjustment to the disease and its treatment also may be
hampered or faciliated by the relationships within the
family (Rakel, 1977). That is, positive family
relationships may lead to positive self image which may lead
to a desire to comply with the therapeutic regimen to attain
or maintain health.

The relationships or the interaction of the family

32

I B
it.»

CCHC

_,—’-\

(I)

an

. N

\..~O.
1

members may be referred to as family functioning. The
concept, although frequently referred to, is variably
defined. For example, Rakel (1977), equates family
relationships or interaction of family members with family
functioning. That is, according to Rakel (1977), families
in which members exhibit love, discipline, tolerance,
adaptability, and ease of communidation function at a more
stable and flexible level than do families that lack one or
more of these features. Pless and Satterwhite (1973), in
contrast, define family functioning as the dynamics of
everyday life: i.e., the way in which the family as a unit
operates across dimensions such as intra-family
communication, cohesiveness, decision—making, marital
satisfaction, and happiness and closeness of the family
unit. Geismar and associates (1962) do not give a specfic
definition for family functioning but do list the categories
of family functioning. These categories are: (1) family
relationships and family unit, (2) individual behavior and
adjustment, (3) care and training of children, (4) social
activities, (5) economic practices, (6) household practices,
(7) health conditions and practices, (8) relationship to
social workers, and (9) use of community resources.
Smilkstein (1978) defines family functioning as that
behavior which promotes emotional and physical growth and ‘
maturation of all members. For Smilkstein, the concept

encompasses five basic components: adaptation, partnership,

33

Srci

0th

the

h
0

his

growth, affection, and resolve. According to Smilkstein,
these components represent the common themes in the social
science literature dealing with families. In the paragraphs
that follow, Smilkstein's definition of the components of
family functioning are defined and compared with the other
authors' concepts of family functioning.

Adaptation is defined by Smilkstein (1978) as family
members aiding each other in times of need and accepting
help or assistance from friends and community agencies for
problem-solving. To assess adaptation within a family one
might for example, ask the following questions. Do the
family members pull together and give each other assistance
when a crisis develops within a family? Is the family
willing to turn to community resources for assistance when
the resources within the family are not sufficient to keep
the family functioning?

Given these examples, it can be seen that
Smilkstein's definition of adaptation is consistent with
other definitions. For example, Rakel (1977) refers to
adaptation as adaptability. He states that adaptability
makes it possible to adjust and respond to sudden or
relatively unexpected changes. Further, Pless and
Satterwhite (1973) refer to adaptation when they consider
the ability of individual member's to turn to other family
members to help solve problems. And, although Geismar and

his associates (1962) do not discuss adaptation

34

U56

91“
Ill

tim

specifically, they do imply the concept when they state that
use of community resources is important to family
functioning. Each of these resources can be useful in the
time of crisis as well as for everyday living. Thus, it
appears that each of the author's views are basically
consistent with Smilkstein's view of adaptation.
Partnership, according to Smilkstein (1978), is
defined as the way in which family members communicate with
each other and make decisions about such matters as
vacation, finances, medical care, large purchases, and
personal problems. Rakel (1977), while not using the term
partnership, refers to free communication and states it
contributes to the consistency of open and honest
relationships among members, i.e., open and honest
relationships contribute to partnership. Similarly, Pless
and Satterwhite (1973) do not identify partnership as such,
but they too consider family communication and
decision-making as vital components of family functioning.
Finally, Geismar and associates (1962) consider family
relationships and family unity, including marital
relationships, parent-child relationship, relationships
among children, and family solidarity, to require family
communication and decision-making. Thus although Rakel
(1977), Pless and Satterwhite (1973), and Geismar and
associates (1962) do not identify partnership as such, each

has identified areas of family functioning that are

35

consistent with Smilkstein's partnership component.

Growth, the third component of family functioning,
is defined by Smilkstein (1978) as the way in which family
members accept change, promote independence, and share
nurturing. Nurturing often refers to the training or
rearing of children. But, the concept also can be used to
refer to freedom available to the individual within the
family to make a change as well as the independence he/she
is allowed to change. In this sense, chnage is equilvalent
to growth. Rakel (1977) refers to this concept as tolerance
which he says allows for individual freedom and development.
In addition, Rakel (1977) alludes to the concept when he
considers discipline which provides guidelines for
acceptable activity and prepares the individual for future
needs. Geismar and colleagues (1962) also consider growth
when they evaluate the family's ability to care for and
train children as well as the social activities in which the
members are involved. That is, social activities promote
growth as well as independence. Each author, then, appears
to consider promotion of growth as an important component of
family functioning, although each refers to it differently.

Affection, according to Smilkstein (1978), is
defined as the way in which members of the family respond to
emotional expressions such as affection, love, sorrow, or
anger. In considering affection, the question becomes:

what is the family member's satisfaction with the intimacy

36

*4.

5»

r1

(I)

Ph

PI‘

In

and emotional interaction that exists within a family?
Rakel (1977) considers affection when he refers to love. He
states that love provides an atmosphere of warmth,
acceptance, and support. So too do Pless and Satterwhite
(1973) consider affection when they address marital
satisfaction, happiness, and closeness. In contrast,
although Geismar and associates (1962) do not single out the
component affection, they do consider it when they view
family relationships and family unity as well as when they
consider the emotional care of children. Each author, then,
in his or her own way, indicates the importance of affection
as a component of family functioning.

Finally, resolve, as defined by Smilkstein (1978),
refers to the way in which members of the family share time,
space, and money. Degree of resolve is defined by the
individual's satisfaction with the amount of time that
family members are willing to share and her/his satisfaction
with the amount of money and space allowed for personal use.
Rakel (1977) does not refer to anything that may be equated
with Smilkstein's resolve. Pless and Satterwhite (1973),
however, consider resolve when they refer to the family's
cohesiveness, that is, the time the family spends together
such as on vacations or on weekends. Geismar and associates
(1962) also do not specifically speak to resolve, but they
do consider the family's economic practices and their

Inausehold practices, practices equivalent to Smilkstein's

37

d1"
‘4

re;

C02

1282

CO'

Ra

he

in

me

[8

V6

cC

Th

reference of sharing of money and space. Thus, other
authors give variable emphasis to resolve as defined by
Smilkstein.

In summary, it can be seen that various authors
define the components of family functioning somewhat
differently. Nevertheless, there appear to be common themes
represented in all their writings. Thus, since Smilkstein's
components of family functioning encompass the concepts
mentioned by the other authors and since his definition of
each component is clear, .his definitions of the five
components will be used as a basis for this study.

Smilkstein (1978), Pless and Satterwhite (1973), and
Rakel (1977) indicate that family functioning influences the
health of patients. Other authors have considered the
influence of the family on the health of the individual
members as well. For example, Schmidt (1978) conducted a
review of the family literature and found affirmation of the
importance of the family to the individual's health. .In
addition, Litman (1974) states that the family constitutes
perhaps the most important social context within which
illness occurs and is resolved and, consequently, serves as
a primary unit in health and medical care. In a similar
vein, Freeman (1970) writes that the home environment is
conducive to health maintenance and personal development.
That is, the extent to which the family functions as a unit

affects its capability in health matters. Geyman (1977)

38

furthermore states that disorders in the family unit can
precipitate illness in individual members. Thus, there 18'
agreement that family functioning affects the health of its
members.

It appears, then, that family functioning may affect
the health of its members and we also know that health is
related to compliance behavior. Therefore, it might be
expected that the family would influence compliance
behavior. Indeed, some studies have been conducted to
ascertain the influence of the family on adherence to a
regimen and attention now turns to them and their
relationship to Smilkstein's (1978) components of family
functioning.

Litman (1966), in a study of patients who suddenly
found themselves with a severe orthopedic disability, found
good response to physical rehabilitation among those
patients who received positive reinforcement from their
families and poor response to physical rehabilitation among
those patients who did not obtain this encouragement. That
is, patients whose families were able to adapt to the crisis
and give support were more likely to comply with the
required regimen. Such ability to adapt in time of crisis
is defined by Smilkstein (1978) as the adaptation component
of'his Family APGAR.

In a similar vein, a study conducted by Cakes and

associates (1970) found that perceived family expectations

39

2.11

In

were strongly related to the arthritic patient's compliance
with the use of a hand resting splint. In addition,
Heinzelmann and Bagley (1970), in a study of a group of
sedentary middle-aged men who were at high risk for
developing heart disease, found that those men .whose wives
had positive attitudes toward a physical conditioning
program were more apt to comply with their physicial
conditioning program. The perceived family expectations
considered in Cakes and associates' study (1970) and the
positive attitude of the wife discussed in Heinzelmann and
Bagley's study (1970) suggest that communication, verbally
or nonverbally, took place between the family members or
partners and encouraged compliance. Smilkstein (1978)
considers communication between family members an aspect of
partnership and partnership is a component of family
functioning. In this sense then, Heinzelmann and Bagley's
(1970) and Cakes and associates' (1970), findings suggest
that partnership is related to compliance behavior.
Furthermore, a study by West (1977) suggests that receiving
support from other family members can result in an increased
probability of compliance behavior. In a survey of smokers
who had attended a smoking clinic five years earlier, West
(1977) found that two-thirds of the quitters had spouses who
made it easier to quit while only slightly more than one-
third of the nonquitters received this support. Quitting

smoking, of course, requires a change in life habits.

40

According to Smilkstein (1978), family members who encourage
and/or accept change and give mutual support and guidance
promote growth of the family members. Families in the West
(1977) study who gave support and accepted change
(Smilkstein's growth component of family functioning)
resulted in a better probability of compliance behavior.

Moreover, a review of research literature by
Baekeland and Lundwell (1973) found that in 19 out of 19
studies, social isolation and/or lack of affiliation with
another person was a major cause for the discontinuation of
treatment among psyciatric patients.‘ Such patients, then,
lacked the emotional and functional support potentially
available to those who are in close contact with significant
others. In this sense, the findings of Baekeland and
Lundwell (1973) suggest that Smilkstein's concept of resolve
-- i.e., a commitment to devote time, wealth, and space with
other members of the family for physical and emotional
nurturing -- is related to compliance behavior, or, _at
least, continuation with treatment.

Finally, each of the studies cited suggests that
emotional support and caring by other family members is
conducive to compliance behavior. That is, by being present
in times of need and sharing emotional experiences, family
members convey support, thereby increase the likelihood that
a patient will comply with treatment regimens. Such

emotional support and caring are referred to by Smilkstein

41

(1978) as the affection component of family functioning.

It appears then, that each of the individual
components of family functioning, as defined by Smilkstein
(1978), do have an influence on compliance behavior.
However, the relationship of compliance behavior to an
individual's perception of the total concept of family
functioning has not been studied to the best of the authors
knowledge.

In summary, the way in which family functioning
influences the health state is a question that has been
explored. But, the total concept of family functioning as
it relates to compliance behavior has not been studied. The
purpose of this study is to investigate the relationship
between perceived family functioning and perceived
compliance behavior of the adult onset diabetic, a

heretofore neglected subject.

Summazx

A review of the literature showed that the
chronically ill patient often has trouble complying with
prescribed treatment. Furthermore, in chronic illness, such
as adult onset diabetes, compliance behavior is necessary
for an individual to reach her/his maximum potential for
daily living. Moreover, research showed that certain
aspects of family functioning appear to influence compliance
behavior which affects the health state of patients with

chronic illnesses such as hypertension, arthritis, and

42

orthope
expect
concep

patien

indie:
perce;
of
seas:
outcc
teas:
the:
can;
Fur:
ind;

the

orthopedic disabilities. Therefore, it is reasonable to
expect that there may be a relationship between the total
concept of family functioning and compliance behavior in
patients with the chronic illness, diabetes.

To continue .then, perception is defined as an
individual's image of reality and since an individual's
perception may influence her/his behavior, for the purpose
of this study, perceived family functioning will be
measured. Since objective measures such as clinical
outcomes have limitations in studies based on a one time
measurement, perception of compliance will be measured since
there appears to be a relationship between perceived
compliance behavior and actual compliance behavior.
Furthermore, since nursing is concerned with helping
individuals reach their maximum potential for daily living,
the relationship of these variables to nursing theory will

be discussed.

Nursins_ThsnL¥

Man and his environment is the central focus for the
framework of nursing (King, 1971). The aim of nursing is to
give assistance to individuals to help them cope with a
health problem or adjust to interference in their health
state (King, 1971). In other words, the aim of nursing is
to help individuals reach their maximum potential for daily
living. King's (1971) theory is based on four universal

concepts. These concepts are perceptions, social systems,

43

id

he

be

interpersonal relationships, and health. These four
universal concepts from King's theory were used as a frame
of reference for this study which has as its focus perceived
family functioning and its relationship to perceived
compliance behavior in chronic illness.

Chronic illness, such as diabetes, causes an
interference in an individual's health state. Compliance
with the therapeutic regimen for diabetes, prescribed by the
health care provider, will aid the individual to reach
and/or maintain her/his maximum potential for daily living.
However, as stated previously, compliance with a therapeutic
regimen, especially in chronic illness, is not easy. The
nurse realizes there are many barriers to compliance. One
facet of the nurse's role is to help the individual to
identify and overcome barriers, and enable her/him to use
her/his potential ability to function.

To assist an individual to identify and overcome
barriers to compliance, the nurse utilizes the nursing
process. The nursing process involves assessing the
problem, planning an intervention, implementing the
intervention, and evaluating the outcome.

In chronic illness, there is a need for the nurse to
assess if the patient may be having a problem with
compliance behavior. A cue that a problem may exist could
be obtained by reviewing the objective parameters from the

patient's records such as blood glucose, degree of

44

glycc
to m
meas

are

glycosuria, alpha hemoglobin, or obesity. It may be a cue
to noncompliance with the diabetic regimen if repeated
measurements of one or more of these objective parameters
are not within normal limits.

It is not only necessary to assess the objective
parameters in relation to compliance behavior, but an
integral part of an assessment is an exploration of the
patient's perceptions since man functions in terms of his
perceptions (King, 1971). According to King (1971),
perception is each individual's image of reality or her/his
conscious awareness. Therefore, the nurse should also
assess the patient's \perception of her/his compliance
behavior. The patient's perception of her/his compliance
behavior is important since if the patient does not perceive
herself/himself as non-compliant, the patient will not
realize a need to plan strategies to improve compliance.
Thus, perception is a necessary concept for the nurse to
deal with in the delivery of health care. .

In addition, the patient's perception of his/her
social system also influences her/his behavior (King, 1971).
King (1971) states ”the social system is composed of groups
of individuals joined together in a network or system of
social relationships to achieve common goals developed about
a system of values with an organized set of practices and
the methods to regulate practices and administer the rules".

The family unit is a social system in that it is a group of

45

:0.

individuals joined together in a network. King (1971)
states that man's adaptation to life and health is
influenced by her/his perception of environmental factors.
Thus, since the family unit is a part of the environment of
the patient, it is reasonable that the individual's
perception of her/his family would influence adaptation to
life and health.

Further, according to King (1971), man functions in
a social system (family unit) through interpersonal
relationships, i.e., interaction of individuals for some
purpose or goal. Rakel (1977) states that the interaction
of family members determines how the family functions.
Therefore, the interpersonal relationships between and among
family members determines how the family functions. Thus,
since it seems the family influences adaptation to life and
health, it is reasonable to expect that the interpersonal
relationship among family members, i.e., family functioning,
also would influence adaptation to life and health.

To continue then, man's perception of his
interpersonal relationships (family functioning) within the
social system (family unit) influences the individual's life
and health. The family unit may be a source of strength to
an individual or it may have a negative influence on the
patient's health. Health is the dynamic state in the life
cycle of an organism which implies continuous adaptation to

stresses in the internal and external environment through

46

to
of

HE

'(J

n:

optimum use of one's resources to achieve maximum potential
for daily living (King, 1971). An unsatisfactory perception
of family functioning by the adult onset diabetic may be a
negative stressor in the individual's life. The individual
may have difficulty adapting to this negative stressor and
in turn the unsatisfactory perceived family functioning may
be affecting the patient's compliance behavior. Compliance
behavior in the adult onset diabetic is essential to
attaining or maintaining health.

The nurse should further recognize that human
behavior has multiple determinants and many factors may
influence the compliance behavior of adult onset diabetic
patients. The limitations of research usually confine a
study to examine only one or two determinants of behavior at
a time. Nevertheless, the nurse must view the patient
holistically and evaluate not only the relationship between
perceived family functioning and perceived compliance, but
must consider' the effect of confounding variables on the
perception of compliance behavior as well.

The confounding variables examined in this study
were: (1) complexity of regimen: (2) number of chronic
illnesses other than diabetes: (3) number of years with
diabetes: and (4) hospitalization for diabetes. These are
only a few of the additional variables that may influence
compliance behavior. If the confounding variables examined

in this study do not appear to alter the perception of

47

conpliar
the pat.
patient
the nu:
develox

fnncti

trate
Strati
{anil'
perce
conpl
be be
urse
tOge‘

was

compliance behavior, and a relationship can be shown between
the patient's perception of family functioning and the
patient's perception of her/his compliance behavior, then,
the nursing action may center on assisting the patient in
developing strategies to improve perceived family
functioning.

Following the patient's implementation of planned
strategies, it is necessary for the nurse to evaluate if the
strategies improved the patient's perception of her/his
family functioning and to determine if the improved
perceived family functioning resulted in improved perceived
compliance behavior. The improved compliance behavior may
be based on the perceptions of both the patient and the
nurse. Finally, the nurse and the patient can evaluate
together if the patient's maximum potential for daily living
was reached. The entire process is summarized in Figure 1.

In this chapter the concepts reviewed were
compliance behavior in chronic illness and family
functioning and the family's influence on compliance
behavior. These concepts were then reviewed in relationship
to King's nursing theory. In Chapter III the review of the

literature will be presented.

48

FIGURE 1. An Integration of the Nursing Process as it Relates to Fam-
ily Functioning and Compliance Behavior in Chronic Illness.

 
 
  
   
 
 
  
   
   

Nurse
*Complexity of
Regimen
A
55955 *Other Chronic
Illnesses

  

Chronic
Illness

 
 

Diabetes

      
  
  
   
   
 
 

    
  

  
 

   

  

Patient's Patient's
Perception Perception
of {--—-'------b’ of
Family Compliance
Function Behavior

*Number of Years
with Diabetes

Plan Intervention *Hospitalization
for Diabetes

 

Implement Intervention

Evaluation of Outcome

 

 

Improved
Improved Family Improved Health
Functioning a Compliance ’ State
Maximum Potential
*Confounding Variables for Daily Living

49

CHAPTER III

REVIEW OF THE LITERATURE

Winn

The review of the literature includes a discussion
of research studies and papers that are related to
compliance behavior, compliance behavior in chronic illness,
and family functioning. First, the magnitude of the problem
of compliance, the variables that have been investigated
concerning compliance, and approaches to measure compliance
will be discussed. Then, studies concerning compliance and
chronic illness focusing mainly on diabetes will be
reported. Finally, a review will be presented of the
literature concerning the identification of components of
family functioning, measurement of family functioning, and
the family's influence on compliance as related to the
components of family functioning. The organization of the
literature will be guided by the following order: (1)
synopses of the literature pertinent to the major subjects
will be presented within individual sections: and (2) each
section will be summarized and the strengths and limitations

of the studies reviewed discussed.

W
In this section the magnitude of noncompliance, the

variables that have been investigated, and the approaches

50

that have been used to measure compliance behavior of the

adult onset diabetic will be discussed.

MW

Past research seems to suggest that compliance with
a therapeutic regimen is a problem to both the patient and
the health care provider. Haynes and his colleagues (1979)
reviewed 537 studies concerning compliance behavior and
found that one-third of ambulatory patients with chronic
illness do not follow their prescribed treatments.
Blackwell (1973) reviewed over 50 studies on medication
compliance and found that failure to take medication
occurred in 25 to 50 percent of all outpatients. By
contrast, Davis (1966) states that taking medication is one
of the easier aspects of the regimen to follow as compared
with restrictions on behavior or changes in personal habits.
Davis' (1966) statement suggests that noncompliance with
restrictions on behavior or with changes in personal habits
may even surpass the percentage of 25 to 50 percent
noncompliance reported by Blackwell. Indeed, Davis (1966),
in his review of the literature, revealed that noncompliance
rates may soar as high as 93 percent although in almost half
of the studies he reviewed, the noncompliance rate fell
within the range of 30 to 35 percent.

According to Davis (1966) the wide range in
compliance rates reported in different studies is due to the

various populations considered, the various methods used to

51

collect data, and the variety of medical problems and
regimens investigated. Moreover, Marston (1970) states that
it is usually misleading to compare compliance rates from
different studies due to the wide variations in operational
definitions of compliance, the lack of truly objective
measures of compliance, and the problem of precision of
analysis that occurs where compliance is based on several
different medical recommendations. Nevertheless, regardless
of the shortcomings inherent in comparisons of compliance
rates across studies, it seems that the problem of
compliance is a substantial one and there is still much that
needs to be learned concerning why patients do or do not

follow treatment regimens.

lariablsueximd

The literature shows that many variables have been
studied to ascertain their influence on compliance behavior.
For the purposes of this research selected variables will be
discussed, i.e., demographic data, knowledge, features of
the disease, and features of the regimen. It is realized
that a considerable body of published data exists on the
relationship between compliance and health beliefs. Health
beliefs are described as the perceived susceptibility to
disease and its complications, the perceived severity of the
disease, the perceived benefits of care, and the perceived
barriers such as physical, psychological, or financial costs

related to the therapeutic regimen. These beliefs have been

52

shown to correlate positively with compliance behavior.
However, many of the studies that examined this relationship
have been conducted on mothers of children with acute
illnesses (Becker et al., 1972: Charney et al., 1967:
Francis et al., 1969). More recent studies, concerning the
health beliefs of adults with chronic illness and compliance
behavior, have revealed conflicting data. (Cerkoney and
Hart, 1980: Ferguson and Boles (1979): Green et al., 1979:
Hershey et al., 1980: Taylor, 1979). Studies have also
shown that doctor-patient relationships have an effect on
compliance behavior (Haynes et al., 1979). However, in an
effort to narrow the scope of the study, doctor-patient

relationships and health beliefs will not be investigated.

Damngrauhi£_¥ariahles. Demographic variables studied in

relationship to. compliance have included age, sex,
education, socioeconomic status, income, occupation, marital
status, and race. Haynes and associates' (1979) extensive
review of the literature revealed that the vast majority of
studies found no significant association between demographic
variables and compliance behavior. However, a few studies
did suggest there may be some relationship between specific
demographic variables and compliance. For example, Bille
(1977), in a study of myocardial infarction patients, found
that the older person was more likely to follow the medical

regimen although he offered no explanation of this finding.

53

3—1

By contrast, Blackwell (1973), who has conducted much
research on compliance and drug therapy, found that
medication errors and noncompliance occur most often at
extremes of age, i.e., the very young do not comply possibly
due to the taste of the medication and the very old have
difficulty complying possibly due to forgetfulness or
self-neglect.

Furthermore, Vincent (1971) found correlations
between age, sex, and compliance behavior. In her study of
patients with glaucoma, Vincent showed that women from 45 to
. 64 years of age were more likely to comply with treatment
regimens than men. ' Vincent suggested that the difference
between sex and compliance was possibly due to the fact that
middle age women usually have more time to give attention to
the specifics of the treatment regimen because their
children are grown, while, men still have occupational
obligations that require their attention. No recent studies
have been located, however, that substantiate Vincent's
finding.' Moreover, Vincent's (1971) findings may no longer
be true since more and more women also have occupatibnal
obligations. By contrast, a review of the literature by
Davis (1966) also revealed that in three out of four studies
of tubercular patients, females were more likely to default
than males concerning medications, but females were more
likely to keep clinic appointments. Davis, however, did not

suggest a reason for the difference.

54

Similarly, some research has shown a relationship
between socioeconomic status, including income and
occupation, and compliance. That is, Davis (1966), in a
review of studies assessing socioeconomic status, reported
that most found that lower class individuals were least
likely to comply. Marston (1970) and Haynes and associates
(1979), hbwever, in more recent reviews of the literature,
suggest that socioeconomic status may not be as strongly
related to compliance as previously thought.

In summary, although a few studies suggest there may
be a relationship between specific demographic variables and
compliance, the majority of studies found that when
demographic variables are examined apart from other
variables they are rarely predictive of compliance behavior.
Demographic variables, however, are useful for descriptive

purposes to depict the sample studied.

Knowledge. Conflicting data exists regarding the
relationship between compliance and knowledge. A positive
association was found by Watkins and associates (1967) in a
study based on direct observation and interview of diabetic
patients. That is, those who had higher scores on knowledge
tests concerning various aspects of diabetes had better home
management scores and more accurate compliance with
treatment than those with low knowledge scores. Williams
and associates (1967) also found that the more patients knew

about their disease and its management the better they

55

carried out recommended therapy. However, these researchers
found no significant correlation between performance and
control of diabetes as measured by such parameters as blood
sugar levels, urine glucose levels, and body weight.
Williams and his colleagues (1967) suggested that a possible
explanation for this finding was that patients who were in
poor control may have been hospitalized more and may have
had more instruction on caring for diabetes. Thus, more
frequent instruction may have resulted in the patients'
increased knowledge concerning diabetes and, accordingly,
more accurate compliance with most aspects of treatment.

In a similar vein, Tagliacozzo and associates
(1970), in a study of 159 medically indigent blacks with
chronic illness, found that knowledge of illness and its
complications, as measured by a paper and pencil test, was
related to the likelihood of continuation of attendance at
the medical clinics over a prolonged period of time. Those
with knowledge of illness and its complications were more
likely to continue in care than those with low knowledge who
tended to terminate care prior to their fourth visit.

In contrast to the findings just discussed, studies
conducted by Bille (1977) on myocardial infarction patients,
Blackburn (1977) on chronic hemodialysis patients, and
Vincent (1971) on glaucoma outpatients, all found no
association between knowledge of the disease and regimen and

compliance behavior. Moreover, Donabedian and Rosenfeld

56

(1964) found no association between understanding the
regimen and compliance behavior, as measured by the accuracy
and completeness with which the patient could describe the
recommendation. However, the researchers report this
negative finding may be due to inadequacies of their
assessments of understanding since the finding was at
variance with the reasons actually given for lack of
compliance. That is, reasons for noncompliance frequently
seemed to reflect poor understanding on the part of the
patient.

In summary, data regarding the relationship between
knowledge and compliance is conflicting. This conflict
suggests that transmitting information alone may not be
enough to overcome noncompliance. Knowledge may be
necessary but not sufficient to insure compliance behavior.
That is, an individual can know something and still not

consider its implication for her/his behavior.

Disea§e_£eatures. The features of the disease to 'be

reviewed in this section are diagnosis, severity of disease,
and degree of disability. Of the studies reviewed, two
assessed diagnosis and its relationship to compliance
(Diamond et al., 1968: Hulka et al., 1975). Diamond and
colleagues (1968) investigated patients admitted for
rehabilitation service who had suffered fractures,
amputation, spinal dysfunction, or peripheral neuropathy.

Hulka and associates' (1975) research dealt with adult onset

57

diabetics and congestive heart failure patients. Neither
study showed an association between diagnosis and compliance
behavior.

Severity of disease was investigated in several
studies of patients with general medical complaints (Davis,
.1968), patients admitted with orthopedic or neurological
problems for rehabilitation (Diamond et al., 1968), patients
with hypertension (Podell et al., 1976), and patients with
adult onset diabetes and congestive heart failure (Hulka et
al., 1975). None of the studies found an association
between severity of disease and compliance behavior.

Two studies reviewed considered the degree of
disability as a function of compliance. Hulka and
associates (1975) in their investigation of adult onset
diabetes and congestive heart failure patients found no
association between degree of disability and compliance
behavior. By contrast, research conducted on chronically
ill patients by Donabedian and Rosenfeld (1964) showed that
the more severe the disability, the more compliant the
patients. Haynes and other (1979) believe that the positive
association found can be explained by the fact that
increased supervision accompanies increased disability. As
a result, people with increased disability are more
compliant.

In summary, all but one of the studies reviewed

(Donabedian and Rosenfeld, 1964) found that no association

58

exists between compliance behavior and diagnosis, severity
of disease, and degree of disability. Therefore, the search
for factors that influence compliance behavior has

continued.

Features_gf_the_negimen. Several aspects of the treatment

regimen appear to have an influence on compliance. In this
section, duration of treatment, complexity of regimen, and
types of regimen will be considered.

The duration of treatment has not been shown to have
a consistent effect on compliance. For example, Blackburn
(1977), in an investigation of chronic hemodialysis
patients, found that patients' adherence to their diet
decreased with time, as measured by blood levels of
potassium and phosphorus. Similarly, Haynes and associates
(1979), in a review of the literature, also found that
adherence to treatment decreases with time. By contrast,
Ferguson and Boles (1979) found no significant relationship
between length of illness and compliance in patients with
rheumatoid arthritis.

Data do exist, however, that suggest that complexity
of medical regimen does influence compliance behavior. For
example, Hulka and associates (1975), in a study of adult
onset diabetic and congestive heart failure patients, found
that compliance was lower for patients for whom both drugs
and other recommendation were made. In a similar vein,

Green and associates (1979) found that the more drugs the

59

hypertensive patients were prescribed, the less likely they
were to say they complied with treatment or to meet
criterion of blood pressure control.

Furthermore, there are indications in the literature
that some aspects of a medical regimen are less likely to be
followed than others. For example, in a study of
hypertensive patients adherence to antihypertensive medical
regimens, Kirscht and Rosenstock (1977) found that patients'
compliance with their dietary regimen was considerably less
than their compliance with their medication regimen.
Donabedian and Rosenfeld (1964) suggest, based on a review
of the literature, that the differences in compliance with
the two treatments are due to the difficulties inherent in
inducing change in patients dietary habits. That is,
restrictions which necessitate relinquishing personal
habits, i.e., dietary changes, are more difficult to follow
than those that do not require general modifications in
behavior, i.e., taking medications (Davis, 1968).

In summary, duration of treatment, complexity, and
type of regimen may affect compliance. Therefore,
continuing to investigate the features of the regimen as
related to compliance behavior may be worthwhile pursuing.
In a subsequent section concerning chronic disease and
diabetes, it will be shown how these aspects of the regimen

relate to chronic illness.

60

WW

Several approaches have been employed to measure the
adult onset diabetics compliance with taking medication and
following diet restriCtions. These approaches include: (1)
objective measures such as pill counts, drug excretion
tests, and alpha hemoglobin: (2) clinical outcomes such as
blood sugar levels, urine glucose levels, and weight: and
(3) subjective measures such as direct observation,

physician report, and patient report.

Qbie9t12e_ueasnres. According to Gordis (1979), research

has shown that pill count as an indirect measure of
compliance behavior is open to serious question. Gordis
(1979) suggests that this is so because patients may throw
the remainder of their medication away in order to appear
compliant. Gordis (1979) also notes that drug excretion
tests, another objective measure of compliance, have
limitations. These limitations are the result of the fact
that assessment cannot be made accurately if the absorption
and excretion patterns of the patients regarding the
medication is not known or if variations among patients as
to time the medication was taken and metabolism of the drug
are not known. Moreover, Gordis (1979). notes that drug
excretion tests are limited in that often it is not possible
to detect medication or their by-products without a

detectable label. Furthermore, drug excretion tests should

61

be based on periodic assessment.

Alpha hemoglobin, however, has been found to be a
somewhat better indicator of compliance behavior of the
diabetic than other objective methods. This is so because
elevated blood sugar has a cummulative effect on the alpha
hemoglobin because glucose attaches to the hemoglobin and
the glycosylated alpha hemoglobin remains elevated for four
to six weeks. Moreover, Tattersall and his associates
(1980) state that the test has the advantages (1) of being
objective, (4) of being independent of the patient's
cooperation, (3) of and having the ability to indicate
diabetic control by a single number that reflects prevailing
blood glucose in the preceding few weeks. Nevertheless,
Leslie and associates (1979) state that a single alpha
hemoglobin estimation may relate to a shorter period of
hyperglycemia than once was suggested, and therefore,
cautioned against drawing conclusions concerning compliance
behavior from a single alpha hemoglobin estimate.

To sum up, then, objective measures have the
advantage of being independent of the subject's cooperation.
However, objective measures have several limitations as well
such as patients throwing pills away to appear compliant in
the case of pill counts, difference in individual metabolism
for drug excretion tests, and the questionable accuracy of
basing an estimation of compliance behavior from a single

measurement of alpha hemoglobin.

62

C1inical_Qutcnms__Measures. Clinical outcome measures

frequently have been used in an attempt to detect compliance
behavior in the diabetic patient. The rationale behind the
use of such measures is that diabetic control is the desired.
outcome and blood sugar and urine glucose levels, parameters
used to view diabetic control, indicate compliance.
Diabetic control, however, is limited as a measure of
compliance behavior since factors such as stress, infection,
exercise, and individual metabolism, as well as appropriate
medication and dosage, all may affect the individual's
response to her/his therapeutic regimen. In addition, blood
sugar gives only a fragmentary glimpse of the patient's
metabolic status (Bondy & Felig, 1971). Futhermore, urine
glucose is only a gross estimate of diabetic control because
blood sugar is well above normal limits before it reaches
the renal threshhold and glucose is spilled into the urine
(Tattersall et al., 1980). At the same time, the renal
threshhold for glucose varies more widely in diabetic
patients than is generally recognized (Tattersal et al.,
1980).

Still another clinical outcome measure used to
detect compliance behavior is weight maintenance or weight
loss. This is so because weight maintenance or weight loss
gives an estimate of compliance with the diabetic diet
frequently prescribed for the generally overweight adult

onset diabetic (Cohen & Etzwiler, 1976). It is necessary,

63

however, to view the weight pattern over an extended period
of time because a patient may weigh 250 pounds and not
appear to be complying, yet, a year ago he/she may have
weighed 350 pounds thus indicating he/she is complying.

In short, then, clinical outcome measures are useful
in that they are objective. However, a single measurement
of clinical outcomes is generally not adequate to base an
estimation of compliance behavior because other factors
besides adherence to a therapeutic regimen may influence the

results of the outcome measures.

Subjectiye_neasnres. Subjective measures such as direct

observation, physician report, or patient report also are
not without their limitations. Direct observation involves
the patient's compliance with diet, medication, and
follow—up appointments by someone other than the patient,
e.g., a trained observer, family member, or friend.
According to Dunbar (1979), for such methods to be accurate,
not only must the observer have clear instructions on what
behaviors are to be observed and recorded but the patient
also should be unaware that he/she is being observed.
Direct observation has an advantage in that it may detect
misinformation regarding the therapeutic regimen. However,
this method does have several disadvantages. For example,
the observer's emotions, attitudes, and prejudices may
result in inaccurate report of findings. In addition, the

observer may see what he/she wants to see or what he/she

64

expects to see, thereby recording a more positive rating
than a person with a less positive general impression might
give.

Still other subjective measures frequently used to
estimate compliance are physician and patient reports.
Gordis (1979) states that physicians' estimates of
compliance are of very limited value since physicians are
unsuccessful in identifying noncompliant patients and tend
to overestimate compliance behavior as well. By contrast,
Haynes and his colleagues (1979) state that self-report of
compliance by the patient is as effective a means of
measuring compliance as more objective means. They note,
however, that patients tend to overestimate their compliance
with a therapeutic regimen. Nevertheless, studies conducted
to substantiate the veracity of patient report, found a
significant relationship between self-report of compliance
with taking medication and blood pressure control (Green et
al., 1979: Hershey et al., 1980).

The findings of these researchers were supported by
Ferguson and Boles (1979) in an investigation undertaken to
verify patient-report of compliance behavior. Ferguson and
Boles examined the relationship between lack of belief in
benefits of treatment and compliance behavior of 40
rheumatoid arthritis patients with taking aspirin and
following an exercise program. Patients self-reported

compliance with taking aspirin and following an exercise

65

program and verification of these self-reports was achieved
(1) by obtaining salicylate levels on 12 of the 40 patients
participating in the study: and (2) by interviewing a
family member of nine of the forty patients about patient's
compliance with the exercise program. Ferguson and Boles
(199) found that in ten of the twelve patients the blood
levels confirmed the patients' reports. Moreover, they
found that 6 of the 9 family members interviewed also
verified the patient report providing support for Haynes and
his colleagues' (1979) finding that patients tend to
overestimate compliance behavior.

Nevertheless, even though patients may tend to
overestimate their compliance behavior somewhat the above
studies suggest a correlation between perceived compliance
behavior and actual compliance with the therapeutic regimen
may exist. Thus, further investigation regarding the
relationship between perceived compliance behavior and

actual compliance would be beneficial.

§ E C 1‘ III I E .

The review of the literature revealed that
approximately one-third to one-half of (all patients on
medical regimens do not comply with their treatment
regimens. As a result, many studies have been conducted to
ascertain which variables influence compliance. Most of the
studies reviewed were cross sectional and descriptive in

design. In addition, the sample sizes were small and from

66

convenience samples, therefore, the results cannot be
generalized. Nevertheless, most of the findings were
consistent from one study to the next.

The review further revealed that demographic
variables such as age, sex, education, socioeconomic status,
income, occupation, marital status, and race when examined
apart from other variables, were rarely predictive of
compliance behavior. However, demographic variables were
useful for descriptive purposes.

It was found in the review that conflicting data
exists concerning knowledge of the disease and regimen and
compliance behavior, i.e., either there was a positive
association between knowledge and compliance or there was no
association between knowledge and compliance. By contrast,
it was found that the majority of the studies that
investigated the relationship between the features of the
disease ( i.e., diagnosis, severity of disease, and degree
of disability) and compliance behavior, showed .no
association between these two variables. On the other hand,
features of the therapeutic regimen (i.e., complexity,
duration, and type of regimen) have been shown to have an
effect on compliance behavior. In short, then, it may be
worthwhile to continue to investigate the possibility of a
relationship between compliance and features of the
therapeutic regimen.

Several approaches were reviewed that could be used

67

to measure compliance behavior. It was found, however, that
each method of measurement has its shortcomings. For
example, direct observation is subject to bias, physician
report has been shown to be unreliable, a one time
measurement of clinical outcomes is not sufficient to
estimate compliance behavior, and patient self-report of
compliance behavior tends to overestimate compliance.
Nevertheless, data suggest that a relationship between
perceived and actual compliance behavior exists and that
this may be a preferred method of measuring compliance
behavior in the absence of more definitive measures.

In the next section literature concerning chronic
illness and its relationship to compliance behavior will be

considered.

g] . 1]] _ D' l I

Research suggests that compliance with a therapeutic
regimen is a problem to both the patient and the health care
provider. The research also suggests that compliance
behavior is more of a problem when compliance with a
therapeutic regimen is extended over a period of time. In
this section, the review of the literature concerning the
nature of chronic illness will be discussed, as well as the
types of chronic illnesses that have been studied in
relation to compliance behavior. The studies that
investigate the chronic illness, diabetes, in relation to

compliance behavior, will be discussed in more detail.

68

Chronic illness encompasses a variety of disorders
and may include any impairment of bodily function that
extends over a period of time. In contrast to the patient
with an acute illness of limited duration, the patient with
a chronic illness is faced with long-standing conflicts and
adjustments to which he/she must adapt (Abram, 1972).
Chronic illness frequently requires. long-standing or
permanent restriction of activities, long-standing or
permanent changes in diet, long-term administration of
medications, and a long-standing or permanent strain on the
individual's financial resources.

A wide range of studies exist that have assessed
chronic illness and compliance. Of the studies reviewed,
the chronic illnesses investigated were: myocardial
infarctions and congestive heart failure (Bille, 1977:
Hulka et al., 1975 a 1976 ): renal disease (Blackburn,
1977: Steidl et al., 1980): neurological and muscular
disorders (Diamond et al., 1968): arthritis (Ferguson. &
Boles,1979: Oakes et al., 1970): hypertension (Green et
al., 1979: Podell et al., 1976: Taylor et al., 1979):
glaucoma (Vincent, and 1971): and diabetes (Bowen et al.,
1961: Cerkoney & Hart, 1980: Hulka et al., 1976: Williams
et al., 1967: Watkins et al., 1967). Most of these studies
focused on a population with a particular diagnosis or
compared patients with similar types of diagnoses such as

congestive heart failure and adult onset diabetes. No

69

studies were found that compared compliance among patients
with acute versus chronic illness.

In spite of the lack of existent comparative
studies, most authors agree it is difficult to comply with a
regimen prescribed for a chronic illness (Barsky, 1976:
ABlackwell, 1973: Davis, 1968a). This is so because it has
been found that noncompliance increases when: (l) regimens
require more extensive behavioral changes (Davis, 1968b:
-Donabedian & Rosenfeld, 1964: Kirscht & Rosenstock, 1977):
(2) are more complex (Green et al., 1979: Hulka et al.,
1975): and (3) must be continued over time (Blackburn,
1977: Hershey et al., 1980).

These findings seem particularly applicable to the
chronic disease diabetes. First, the treatment regimen for
the disease requires that the patient make extensive
behavioral changes. For example, dietary habits and
activity patterns must often be altered. Second, the
regimen for the disease frequently is complex, requiring
dietary restriction, taking medication (insulin or oral
hypoglycemics), testing urine, engaging in regular exercise,
and practicing good ~general hygiene including special foot
care. And third, the diabeticj regimen must be continued
throughout the patient's lifetime in order to maintain
control of her/his diabetes.

Because compliance is critical to diabetic control,

several studies have used the clinical outcome of diabetic

70

control as a surrogate to measure compliance behavior. The
underlying assumption of these studies has been that the
patient whose diabetes is controlled is a compliant patient.
For example, Hulka and her associates (1975) investigated
doctor-patient communication and outcomes among.242 diabetic
patients who sought care from private physicians to assess
compliance. In the study, both patients and physicians were
questioned on instructions provided for diabetic management
and self care. In addition, clinical data were collected
concerning frequency of hypoglycemic reactions, level of
blood sugar, frequency of urine sugars greater than 2+,
maintenance of appropriate weight in relation to height, and
the occurrence of acidosis resulting in hospitalization.
And, as an additional measure of compliance, information was
gathered on appointments kept and the appropriate use of
prescription drugs. The results of the study showed that no
relationship was found between good doctor-patient
communication and diabetic control.

In a similar vein, two intervention studies have
been conducted to measure the effects of patient teaching on
diabetic control. An early study by Bowen and colleagues
(1961) investigated the effects of participation in a
planned program instruction taught by registered nurses on
51 insulin dependent diabetic patients. Following the
programed instruction, tests were administered to patients

during individual interviews to assess their knowledge,

71

skills, and attitudes. In addition, data on blood sugar
levels, weight gain, and levels of sugar in the urine were
abstracted from the patients' medical records. The findings
of the study showed no significant difference between
knowledge of diabetes and its management and diabetes
control.

The more recent intervention study, conducted by
Etzwiler and Robb (1972), investigated the acceptance and
effectiveness of a programmed machine instructions on
juvenile diabetics and their parents. It was found that
both young diabetics and their parents had increased
knowledge of diabetes and its management following
programmed instruction, and the information acquired was
retained for at least three months. However, the findings
showed that no significant change in fasting blood sugar
levels or the amount of urine glucose accompanied the
acquisition of knowledge.

Still another intervention study, conducted .by
Tagliacozzo and associates (1974), evaluated patient
teaching and its effects, not on clinical outcomes, but
rather on behavioral outcomes. The subjects in the study
were black clinic outpatients with diabetes, essential
hypertension, cardiovascular disease, and gastrointestinal
illness who received instruction on diabetes in a classroom
situation. Patient compliance was measured by regularity of

adherence with scheduled visits over a prolonged period of

72

time, adherence with the request to take medications, and
adherence with requests for laboratory tests and visits to
other clinics. It was found that the effects of nurse
intervention in the form of patient teaching were only
slightly associated with patient compliance behavior as
measured by behavioral outcomes.

Descriptive studies also have been conducted to
investigate the relationship between control of diabetes and
knowledge (Williams et al., 1967: and Watkins et al.,
1967). In Williams and his associates' (1967) study, a
sample of two hundred and eleven adult insulin-dependent
diabetics were drawn from four settings and their knowledge
of diabetes was tested and correlated with diabetic control,
as measured by body weight, presence or absence of insulin
reactions, and blood sugar or urine glucose levels contained
in medical records. In addition, data concerning urine
glucose, insulin reactions, and episodes of acidosis were
collected from patients via interviews. The findings showed
knowledge about diabetes was inversely correlated with
control, i.e., patients who knew more about diabetes were in
poorer control. Williams and his associates (1967) state
that a possible explanation for this is that the patients in
poor control have more experience with the problems,
resulting in more attention being paid to their disease.

This increased attention may facilitate increased learning.

73

The Watkins and associates' (1967) study
incorporated subjective measures of the variables of
interest rather than objective measures only. Sixty of the
subjects included in the above study were visited in their
home by two public health nurses and a medical student who,
via interview and observation using a structured protocol,
assessed the subjects' over—all knowledge of diabetes and
actual home management. The analysis of these data showed
no significant correlation between compliance and control
but a relationship between knowledge and compliance. In
other words, the more patients knew about the disease, the
better they carried out recommended therapy. But, their
compliance was not reflected in better outcomes. This would
suggest that other variables besides compliance behavior may
influence the outcome.

In short, both the Williams and associates' (1967)
and Watkins and associates' (1967) studies found a high
number of patients in poor and very poor diabetic control.
Moreover, Williams and associates (1967) found several
factors to be associated with poor control. First, the
strongest association was found between age of onset of
diabetes and control. That is, the juvenile onset diabetic
was much more difficult to. control than adult onset
diabetics. Second, psychic stress affected diabetic
control, i.e., the greater the degree of family stress, the

poorer the control of diabetes. Third, the greater the

74

satisfaction with the physician the better the diabetic
control. Fourth, in general low socioeconomic status does
not appear to limit control: in fact, patients of higher
socioeconomic status tended to be in poorer control. Fifth,
the larger the household size the poorer the patient's
diabetic control. This association was most marked among
married males who lived in large households. The authors
did not suggest an explanation of these findings. However,
Williams and associates (1967) did suggest that perhaps
there is still inadequate understanding of the nature of the
disease which results in difficulty making appropriate
medical recommendations. Thus, there continues to be a need
to investigate compliance behavior as it relates to control
of diabetes.

In summary, many studies have been conducted to
assess compliance behavior in chronic illness. A few of
these studies investigated the diabetic patient. Most of
these studies have investigated the relationship between
knowledge and compliance behavior as measured by outcome,
i.e., diabetic control. The results of these studies showed
but a tenuous relationship between knowledge and diabetic
control. The one study that found a _relationship between
knowledge and compliance behavior measured compliance by
direct observation and self-report. But, the researchers
who conducted this study found no relationship between this

subjective measure and control of diabetes.

75

The lack of relationship between compliance behavior
and control of diabetes may be due to other variables that
influence diabetic control: such as infection, stress, the
right medication, and/or the right amount of medication. Or
possibly the lack of relationship may be due to the outcome
measures that were used to measure control. Bondy and Felig
(1971) state that appropriate outcome measures used to
indicate control are difficult to identify. Their assertion
is substantiated by Malone's (1976) research in which the
accuracy of urinalyses and blood sugar tests performed on
juvenile diabetic campers, were analyzed to detect diabetic
control. The results of Malone's analysis showed that
unreliable testing of urine by patients, inconsistent
correlations between the 24 hour excretion of sugar in urine
and blood sugar, plus variability in an individual's control
as defined by conflicting same-day test results, did not
reflect the true metabolic state, i.e., diabetic control.
The inadequacy of such objective measures may reflect the
fact that they are collected at one point in time. And, as
Bondy and Felig (1971) note, it is difficult to obtain
entirely satisfactory control measures for clinical studies,
especially when a study is based on a one time measurement
of clinical parameters.

The research has provided some useful data
concerning compliance and diabetes, i.e., lack of

relationship between knowledge of disease and treatment

76

regimen and diabetic control. Yet, no studies were located
that evaluated the family's influence as related to family
functioning on compliance behavior of the adult onset
diabetic patient. In the next section family functioning
and the family's influence on compliance behavior will be
discussed.
. . . . ,
Eam1lx_£unctingingiand_fh§_fam11¥_s_Influence
Management of a patient with chronic disease usually
requires that the patient make substantial alterations and
adjustment in her/his life style. An alteration in life
style affects not only the individual with the chronic
disease but, not infrequently, her/his family as well. For
example, an alteration in diet of one individual within the
family may affect all individuals within the family. At the
same time, the manner in which a family responds to the
patient's illness may have a marked effect on the course of
the illness (Rakel, 1977). That is, patients who are
encouraged by their family members to adhere to their
regimens may have fewer complications from their disease
than those who do not receive such encouragement. Thus,
many authors agree it is necessary to treat not only the
individual but the individual within the family unit as well
(Abram, 1972: Geyman, 1977: Litman, 1974: Schmit, 1978).

In this section identification of components of

family functioning, tools to measure family functioning, and

77

the families influence on compliance behavior as it relates

to components of family functioning will be discussed.

E .1 E l' .

The interactional framework is a systematic approach
used to study families and is frequently encountered in
studies of this basic human unit. According to Nye and
Bernardo (1968) the interactional framework is a system of
viewing personal relationships within the family. Within
this framework, the individual's definition of the
situation, rather than the objective situation, is stressed
(R. King 1969) and attention is focused on love and
affection, adaptability, common interests, communication,
conflict, problem-solving, decision-making, and stress
reaction (R. King, 1969: Nye & Bernardo, 1968). These
. components of the interactional approach have been utilized
in early studies of the family.

For example, Cavan and Ranck (1938) studied families
during the economic depression of the 1930's to identify
elements of family unity. The major elements of such-unity
identified were: (1) unity in family objectives: (2)
subordination of personal interests to the good of the
family as a group: (3) adherence to family ideals: (4) the
capacity of the family to satisfy within the family circle
the personal interests of its members: and (5) the
importance of reciprocal roles and their acceptance by the

family members. On the bases of the work of Cavan and Ranck

78

(1938), Hill (1949) formulated a scale of family integration
which he used to study families under stress. Hill's scale
included five items: (1) degree of affection: (2) amount
of joint activity of family members: (3) willingness to
sacrifice to attain family objectives: (4) degree of family
.pride: and (5) the degree to which solidarity is present.
Hill's work was followed by that of Blood and Wolfe (1960)
who, in a classic study, investigated the interaction
patterns of husbands and wives. Through interviews with 909
urban and farm wives, they studied the personal
relationships of the husband, wife, and family as a whole.
On the basis of their data, they identified decision-making
within the family, the division of labor, economic function,
affection, understanding, and emotional well-being to be
important aspects of relationships among family members.

Each of the above studies identified factors that
appear to influence the well-being of the family. Moreover,
even though each study applies somewhat different labels to
these factors, each group of factors appear to be similar
from one study to the next. In addition, the components
studied seem to be consistent with the components identified
in the interactional approach.

More recently Smilkstein (1975) developed a format
to assess the family that appears to be based on the
interactional approach. That is, Smilkstein (1975)

developed a system of assessing the individual's view of

79

her/his interactions within her/his family in an attempt to
enhance the identification of family problems affecting the
patient under the physician's care. Based on family
literature, such as that cited above, Smilkstein (1975)
identified the following components which he considered
inherent in family functioning. These components are: (l)
commitment (the pledge to be responsible to and for other
members of the family including the contribution of time and
money): (2) adaptation (the capability for behavior
modification in the times of crisis or stress): (3)
mutuality (the sharing of nurturing needs by family
members): (4) differentiation (the measure of individual
maturation and development that is allowed within the family
structure): and (5) intimacy (the caring or loving
relationship that exists among family members). The
components of his format appear to be based on many of the
factors identified in the interactional framework. On the
basis of his format, Smilkstein developed the Family APGAR,
‘ an instrument used to assess family functioning to be
discussed in the next section.

In summary, early studies identified areas that
contributed to family unity. Subsequent studies used these
identified areas to develop scales to study family
functioning. These studies were descriptive and contributed
to the theoretical body of knowledge but no empirical tests

were performed. In the next section, health care studies in

80

which tools were developed .and tested to measure family

functioning will be discussed.

M l E E .1 E Ii .

In the administration of health care there is a need
to assess both the functioning of the family and that of the
individual. This is so because the health of the family as
a functioning unit and the health of the individuals are
assumed to be interdependent (Mechanic, 1962). To establish
measures of family functioning it has been necessary to
identify the component parts that make-up family
functioning. This portion of the literature ‘review will
look at research that was concerned with the development and
testing of family functioning instruments.

The earliest study reviewed was that by Geismar and
associates (1962) who attempted to devise a technique to
measure family disorganization that did not rest solely on
demographic data. In the study, family functioning was
defined by role performance of family members in nine
categories: (1) family relationships and family unity
including marital, parent-child, and sibling relationships
as well as family solidarity: (2) individual behavior and
adjustment, including the roles of husband, wife, and
children as well as the mental and physical health of family
members: (3) care and training of children: (4) social
activities, formal and informal: (5) economic practices:

(6) household practices: (7) health problems and practices:

81

(8) relationship to social worker: and (9) use of community
resources.

On the basis of these categories of family
functioning, Geismar and associates' (1961) measured family
disorganization by means of rating of role performance of
family members in each of the nine categories. One hundred
and fifty families with multiple problems were rated on a
seven point continuum ranging from inadequate family
functioning to adequate functioning, although no mention was
made of the person who rated the families. On the basis of
their findings Geismar. and his colleagues' (1962) analysis
suggest that multiproblem families have greatest difficulty
with interpersonal relationships within the family, and have
the least problems in the area of household practices,
health conditions, and economic practices. In other words,
as suggested by Geismar and associates (1962), an increased
degree of incompetence in managing interpersonal
relationships within the family tended to affect the
family's ability to provide for adequate health, food, and
shelter. Despite the usefulness of the data concerning
areas to consider as components of family functioning,
generated by the study, the methods used in this study were
not explicit and the instrument did not prove to be
reliable.

Accordingly, Haggerty (1965), building on this

earlier work, developed an extensive questionnaire and a

82

focused interview schedule to study family functioning.
Haggerty's (1965) questionnaire consisted of 75 questions
and covered areas such as demographic characteristics, past
medical experience, and internal and external functions of
the family.. Included within the questions on past medical
experiences of the family were several that elicited
information about causes of illness in the family, use of
medical and social resources, and types of preventive and
therapeutic measures used. Those that dealt with internal
functions of the family elicited information on relationship
to family of origin, relationship of present nuclear family,
stages of family life cycle, division of work in the family,
dominance in family, child-rearing practice, maternal
nurturing, as well as parental attitudes towards problem
solving, education, religion, time orientation, and adequacy
of the physical environment. Finally, the questions related
to external functions of the family elicited information on
social mobility and isolation, recreational‘ activities,
attitude towards authority groups in the community,
occupational stress and satisfaction, and relation to
neighbors.

Haggerty selected 39 families of children that had
been hospitalized to test his questionnaire and focused
interview of family functioning. Before leaving the
hospital, each parent was asked to fill out the

questionnaire. Then, an identical questionnaire was sent to

83

the family for completion four weeks later to test the
reliability of the instrument. To analyze the reliability
of the instruments, Haggerty (1965) also examined the
agreement of response between fathers and mothers.
Moreover, he had a nurse, who knew the family, rate them on
a separate questionnaire. A comparison of the these three
sets of data revealed that standard demographic
characteristics showed 88-96 percent agreement between the
different methods of obtaining data and the different
observers. However, there was an larger degree of variation
between responses about past medical experience, internal
functions of family, and external functions among the
different respondents. In short, Haggerty's efforts
resulted in the identification of family functioning similar
to those of Geismar and associates (1962), e.g.,
childrearing and nurturing, and social activites. But, just
as Geismer (1962), Haggerty (1965) too was not successful in
his attempt to develop a reliable instrument to measure
family functioning.

Thus, the effort to develop such an instrument was
continued by Pless and Satterwhite (1973) who realized that
too often many potentially important items of information
about the family are not sought or not recorded by the
health care provider, despite the fact that family
functioning is inextricably entwined with the health of the

individual. For example, according to the authors, the

84

knowledge of a family's interpersonal relationships (i.e.,
family functioning) may enable the health care provider to
appreciate more fully the significance of symptoms, to make
her/his recommendation for treatment more realistic, or
enable her/him to identify high risk situations promptly.

For the purpose of their research on the adjustment
of families of children with chronic physical illness, Pless
and Satterwhite (1973) defined family functioning as the
dynamics of everyday life: the way in which the family, as
a unit, operates across several dimensions. Based on this
definition they developed a family functioning index that
included questions dealing with intrafamily communications,
cohesiveness, decision-making, marital satisfaction, and a
general assessment of the level of happiness and closeness
of the family unit. In addition their instrument was
designed to assess the psychological adjustment of all
children, sick and well. The index was administered to the
parents of 399 school age children. Two hundred and nine of
these children had chronic disorders and one hundred ninety
were healthy.

The results of their analysis showed that the total
index score was not affected significantly by the severity
of the child's physical handicap and, therefore, the authors
believe the index can be used with families with healthy
children as well as those with chronic disorders. In

addition, analysis, based on a comparison of index scores

85

with rating of the same families by experienced social
workers, showed the family functioning index to be valid.
Moreover, the high correlation (0.72) found between the
scores of the husband and wife concerning each one's view of
her/his family functioning demonstrated the reliability of
the index. Pless and Satterwhite (1973) viewed the
husband-wife comparison as a special example of test-retest
reliability and considered it to be a more rigid criterion
of reliability.

On the basis of their finds, Pless and Satterwhite
(1973) concluded the instrument would identify chronically
ill children who were likely to experience secondary
psychological difficulties. Moreover, they believe the
instrument is a good tool to acquire knowledge of family
life that may influence health behavior since it is easily
administered in the clinical setting. Despite these claims
to its utility, however, the instrument does have
limitations, i.e., it is geared towards families with young
children. 7

The research for an instrument to measure family
functioning has been continued by Smilkstein (1978) who
defined family functioning as that behavior which promotes
emotional and physical growth and maturation of all members.
His Family APGAR represents the most recent attempt to
develop a practical instrument to assess family functioning

and is a brief screening questionnaire designed to elicit a

86

data base that reflects the patient's view of the functional
state of her/his family.

Smilkstein used the acronym APGAR to outline the
five components which appear to represent the most common
themes in the literature that deal with family functioning.
In brief these are: Adaptation, Partnership, Growth,
Affection, and Resolve. Adaptation refers to the family's
ability to utilize intra and extrafamilial resources for
problem-solving when family equilibrium is stressed during a
crisis. Partnership pertains to the sharing of
decision-making and nurturing responsibilities by the family
members. Growth refers to the physical and emotional
maturation and self-fulfillment that is achieved by family
members through mutual support and guidance. Affection is
the caring or loving relationship that exists among family
members. Resolve pertains to the commitment to devote time
to other members of the family for physical and emotional
nurturing and the sharing of money and space as well. The
questionnaire designed by Smilkstein measures family
members' satisfaction with each of the five basic components
of family functioning and identifies problem areas within
the family unit of both nuclear and alternative life-style
families.

The Family APGAR Index has been tested by Good and
her associates (1979) for construct validity. In their

study, Good and her colleagues (1979) administered the

87

questionnaire to a total of 33 subjects from two sample
groups: one nonclinical group of ”normal families" and one
clinical group of psychiatric outpatients. In addition,
they administered the Family Functioning Index (FFI)
developed by Pless and Satterwhite (1973), for which
validity and reliability was already established, to the two
groups. And finally, they had clinical therapists rate "low
functioning families" after having met with these patients
on the average of 5.5 times the preceding year.

To analyze the three sets of data, Good and her
associates (1979) used the comparative method. They found
that the Family APGAR scores showed a strong correlation of
.80 with the FFI scores. By contrast, they found only a
correlation of .64 between the Family APGAR Index score and
the therapists' family evaluation score and a moderate
correlation of .67 between interspouse scores.

On the basis of their findings, Good and associates
(1979) concluded that, the Family APGAR is an appropriate
tool for health care providers to use in screening their
clients for family difficulties that may affect their health
behaviors. In addition, they suggested that the tool may be
a useful instrument for research. However, since to date
the tool only has been used with young, student families the
authors suggested that the instrument should be further
tested on other types of family units. That is, the

usefulness of the instrument for families in different

88

stages of the life cycle is yet to be tested. Despite the
lack of empirical data to support its utility across the
developmental span of the family, the Family APGAR Index has
several advantages that indicate it may be a useful research
tool. That is, it is short and easy to score and,
furthermore, applicable to both nuclear and alternative
life-style families.

In summary, attempts to develop instruments that can
be used to assess family functioning have been reviewed.
Geismar and associates' (1962) instrument was neither
explicit nor reliable. Haggerty's (1965) instrument was too
.extensive for clinical use and did not show reliability.
Pless and Satterwhite's (1973) Family Functioning Index
showed validity and reliability, and was easily administered
in the clinical setting. But, this instrument was designed
for use with families in which there are children.

In contrast, Smilkstein's (1978) Family APGAR Index,
which showed reliability and validity, is short, easy .to
score, contains clear and concise definitions of family
functioning components, and is applicable to the situation
of both nuclear and alternative life-style families.
Moreover, the components included in Smilkstein's index
essentially encompasses other authors' concepts of family
functioning. Nevertheless, Smilkstein's tool does have
limitations. First, the applicability of the index for

families in different stages of the life cycle has yet to be

89

tested. In addition, the instrument is only a screening
tool and would have to be expanded if used as the basis of a
study instrument. Yet, it appears to be the best tool thus
far developed and it is believed that it has potential, if
expanded, for use in this research project.

In the next section the review of the literature
will attempt to elucidate why this is so. That is, it will
contain a review of studies that focus on the family's
influence on compliance behavior, as well as studies that
have shown a positive relationship between family
functioning and compliance. In addition, in the next
section these studies will be related to the five components

of family functioning.

WW

Although little attention has been given to the
family's influence on the compliance of the adult onset
diabetic patient, several studies exist that have considered
the family's influence on compliance behavior in general.
These studies have focused on the family's influence on:
(1) psychiatric patients dropping out of treatment: (2)
drug, alcohol, and smoking withdrawal programs: (3)
preventative programs such as family planning and physical
fitness programs for prevention of heart disease: and (4)
patients with chronic problems such as hypertension,
arthritis, orthopedic disabilities, kidney problems, and

other chronic diseases. In the following sections, several

90

of these studies will be reviewed and related to
Smilkstein's (1978) components of family functioning.

As noted above, five components are included in
Smilkstein's (1978) Family APGAR. The first component,
adaptation, is defined as the degree to which individuals
aid each other in the time of need, i.e., the ability of
family members to adapt to the situation and give support.
Studies have been conducted that have assessed such
adaptation. For example, Litman (1966) conducted a study of
100 patients who suddenly found themselves with a severe
orthopedic disability and were undergoing physical
rehabilitation. In his study, Litman attempted to evaluate
the influence of family adaptation and support on the
patient's acceptance of her/his condition as perceived by
the staff, her/his general physical state, her/his
rehabilitation potential and her/his overall motivation and
performance. Reinforcement by the patient's family was
measured by a paper and pencil test to ascertain the degree
to which the patient perceived: (l) agreement of family
members: (2) cooperation: (3) concern for each others'
welfare: (4) enjoyment of association: (5) affection: (6)
esteem or admiration for each other: (7) interest,
confidence, and respect for each other: and (8)
satisfaction with family relationships. The perceptions of
the staff were collected via a set of staff evaluation

questionnaires completed by the attending physician, the

91

physical therapist, and the occupational therapist.

Analysis of the data collected showed that 73
percent of the 58 subjects with a good response to physical
rehabilitation, as perceived by the health care staff,
received positive reinforcement from their families. By
contrast, 77 percent of the 42 subjects with a poor response
did not obtain this encouragement from their families. As a
result of these findings, Litman concluded that cooperation
and concern for others' welfare were especially necessary
for adaptation in the time of crisis.

The second of Smilkstein's components, partnership,
also has been examined in two studies of compliance
behavior. According to Smilkstein (1978), partnership is
defined as the degree to which family members communicate
with each other about matters such as medical care or
finances. This concept was included in a study by Cakes and
associates (1970) in which it was found that perceived
family expectations were strongly related to arthritic
patients' compliance with the use of a hand splint. In
their study, Cakes and associates (1970) collected data from
66 rheumatoid arthritis patients by asking them two
questions: one asked whether their family members expected
them to wear their splint and the other asked the percentage
of time in the last 18 months they had worn the splint.
Additionally, a family member was asked to respond

independently to the question regarding the extent of time

92

the patient used the hand rest splint.

Another study that investigated partnership was
conducted by Heinzelmann and Bagley (1970) among a group of
sedentary middle-aged men who were at high risk for
developing heart disease. In the study, participants were
asked to exercise for one hour three times a week for
eighteen months. In addition, prior to beginning the
exercise program a telephone health behavior survey was
administered to potential participants. Then, once subjects
had been enrolled into the study, additional data concerning
their attitudes, beliefs, and behavior were obtained by
interview as well as questionnaires. This additional data
was obtained at 3 to 4 month intervals during the program
and at the time of termination. Analysis of the data showed
that those men who had wives with a positive attitude toward
a physical conditioning program made the chance of
compliance with the program 80 percent. By contrast, if the
wife's attitude was neutral or negative, the chance of
participation fell to 40 percent.

In short, both Cakes and associates (1970) and
Heinzelmann and Bagley (1970) found that perceived family
expectations and support resulted in increased chance of
compliance. These findings suggest that verbal and /or
nonverbal communication between the family members or
partners have a positive influence on compliance. Moreover,

since communication about matters such as health care is an

93

aspect of partnership and partnership is a component of
family functioning, it might be expected that family
functioning will affect compliance behavior.

The third component of family functioning defined by
Smilkstein (1978) is growth. According to Smilkstein,
(1978), growth in individuals results when family members
encourage and/or accept change, and give mutual support and
guidance. One example of such growth that is related to
health behavior is quitting smoking since it requires a
change in life habits and therefore, reflects change or
growth. A study that supports this concept was conducted by
West and colleagues (1977) who evaluated the effects of
family support upon quitting smoking. Eight hundred
volunteer subjects attended a smoking clinic which used
medication and an educational program to assist patients in
quitting smoking. Five years later the researcher was able
to contact successfully 559 of the original 800 patients and
have them complete a mailed questionnaire or a questionnaire
administer via telephone. The patients were asked about
their present smoking habits and those of the people with
whom they associated. In addition, they were asked about
the support they perceived they received from their spouses
as they tried to quit. The results of the analysis showed
that patients who had quit smoking were more likely than
those who were still smoking to have spouses who had never

smoked or had quit smoking. Moreover, two-thirds of the

94

quitters perceived that their spouses had provided support
that made it easier for them to quit smoking. By contrast,
only about one-third of the smokers perceived that they
received this support.

Still another component of family functioning
defined by Smilkstein (1978) is resolve which he defines as
the commitment to devote time to other members of the family
for physical and emotional nurturing and to share wealth and
space. This component has been examined in a number of
studies and shown to influence compliance. For example,
Baekeland and Lundwell (1973), _in a review of research
literature on dropping of the treatment, found that in 19 of
the 19 studies examined, social isolation and/or lack of
affiliation was reported to be a major cause for
discontinuing treatment in psychiatric patients. That is,
patients social isolated and/or who lacked affiliation with
anyone, were void of the influence of family functioning
either positively or negatively. Those in social isolation
lack support from family members for physical and emotional
nurturing.

The last component of Smilktein's (1978) to be
discussed is affection which he defined as emotional support
and caring. All of the above studies suggest that emotional
support and caring from other family members is conducive to
compliance behavior. In addition, several studies have been

conducted that have explored the relationship of social

95

support and/or family support to compliance behavior. For
example, Caplan and associates (1976) who defined social
support as encouragement provided by another person or group
(friend, family acquaintence, health care provider) which
moves the person .toward a goal, studied a group of
hypertensive patients to identify factors which determine
why patients do not adhere to their regimens. The findings
of their study confirmed the influence on compliance
behavior of the patient's perception of support from family
members, especially the spouse, as well as support from the
health care provider. The findings showed, however, that
social support from other sources, such as best friends, was
not related to compliance behavior.

However, the findings were not supported in a later
study conducted by Ferguson and Boles (1979) which showed
that family support did not routinely produce compliance
behavior. In their study, Ferguson and Boles (1979)
questioned 40 patients with arthritis to determine the
relationship between family support, health beliefs and
compliance with taking aspirin and following an exercise
program. They found that the presence of family support and
understanding as perceived by the patient did not routinely
‘produce compliance. The authors suggest that this
unexpected finding may be due to the fact that the overall
number of nonsupportive families was small. And

furthermore, the family groups were stable (as measured by

96

patient report and investigators' clinical impression).

In summary, the majority of studies reviewed showed
a positive association between specific aspects of family
functioning and compliance behavior. Nevertheless, none
investigated the total concept of family functioning as it
relates to compliance behavior. Only one study was located
that has examined these two concepts conjointly. This lone
study, by Steidl and colleagues (1980), investigated the
adherence to treatment and family functioning of patients
receiving long-term kidney dialysis treatment. In the
study, family functioning was assessed by experienced family
therapists. An analysis of the data revealed no significant
relationship was found between family functioning and
compliance behavior in the kidney dialysis patient.
However, Steidl and colleagues did find a relationship
between compliance and family functioning for those patients
in which family members: (1) exhibit respectful, shared
adult leadership and parental coalitions (partnership); (2)
have the ability to take responsibility and have an open
responsive attitude towards the opinions of others (growth):
and (3) demonstrate effective problem-solving skills
(adaptation).

In summary, the majority of studies reviewed showed
a positive association between family influence and
compliance behavior. Moreover, in several studies certain

components of family functioning were shown to influence

97

compliance behavior. Only one study reviewed, however,
investigated the total concept of family functioning as it
relates to compliance behavior. And, in that study, family
functioning was based on the perception of a family
therapist rather than on the perception of the patient. To
date then, neither the relationship between perceived family
functioning and compliance behavior nor the relationship
between perceived family functioning and perceived
compliance behavior in the adult onset diabetic patient has

been studied.

SummarY

In this chapter one dimension of the literature
reviewed related to compliance behavior. It was found that
the majority of studies showed little or no association
between demographic variables and compliance levels when
each variable is independently examined. In addition, those
studies that had investigated the relationship between
knowledge of the disease and regimen and compliance behavior
produced conflicting results. Furthermore, studies of
disease features such as diagnosis, severity of disease, and
degree of disability showed that these variables apparently
'have little or no association with compliance behavior.
However, studies designed to examine features of the
regimen, such as duration of treatment, complexity and type
of regimen all found that the nature of the treatment

regimen affects adherence to a regimen. That is, the longer

98

the duration of treatment and the more complex the regimen
the less likely the patient is to comply with it.

The research reviewed also indicated that several
approaches have been employed to measure compliance
behavior, i.e., direct observation, physicians report,
clinical outcomes, and patient self-report. Each of these
approaches have their limitations, eg., clinical outcomes
are affected by other things such as infection, metabolism,
and stress as well as compliance behavior. Futhermore, the
literature review suggests that it is difficult to identify
appropriate outcome measures that indicate diabetic control.
However, research has shown a relationship between perceived
(self-reported) and actual compliance behavior exists.

In addition, the literature reviewed revealed that
patients have more difficulty in compliance when the illness
is chronic than when it is acute. Since diabetes, a chronic
illness that requires extensive behavioral changes is the
focus of this study, only investigations in which diabetics
were studied were reviewed. This review showed that only a
tenuous relationship exists between knowledge of the disease
oand its regimens and diabetic control. That is, only one
study reported a relationship between knowledge and
compliance behavior and this finding was based on measures
of direct observation and self-report.

Finally, literature was reviewed on the measurement

of family functioning and the family's influence on

99

compliance behavior. It was found that only a limited
number of instruments have been developed to measure family
functioning. Moreover, almost all of these instruments had
limitations, e.g., not reliable, too long, geared towards
families with young children. Although there are few
instruments to measure family functioning several studies
did assess the family's influence on compliance behavior.
The majority of studies did show that a positive association
between family support and compliance behavior exists. In
addition, the results of these studies showed that certain
aspects of family functioning appear to be related to
compliance. Only one study, however, viewed the total
concept of family functioning as it relates to adherence to
a regimen. Furthermore, no studies were located that viewed
the total concept of perceived family functioning as it
relates to perceived compliance behavior in the adult onset
diabetic.

In short, the review of the literature showed that
(l) perceived compliance behavior may be an effacious
measure of actual compliance behavior: and (2) a need
exists to study the relationship between compliance behavior
and the basic unit of living with which the patients' health
is inextricably entwined. In Chapter IV the methodology and
procedures used in the study to answer this need will be

presented.

100

CHAPTER IV

METHODOLOGY AND PROCEDURE

Quarries

This descriptive study was designed to identify the
relationship between perceived family functioning and
perceived compliance behavior with selected aspects of the
prescribed therapeutic regimen among ambulatory adult onset
diabetic patients. The research design used for a study of
this nature is the correlational design which determines the
extent of relationship between existing variables. It
permits the free variation of both variables of interest--
perceived family functioning and perceived compliance
behavior. Neither variable was manipulated by the
researcher.

The questionnaire developed for the study consisted
of three parts. The first part was used to obtain
descriptive data (Appendix A). The second part of the
questionnaire measured the variable perceived family
functioning (Appendix B) and was correlated, by means of
Pearson Product-Moment Correlation, with the third part of
the questionnaire which measured the variable perceived
compliance behavior (Appendix C). The population was a
convenience sample of 25 subjects. The subjects were

referred by private physicians in Flint, Michigan.

101

Q l' J D E' 'l' E I] y . I]
E i i E .1 E Ii .

The variable, perceived family functioning, was
reflected by the score on the family functioning section of
the questionnaire (Appendix B, items 1-26). The family
functioning section of the questionnaire was developed by
the researcher and was based on Smilkstein's (1978)
definition of family functioning. To provide a better
understanding of the variable family functioning, the
definition of family was included. Smilkstein (1978)
defined family as a group consisting of the patient and one
or more persons, children or adults, in which there is a
commitment for members to nurture each other. This
definition encompasses all types of families. Smilkstein
(1978) further defined family functioning as that behavior
which promotes emotional and physical growth and maturation
of all members. Smilkstein (1978) established parameters by
which a family's functional health could be measured .by
selecting five basic components of family
function--adaptation, partnership, growth, affection, and
resolve. Following is a definition of each component and
the questions used to measure each component:

Adaptation is the utilization of intra and

extrafamilial resources for problem solving when

family equilibrium is stressed during a crisis.

Adaptation was measured by questions 1, 7, 12, 16,

102

22 (Appendix B). The questions focus on giving and
receiving help in times of need.

Partnership is the sharing of decision-making and
nurturing responsibilities by family members.
Partnership was measured by questions 3, 8, 14, 21,
23 (Appendix 'B). The questions focus on expressing
opinions, discussing problems, sharing
responsibilities, and sharing common interests.
Granth is the physical and emotional maturation and
self-fulfillment that is achieved by family members
through support and guidance. Growth was measured
by questions 4, 6, 9, 19, 20, 24 (Appendix B.) The
questions focus on encouraging independence and
accepting changes within the family unit.

Affeerien is the caring or loving relationship that
exists among family members. Affection was measured
by questions 10, ll, l3, 17, 25 (Appendix B). The
questions focus on expressing emotions i.e., love
and anger and the feeling that other family members
care.

Beanlye is the commitment to devote time to other
members of the family for physical and emotional
nurturing. Resolve was measured by questions 2, 5,
15, 18, 26 (Appendix B). The questions focus on
working together, spending time together, and

sharing space and money.

103

The family functioning section of the questionnaire
was developed to meaSure the adult onset diabetic's
perception of the interactions within her/his family.
Perception was defined as the individual's image of reality
(King, 1979). The score on the family functioning section
of the questionnaire will give an estimate of how the
individual views her/his family functioning. Items were
worded in a way to measure the diabetic's perception (e.g.,

My family spends time together for family activities).

2 . i C 1‘ E l .

The variable, perceived compliance behavior with
selected aspects of the therapeutic regimen of the adult
onset diabetic, was reflected by the score on the compliance
section of the questionnaire (Appendix C, items 1-24).
Compliance behavior is defined as the extent to which the
patient perceives her/his behavior coincides with medical or
health advice (Haynes et al., 1979).

The compliance behavior section of the questionnaire
was designed to measure the adult onset diabetic's
perception of compliance with selected aspects of the
therapeutic regimen. Selected aspects of the therapeutic
regimen are defined as diet, medications (insulin or oral
hypoglycemics), and follow-up appointments. For diet and
medication, the subjects were asked to estimate their
compliance behavior within the two week period prior to

filling out the questionnaire and within a six month period

104

of time for follow-up appointments. Following is a
description of each dimension and the questions used to
measure each:
Qie;_includes adherence to general principles such
as eating the recommended foods in the recommended
amounts, at the appropriate times, and attempting to
lose or control weight. Perceived compliance with
diet was measured by items 1, 3, 4, 6, 7, 9, 12, 15
(Appendix C).
Medieatigna (insulin or oral hypoglycemics) include
taking medications in prescribed doses, at
prescribed times, watching for side effects, and
having prescriptions refilled. Perceived compliance
with medications was measured by items 2, 5, 8, 10,
ll, l3, 14, 16 (Appendix C).
Fellen;np_appnintmenta include keeping well and ill
follow-up appointments and keeping laboratory
appointments for the assessment of blood sugar.
Perceived compliance with follow-up appointments
were measured by items 17-24 (Appendix C).
Items were worded to determine the diabetic's
perception of her/his compliance (e.g., I have followed my

diet as ordered by the doctor).

105

Clini§a1_2a£iahlsfi

. Clinical outcome variables such as blood sugar give
a fragmentary glimpse of diabetic control (Bondy & Felig
1971). Although blood sugar gives an estimate of diabetic
control, blood sugar is inadequate as a measure of
compliance behavior since many variables influence the blood
sugar other than adherence to the diabetic regimen (Guthrie
& Guthrie, 1977). Nevertheless, data concerning the blood
sugar was obtained from the patient's chart to ascertain if
there was a correlation between the patient's blood sugar
and her/his perceived compliance behavior. A positive
finding would offer veracity to the patient's self-report of
compliance behavior.

The blood sugar, drawn within the month the
questionnaire was completed, was obtained from the patient's
record. The subject was considered within normal limits if
her/his random blood sugar was between 70-120 mg/dl. Skyler
and associates (1981) indicate that random blood sugar 'is
acceptable if between 60 and 180 mg/dl although they state
the 2 hour postprandial should be 150 mg/dl or less.
Williams and Alling (1980) state it is desirable for the
diabetic patient to maintain her/his blood sugar levels
within the range of 50-150 mg/dl. For this study the random
blood sugar will be considered acceptable if it is between

60-150 mq/dl. Subjects whose random blood sugar levels are

106

less than 60 mg or greater than 150 mg will be considered
out of control.

Weight in relation to height and sex is another
clinical variable that can offer veracity to a patient's
report concerning compliance with diet. The height, sex,
and most recent weight of each patient was obtained from the
patient's record and recorded on the chart audit sheet
(Appendix D). The weight was evaluated by using the
Metropolitan Life Insurance height, weight chart (Capell,
1976). Weight was categorized arbitrarily into five groups:
(1) weight within normal limits for .height and sex: (2)
less than 20 pounds. overweight: (3) 20-50 pounds
overweight: (4) 51-100 pounds overweight: and (5) more

than 100 pounds overweight.

MW

According to Haynes and his associates' (1979)
extensive review of the literature, research did not reveal
sufficient evidence to consider demographic characteristics
as confounding variables. The confounding variables
investigated in this study were: (1) number of years with
diabetes: (2) number of chronic illnesses the patient had
in addition to diabetes: (3) complexity of the patient's
regimen: and (4) whether the patient was hospitalized for
diabetes within the last six months. Complexity of the
regimen was assessed by (1) number of medications: (2)

number of times per day medications were taken: (3) number

107

of other diet restrictions: and (4) number of treatments.
Information concerning the type of diabetic medication
(insulin or oral hypoglycemic) prescribed for the patient
was also elicited. That is, each of these confounding
variables may have had an effect on the way the adult onset
diabetic responded to items on the questionnaire concerning
perceived compliance behavior. The confounding variables
were collected by means of the confidential data sheet
(Appendix A).

In summary, the major variables in this study
included the adult onset diabetic's perception of her/his
family functioning and the perception of her/his compliance
behavior. Confounding variables included the number of
years the patient had diabetes, the number of other chronic
illness in addition to diabetes, complexity of regimen, and
whether the patient was hospitalized for diabetes within the

last six months.

Sample
The population selected for the study was adult

onset diabetics. Standard criteria used for admission into

the study included the following:

- The subject was living with at least one other
person since this study was concerned with family
functioning. A family consists of a group con-

sisting of the patient and one or more persons,

108

children or adults, in which there is a commitment
for members to nurture each other (Smilkstein,

1978).

The age at diagnosis was 30 years or older. The
subjects were to have adult onset diabetes. Only
rare cases of juvenile-type diabetes appear after

age 30 (Cohen & Etzwiler, 1976).

The subject was a diagnosed diabetic at least one
year prior to time of selection. One year allows
for a time period of adjustment to an illness

(Peretz, 1970).

The age at time of admission to the study was no
more than 70 years of age and no less than 31 years
of age. Thirty-one years at the time of admission
allowed time for the subject to have been diagnosed

for one year.

The subjects were on a calculated diabetic diet. A
calculated diabetic diet is a diet based on a
prescribed number of calories that are obtained from
a prescribed amount of carbohydrate, protein, and
fats. Many physicians have their patients use the
American Diabetic Association exchange list. The
American Diabetic Association devised an exchange

system where a serving constitutes an exchange and

109

is comprised of fixed gram weights of protein, fat,
and carbohydrates. The exchange list indicates the
foods allowed and the size of servings. The number
of prescribed calories dictates the number of
exchanges allowed from each food group. Some
physicians allow their patients to be on a "free"
diet or self-chosen, unmeasured diet and tell the
patients to avoid an overindulgence of sweets
(Robinson, 1980). Diet is basic to all diabetic
regimens. Timing of food intake, the caloric value
of food ingested, and the proportions and quality of
carbohydrate, fat, and protein are all important

aspects of the diet (Skillman & Tzagournis, 1977).

The subject was on an oral hypoglycemic or less than
60 units insulin per day. According to Blevins
(1979), diabetics who are not controlled by diet
alone are frequently placed on insulin. Blevins
(1979) further states that oral hypoglycemics are
used for patients who are not at risk for
cardiovascular disease. Cohen and Etzwiler (1976)
state that the complexity of the regimen is greater
for those who are on oral hypoglycemics or insulin
than those on diet alone. Researchers have found
that complexity of the regimen influences compliance

behavior (Haynes et al., 1979). Therefore, in an

110

effort to keep the complexity of the regimen
somewhat similar, only those on oral hypoglycemics
or less than 60 units of insulin were included.
Patients who require over 60 units of insulin were
excluded because these patients are considered
difficult to control and may have juvenile onset

diabetes (Guthrie & Guthrie, 1977).

- The subject was seen at a minimum of two times per
year in an ambulatory care setting for follow-up
appointments for her/his diabetes. Most diabetics
need to be seen at two to three month intervals for
continued management, although the interval varies
with the patient, the nature of the disease, and the

physician (Cohen & Etzwiler, 1976).

- The subjects were able to read and complete the
questionnaire since someone else reading the
questionnaire to the subject may alter the way the

subject answered the questions.

- Pregnant women were excluded from the study since
more health care attention is given to the pregnant
diabetic, thereby creating stronger motivation to
comply.

To be included in the study, then, the subjects had
to meet the nine criteria listed above. But, they were not

excluded if either of the following variables were present.

111

Diabetics with other chronic illnesses were not excluded
since most diabetics have other chronic illnesses (Guthrie &
Guthrie, 1977). Of course, the presence of another chronic
illness further increases the complexity of the regimen,
thereby frequently resulting in increased noncompliance
(Haynes et a1. 1979). But, since it would have been
difficult to obtain a sample of patients with only diabetes,
subjects with multiple diseases were included. The number
of other chronic illnesses was recorded on the patient data
sheet (Appendix A) and was considered a confounding
variable.

Diabetics who were hospitalized within the last six
months for their diabetes were also included. When a
patient is hospitalized due to uncontrolled diabetes the
individual may receive reinforcement teaching that may serve
to improve her/his compliance temporarily. The patient was
asked if he/she had been hospitalized within the last six
months for diabetes or if he/she had had diabetic teaching
while hospitalized for another illness. Hospitalization for
diabetes was considered a confounding variable (Appendix A).

In summary, the subjects included in this study
were: (1) living with at least one other person: (2)
diagnosed at 30 years or older: (3) diagnosed at least one
year prior to the time of selection: (4) between the age of
31 and 70 years of age: (5) on a calculated diabetic diet:

(6) on insulin or oral hypoglycemic: (7) seen at least two

112

times per year in an ambulatory care setting; (8) able to
read and complete the questionnaire: and (9) not pregnant.
Diabetics who had been hospitalized within the last six
months and who have other chronic illnesses were not
excluded. Data concerning hospitalization and other chronic
illnesses were gathered per confidential data sheet

(Appendix A) and were considered as confounding variables.

W

The subjects for this study were recruited from the
practices of four private physicians in the Flint area.
Three of the physicians were internists and one was a
general practitioner. Two of the internists treated many
diabetics. Twenty of the 25 subjects were obtained from
those two offices. Four subjects were from the third
internist's office and one was from the general
practitioner's office.

The sample size was small and the participants
represent a convenience sample. Therefore, the

generalizability of the findings is limited.

W
The researchef gained consent, verbally and in
writing, from eaCh of the four physicians (Appendix E). The
researcher discussed with each of the office nurses the type
of patients needed for the study and distributed to them a

check list of the criteria that must be met to be admitted

113

to the study (Appendix F.) The nurses used the check list
as a guide in selecting the diabetic patients they referred
to the researcher.

As the diabetic patient came to the physician's
office for her/his appointment, the nurse asked the patient
.if he/she would consider participating in a research
project. The nurse gave the patient the introductory letter
explaining the nature of the study (Appendix G) and the
consent form (Appendix H). If the subject agreed, he/she
was asked to sign the consent form.

The researcher contacted each physician's office
every three to four days and went to their offices when the
physicians had some signed consent forms from the patients
who agreed to participate. The researcher collected the
consent forms and obtained information from the patients'
records to ascertain if each subject met the specified
criteria necessary to be a participant in the study
(Appendix F). In addition, a chart audit was done to gain
information concerning the patient's blood sugar values,
dates of office visits in last six months, height, and
weight. This information was placed on the chart audit
sheet (Appendix D). The telephone numbers of those who met
the specified criteria were obtained from the patients'
records.

The researcher contacted the potential subject by

telephone, identified herself, and asked the subject if

114

he/she had any questions concerning the research project.
The researcher then arranged a time and a place for the
patient to answer the questionnaire such as .the patient's
home, physician's office, or place of employment.

Prior to filling out the questionnaire the subject
was assured of anonymity, confidentiality, and the right to
refuse. The subject was requested to complete the
confidential data sheet (Appendix A) and the questionnaire
(Appendix B a C). The researcher was available to answer
any questions the subjects had concerning the questionnaire,
diabetes, or family functioning. The researcher utilized
the time to review sections of the thesis in order to guard
against unnecessary pressure or observation of the subject.
No time limitation was placed on completing the
questionnaire. However, most of the subjects took ten to
twenty minutes to complete it. After completion of the
questionnaire, the subjects were again encouraged to ask any
questions they had regarding the questionnaire, diabetes, or
family functioning. A copy of the consent form with the
researcher's telephone number was given to the subject in
case a question should arise at a later date.

At the completion of the data collection phase, the
data were coded on coding sheets and key punched to prepare

for data analysis.

115

H 8' ll 2 l I'

The measures taken to ensure that the rights of the
participants were protected are presented in this section.
The subjects were not solicited. All participants were
informed of the purpose of the study and that participation
in the study was voluntary. The participants were informed
of their right to refuse participation or to withdraw from
the study without it affecting the care they were presently
receiving. They were assured of anonymity, by the
researcher explaining that there would be no name on the
questionnaire and that the consent form would be separated
from the questionnaire. Subjects were also assured that no
one from the ambulatory care setting would be permitted to
read the questionnaire following its completion. The
subjects were encouraged to ask questions following
completion of the questionnaire.

The subjects were given a copy of the consent form
which informed them of any possible adverse effecrs and the
measures that would be provided if they experienced any of
them. The researcher's name and telephone number were on
the consent form enabling the subjects to withdraw from the
study or ask questions that may have been stimulated by the
study. Precautions were taken to ensure the rights of the

participants.

116

Th£_lnatrnment

The questionnaire utilized for this study contained
two parts. Part A of the questionnaire contained questions
concerning perceived family functioning and part B included
questions concerning perceived compliance behavior. Both
parts were developed by the researcher based on a review of
the literature and from the operational definitions of
perceived family functioning and perceived compliance
behavior. The instrument was administered to three adult
onset diabetic patients who met the specified criteria. The
questionnaire took 10-20 minutes to complete. After the
completion of the questionnaire, the patients were
interviewed to ascertain any difficulties they may have had
with understanding the directions or the questions asked.
None of the patients expressed any difficulty. Since it was
not necessary to alter any questions, the data collected in
pretests were used in all analyses. In this section the
development, scoring, reliability, and validity of the

instrument is presented.

2 . i E .1 E I' .

A review of the literature revealed that several
attempts have been made to establish instruments that would
measure the quality of family life, its style, or level of
family functioning (Geismar, 1962: Haggerty, 1965: Pless
and Satterwhite, 1973: Smilkstein, 1978). The most recent

attempt was Smilkstein's Family APGAR Index (Adaptation,

117

Partnership, Growth, Affection, Resolve), a screening tool
designed to elicit a reflection of the patient's view of
her/his family functioning. Smilkstein's (1978) definition
of concepts, examples of questions on the questionnaire, and
the scoring were made explicit in his published article.

Previously the family APGAR has been tested
primarily on younger families. The researcher believes that
the components of the family APGAR (adaptation, partnership,
growth, affection, and resolve) are applicable to all stages
of development of the adult if the individual lives within a
family unit. This is because throughout a persons life-time
one needs to be able to adapt in the time of crisis, give
and receive support and guidance, share in decision-making
and household responsibilities, give and receive affection,
and feel a sense of belongingness.

Good and associates (1979) showed the Family APGAR
to have construct validity by comparing the instrument's
score with the scores of an established family functioning
test (Pless and Satterwhite's Family Functioning Index,
1973) and the scores of clinical therapists. There was a
strong correlation of .80 between the Family APGAR Index
score and the Family Functioning Index and a moderate
correlation of .64 between the Family APGAR Index score and
the clinical therapists' score.

Smilkstein's definition of family was considered

appropriate to this study since his definition included the

118

nuclear family as well as alternative life-style families.
The concepts used in the Family APGAR represented the common
themes in the social science literature dealing with
families. The family functioning section of the
questionnaire was developed by the researcher for this study
following a review of the literature.- The questions
constructed were based on Smilkstein's work (1978)
concerning the definitions of the components of family
functioning and the open-ended questions used by Smilkstein
to elicit family functioning information (Appendix I).
Refinement of the questions followed the evaluation by
thesis committee members. Changes were made in wording and
arrangement of items on the questionnaire.

The family functioning section of the questionnaire
contained five to six items concerning each component of
family functioning with a total of twenty-six questions.
The questionnaire was in the form of an opinion scale so
there were no right or wrong answers. Some statements were
stated positively and some negatively. The statements were
answered according to a five point Likert scale.

Each subject was asked to indicate the degree to
which he/she agreed with each statement made-- strongly
agree, agree, undecided, disagree, strongly disagree.
Statements stated positively (Items 1, 2, 4, 6, 11, 14, 17,
18, 20, 21, 22, 25, 26 Appendix B) were rated as strongly

agree - 5 points, agree - 4 points, undecided - 3 points,

119

disagree - 2 points, and strongly disagree - 1 .point. The
negatively stated statements (Items.3, 5, 7, 8, 9, 10, 12,
13, 15, l6, 19, 23, 24 Appendix B) were scored in the
reverse. For example, the positive statement, ”My family
accepts my wishes to take on new activites', was scored five
points if answered "strongly agree". The negatively. stated
item, "My family gets upset when I spend time away from
home" was scored one point if answered ”strongly agree”.
The higher the score, the more positive the family
functioning was perceived.

Although the parameter that measures perceived
family functioning was divided into five components
(adaptation, partnership, growth, affection, resolve), only
the total score was used for statistical purposes. This was
done because the purpose of this study was to evaluate the
total concept of perceived family functioning as it related
to perceived compliance behavior. However, the raw scores
of the components were viewed separately to ascertain .if
there was a marked difference in the way the subjects viewed
each component.

WW

Examples of compliance instruments for the diabetic
patient were sought through a review of the literature.
Although several studies have been conducted to evaluate
compliance behavior of the diabetic (Tagliacozzo, 1974:

Williams, 1967: Watkins, 1967: Hulka, 1975), few examples

120

of instruments were available and they were not applicable.
Consequently, the researcher developed the questionnaire
based on selected clinical prescriptions for the adult onset
diabetic.

The dimensions of the clinical prescription that
were considered for this study were compliance with diet and
medication (insulin or oral hypoglycemics) within the past
two weeks, and compliance with follow-up appointments during
the last six months. Questions concerning other aspects of
the regimen such as urine testing, exercise, and general
hygiene including foot care were excluded. The reason for
exclusion is that patients are not usually given a written
prescription regarding urine testing, exercise, and general
hygiene. These aspects of the regimen are frequently not
recommended and/or taught with equal emphasis by all health
care providers.

Refinement of the questionnaire followed the
evaluation of the questionnaire by thesis committee members.
Statements were changed to increase the clarity of some
items. Changes in the arrangement of items on the
questionnaire were also made so a definite pattern of
positive and negative statements would not be apparent. In
addition, the questions concerning follow-up appointments
were separated from the questions concerning diet and
medication because of the difference in time frame to be

considered.

121

The perceived compliance behavior section of the
questionnaire contained eight items concerning each
dimension of the therapeutic regimen (diet, medication, and
follow-up appointments) with a total of twenty-four
questions. The questions were stated positively and
negatively and were scored according to a five point Likert
scale.

Each subject was asked to select the most
appropriate category representing the extent to which he/she
had perceived herself/himself as complying with what the
statement indicated. The subject was to consider the past
two weeks for diet and medication statements and the past
six months for follow-up appointment statements. The
choices included the following: never, seldom, half of the
time, most of the time, all of the time. These categories
were selected to indicate the degree of perceived
compliance. Negative statements (Items 1, 2, 3, 6, 7, 8,
10, 11, 15, 17, 19, 22, 23 Appendix C) were rated as never -
5 points, seldom - 4 points, half of the time - 3 points,
most of the time - 2 points, all of the time - 1 point.
Positive statements (Items 4, 5, 9, 12, 13, 14, 16, 18, 20,
21, 24 Appendix C) were scored in the reverse. For example,
the positive statement, "I have taken my medicine in the
amounts ordered”, if answered ”all of the time" was scored
five points. In contrast, the negative statement, "I have

eaten snacks that are not on my diet", when answered "all of

122

the time" was scored one point. Again, a high score would
indicate high perceived compliance behavior. Perceived
compliance behavior was measured by the total score and the
total score was used to ascertain if there was a
statistically significant relationship between perceived
family functioning and perceived compliance behavior.
However, the raw scores of each dimension (diet, medication,
follow-up appointments) were viewed independently as well.
According to Marston (1970), a more precise answer can be
given regarding compliance behavior by looking at the

dimensions individually.

E ]' I'J'l

The reliability of an instrument is the degree of
consistency with which it measures the attribute it is
supposed to be measuring (Ary et al., 1972). The procedure
used to estimate reliability for this study was the
coefficient alpha. The coefficient alpha is a mathematical
method which determines how all items on the questionnaire
relate to all other items as well as how the items relate to
the questionnaire as a whole. This refers to the internal
consistency or inter-item relatedness.

Crane and Brewer (1973) state coefficient alpha
probably represents the best estimate of internal
consistency because it is free from all the other influences
on reliability that are common to other means of testing

reliability. For example, the test-retest method

123

necessitates access to the same subjects at a later date.
The time lag between the test-retest method may result in
subjects answering the questions differently due to other
intervening events that may have occurred during the time
lag. The intervening event may be a death, a crisis within
the family, or other event that may have changed the
individual's attitudes, behaviors, moods, or knowledge. The
low correlation that may be obtained may be due to the
change that has taken place in the individual rather than
the unreliability of the test. Another test of reliability
is the split-half _method. However, the split-half method
tends to underestimate the reliability of the entire
questionnaire since longer questionnaires are more reliable
than shorter ones (Crano and Brewer, 1973).

For this study it is necessary that a high degree of
interrelatedness exists among all items so the items can be
summed together to produce an overall score that is a
meaningful measurement of the total concept of perceived
family functioning or perceived compliance behavior. The
total items in the family functioning section of the
questionnaire will be evaluated as well as the total items
in the compliance section. If the coefficient alpha is .80
or above then it is a satisfactory criterion of internal
consistency and fulfills the basic requirement for
operationalization of the concept (Crano and Brewer, 1973).

If the scale does not meet this criterion of .80, the

124

researcher must identify the items with a low and/or

negative correlation and delete them from the pool of items.

1111511121:

The only level of validity possible to this study is
content validity. Content validity refers to the degree to
which an instrument samples the content area which is to be
measured (Ary et al., 1972). That is, how well do the items
in the questionnaire represent the multitude of all items
that might be asked? Content validity is based on judgment
of experts. The researcher reviewed the family functioning
and compliance literature to obtain material feasible for
item construction. Following the development of the items
by the researcher, the items were evaluated based upon the
expert judgment of thesis committee members. Further
refinement of the instrument occurred following the
evaluation by thesis committee members. Changes were made
in wording and arrangement of items on the questionnaire.
The instrument was then assumed to be reflective of content
validity.

Two threats to validity were considered when
constructing the questionnaire. One threat to validity is
acquiescence. That is, some individuals tend to agree with
positive statements. To minimize the threat to validity
both positive and negative statements were included. The

second threat to validity is social desirability. That is,

125

some individuals tend to answer questions in the way they
think the researcher expects they should. To decrease this
threat to validity, the researcher asked the subjects to be
as honest as possible and assured them of anonymity.

In summary, this section has included a discussion
of the instrument in terms of its development and scoring,
reliability and validity. In the next section the research
question, hypothesis, subresearch questions, and techniques

used for data analysis will be presented.
;--‘; 9 07-: 0. $190 o‘:: ... o ‘t‘. o O -: o.‘

The research question was stated as follows:

Is there a relationship between the perceived
family functioning and the perceived compliance
behavior with selected aspects of the prescribed
therapeutic regimen of ambulatory adult onset
diabetic patients?

The following hypothesis was tested in this study:

There is a relationship between the perceived
family functioning and the perceived compliance
behavior with selected aspects of the prescribed
therapeutic regimen of ambulatory adult onset
diabetic patients.

Two subresearch questions were asked. They are as follows:

1. How does the ambulatory adult onset diabetic
perceive her/his family functioning?

2. How does the ambulatory adult onset diabetic

perceive her/his compliance with the
therapeutic regimen?

126

These questions were analyzed according to
statistical techniques suitable for descriptive data.

The research questions were answered by analyzing
the raw scores of perceived family functioning and perceived
compliance behavior using the Pearson Product-Moment
Correlation coefficient (Pearson r). The Pearson r is the,
most commonly used correlational index for descriptive
studies. The Pearson r enables the researcher to make an
association but does not necessarily indicate a causal
relationship.

If the two variables, perceived family functioning
and perceived compliance behavior, were highly related in a
positive way the correlation between them would approach +1.
If they were highly related in a negative way the
correlation would approach -1. If there was little relation
between them the correlation would be near zero. The
difference between the coefficient and zero indicated the
degree of relationship. The interpretation of r computed
between the variables in this study was:

1. r from 0.00 to 0.15 or 0.20 represented
negligible, or if close to 0.20, very slight
relationship between the variables.

2. r from 0.20 to 0.40 represented low correlation
present, but slight.

3. r from 0.40 to 0.60 represented moderate or fair

correlation.

127

4. L from 0.60 to 0.80 represented marked, somewhat

high relationship.

5. ,1 from 0.80 to 1.00 represented high to very

high relationship (vanOrmer and Williams 1949,
p. 65).

When a correlation coefficient is derived, the
statistical significance indicates whether or not the
coeefficient obtained is different from zero at a given
level of confidence. A statisically significant correlation
represents evidence of an actual relationship rather than
one due simply to chance. The level of significance for
this study was set at .05. The .05 level is the level of
significance commonly used when product moment correlation
is used and when the sample size is small.

The subresearch questions were answered according to
descriptive statistics also. The range and median were
calculated for each section of the questionnaire to answer
the questions.

In summary, in this section the research question,
hypothesis, subresearch questions, and techniques used for

data analysis were presented.

SummarY

In Chapter IV an overview of the methodology and
procedures involved in the study was presented. The
discussion included a presentation of the operational

definitions of the variables, sample, source of sample, data

128

collection procedures, human right protection, instrument,
and hypothesis and research questions. The data will be

presented in Chapter V.

129

Chapter V

DATA PRESENTATION AND ANALYSIS

Qxemiew

The data presented in this chapter describe the
study population and answer the following research question:
Is there a relationship between the perceived
family functioning and the perceived compliance
with selected aspects of the prescribed
therapeutic regimen of ambulatory adult onset

diabetic patients?
In addition, data are presented to explain the relationship
between the confounding variables (i.e., number of years
with diabetes, number of chronic illnesses other than
diabetes, complexity of regimen, and hospitalization within
the last six months) and their effect on compliance
behavior. Also included in this chapter is an analysis of
the reliability of the perceived family functioning and

compliance questionnaire. The hypothesis and the

subresearch questions that are presented and analyzed are:

mama
There is a relationship between the perceived
family functioning and the perceived compliance
with selected aspects of the prescribed

therapeutic regimen of ambulatory adult onset
diabetic patients.

130

Wartime

1. How does the ambulatory adult onset diabetic
perceive her/his family functioning?

2. How does the ambulatory adult onset diabetic

perceive her/his compliance with the
therapeutic regimen?

D . l' E' i' E I] 5| 1 S J

The sample was drawn from the population of diabetic
patients who sought care from four private physicians in
Flint, Michigan. The study sample consisted of 25
ambulatory adult onset diabetics who ranged in age from 36
to 70 years old with a mean age of 52.5 years. Nine of the
subjects (36%) were 36 to 47 years of age, nine (36%) were
48 to 59, and seven (28%) were 60 to 70 years of age.
Thirteen males (52%) and 12 females (48%) participated in
this study. Almost all were white (96%): only one subject
(4%) was black. The annual family income of subjects ranged
from less than $10,000 to greater than $20,000 per year.
Two subjects (8%) had incomes of less than $10,000, seven
(28%) had incomes of $10,000 to $19,999, and 16 (64%) had
incomes of $20,000 or greater. Data on age sex, race, and
income are summarized in Table 1.

Three (12%) subjects in the study were single,
twenty-one (84%) were married and one (4%) was separated.
The number of people with whom the subjects lived ranged
from one to seven. A majority (56%) lived with one other

person only. Approximately half (48%) of the subjects had

131

TABLE 1. Sociodemographic Data by Age, Sex, Race. and
Income of Participants (n = 25).

 

 

 

 

 

 

 

 

 

Number of
Participants Percent

Age
36 to 47 ‘ 9 36
48 to 59 9 36
60 to 70 7 28
Total ‘33" 100

Sex
Male 13 52
Female 12 48
Total T 100

Race
Black 1 4
White 24 96
Total '35— 100
Income

Less than 10,000/year 2 8
$10,000 - $19,999/year 7 28
$20,000 or greater/year 16 64
Total _2?’ 100

 

children living with them. The number of children in the
families ranged from one to six. These children's ages

ranged from two to thirty-two years of age with a mean of

132

sixteen years and a median of eighteen years. All but one
of these subjects with children was living within a nuclear
unit. The one exception was a woman who was separated from
her husband and who lived with her child. Thus, the
composition of the household varied among the 25

participants. See Table 2 for the summary of these data.

TABLE 2. Living Arrangements of Participants (n = 25).

 

 

 

Number of

Living Arrangements Participants Percent
Live with spouse 10 40
Live with spouse and 11 44

unmarried children
Live with either sibling,

cousin, or child 4 16

Total 25 100

 

The number of years the participants had diabetes
ranged from one to more than twenty years. Eleven subjects
(44%) had diabetes one to five years, nine (36%) from six to
ten years, two (8%) from eleven to fifteen years, one (4%)
from sixteen to twenty years, and two (8%) subjects had
diabetes more than twenty years.

To be a participant in the sthdy the subjects had
to have had either an oral hypoglycemic or insulin
prescribed for their diabetes. Fifteen (60%) of the
subjects were taking insulin and ten (40%) were taking oral

hypoglycemics.

133

One-fifth (20%) of the patients had diabetes only
(see Table 3). In addition to diabetes, of the twenty
patients who had other chronic diseases, twelve (60%) had
one other illness, five (25%) had two other illnesses, and
three (15%) had three other illnesses. The most frequent
other chronic illness mentioned by the twenty subjects was
hypertension (60%). This disease was followed in frequency
by heart problems (16%) and arthritis (16%). In total, the
twenty subjects with other chronic diseases mentioned eight
maladies. These data are summarized in Table 4.

As might be expected then, the complexity of the
regimen that had been prescribed for the patients varied.
For example, seventeen (68%) of the subjects were taking
medications in addition to those prescribed for their
diabetes. The number of these other medications ranged from
one to five (Table 5). Seven of the seventeen patients
(41%) were on one other medication, three (18%) were on two
other medications, three (18%) were on three other
medications, three (18%) were on four other medication, and
one (6%) was on five other medications.

The number of times per day the subjects were
taking medication ranged from one time to four times. Seven
(28%) of the subjects were taking medication once a day, ten
(40%) two times per day, four (16%) three times per day, and
four (16%) four times per day (Table 6).

134

Sixteen (64%) of the subjects indicated that no
other diet restrictions had been prescribed for them other

than their diabetic diet (Table 7).

TABLE 3. Number of Chronic Illnesses of Participants

 

 

(n = 25).
Number of
Chronic Illnesses Number of
Besides Diabetes Participants Percent
0 5 20
1 12 48
2 5 20
3 3 12
Total -25— —100'

 

TABLE 4. Types of other Chronic Illnesses of Participants.

 

 

Number of
Illnesses Participants Percent
Hypertension 15 60
Heart Problems 4 16
Peripheral Vascular Disease 2 8
Arthritis 4 16
Asthma 1 4
Kidney Disease 1 4
Thyroid Disease 2 8
Female Problems 2 3

 

135

TABLE 5. Complexity of Regimen Measured by Number of
Medications Other than Medication for Diabetes.

 

 

Number of Number of

Other Medications Participants Percent

0 8 32

1 7 28

2 3 12

3 3 12

4 3 l2

5 1 4
Total -25_ 100—

 

TABLE 6. Complexity of Regimen Measured by Number of
Times per Day Taking Medication

 

 

Number of
Times per Day Number of
Taking Medication Participants Percent

1 7 28
2 10 40
3 4 16
4 4 16

Total 25 100

 

136

TABLE 7. Complexity of Regimen Measured by Number of
Diet Restrictions Other than the Diabetic Diet.

 

 

 

Number of - Number of
Other Diet Restrictions Participants Percent
0 16 64
l 7 28
2 2 8
Total ‘33— 100

 

Among the nine subjects who had been told to follow a diet
in addition to that prescribed for their diabetes, six
subjects (67%) were on low salt diets, two (22%) were on low
salt and low cholesterol diets, and one (11%) was on a low
cholesterol diet.

Besides additional medications and diet
restrictions for the other chronic illnesses, only three
subjects (12%) mentioned that they had been told to exercise
for their heart problem. Twenty-two (88%) mentioned no
treatments other than what was prescribed for their
diabetes. I

Only one (4%) of the 25 subjects was hospitalized
for diabetes in the last six months and had had diabetic
teaching. Subjects who were hospitalized for other reasons
in the last six months denied having had any diabetic

teaching while hospitalized.

137

B ]' I'J'l E I | |

To determine the inter-item relatedness of items
within the questionnaire, the reliability of the instrument
was measured by computing coefficient alpha. Two
coefficient alphas were calculated: one for the percelved
family functioning section of the questionnaire and one for
the perceived compliance behavior section of the
questionnaire.

In the perceived family functioning section of the
questionnaire, items F1, F7, F16, F19 were found to have a
low inter-item correlation, i.e., below .20, and were
deleted from the pool of items. Following the deletion of
these items, the twenty-two items remaining in the perceived
family functioning section (Appendix J) had a coefficient
alpha of .84, thus demonstrating internal consistency.

In the perceived compliance behavior section of
the questionnaire only two items had to be deleted. That
is, items C7 and C16 were found to have a correlation below
.20, and accordingly were deleted from the pool of items.
Following the deletion of these items, the 22 items
remaining in the perceived compliance behavior section
(Appendix J) had a coefficient alpha of .88, demonstrating
internal consistency.

Each of the dimensions of the therapeutic regimen
(i.e., diet, medication, and follow-up appointments) was

subjected to individual coefficient alpha analysis to

138

evaluate the interrelatedness between the items contained
within them. ‘

The reliability analysis to determine internal
consistency for items concerning diet resulted in the
deletion of items C4, C7, C15 since an inter-item
correlation below .20 was found. Items C1, C3, C6, C9, C12,
remained in the pool with a coefficient alpha of .86 which
- demonstrated internal consistency among the five items for
diet.

Following the reliability analysis on the
medication items, items C2 and C13 were deleted from the
pool because of a low inter-item correlation, i.e., below
.20. Items C5, C8, C10, C11, C14, C16 remained in the pool
with a coefficient alpha of .73. This coefficient alpha
demonstrated internal consistency among the items for
medications.

The reliability analysis to determine internal
consistency for items concerning follow-up appointments
resulted in the deletion of items C19 and C20 from the pool
of items due to a low inter-item correlation, i.e., below
.20. Items C17, C18, C21, C22, C23, C24 remained in the
pool with a coefficient alpha of .93. This statistic
demonstrated internal consistency among the items for

follow-up appointments.

139

D l E | l' E H I] .
Ins_fitatistisal_Technisns

This study was a descriptive study and therefore was
analyzed according to descriptive statistics. Correlation
is the statistical technique used for measuring the degree
of relationship between two variables which results in a
value that ranges from -1 to +1. The Pearson Product-Moment
Correlation (also called Pearson r) is the most commonly
used correlational index for descriptive studies (Ary et
al., 1972). After items that correlated below .20 were
omitted, the Pearson r was used to . determine the
relationship between the raw scores of perceived compliance
behavior and perceived family functioning. For this study,
to accept the hypothesis, the level of confidence was set at
.05. The .05 level is the level of significance commonly

used when product moment correlation is used and when the

sample size is small.

Hypothesis
There is a relationship between the perceived
family functioning and the perceived compliance
with selected aspects of the prescribed
therapeutic regimen of ambulatory adult onset
diabetic patients.
Perceived family functioning and percelved
(compliance behavior were measured by statements contained

vvithin a self-administered questionnaire. Responses to

statements were recorded on a five point Likert scale. Some

140

statements were stated positively and some were stated
negatively to prevent the development of a response
situation.

To test the hypothesis a Pearson Product-Moment
Correlation was performed. First, a numerical value was
assigned to each response that indicated the degree of
agreement or disagreement with the individual statements
contained in perceived family functioning section and
perceived compliance behavior section of the questionnaire.
Then the values of the items measuring perceived family
functioning and the items measuring perceived compliance
behavior were summed separately. That is, each subject was
assigned two scores: one score represented the value of
summed perceived family functioning responses and one score
represented the value of summed perceived compliance
behavior responses. Finally, the sum of the raw scores of
perceived family functioning was correlated with the sum of
the raw scores of perceived compliance behavior.

The Pearson Product-Moment Correlation between the
perceived family functioning score and the perceived
compliance behavior score was r = .10 and not significant at
the .05 level. Thus, there was no relationship between
perceived family functioning and perceived compliance

behavior.

141

W

The data relevant to the subresearch questions also
were subjected to descriptive statistic techniques. That
is, the range, mean, and median were calculated for the

variables dealt within the questionnaire.

Questign_l: How does the ambulatory adult onset diabetic
perceive her/his family functioning?

The scores of the perceived total family functioning
section showed a range of 73 to 102 with a maximum score
.possible of 110. The mean for the scores was 87.60, with a
standard deviation of 8.26. The median was 87.00. The
average score for each question was 3.98 with a range of 1
to 5. The subjects in this sample had a positive view of

their family functioning.

Qgggtign_z: How does the ambulatory adult onset diabetic
perceive her/his compliance behavior?

The perceived compliance behavior scores showed a
range of 76 to 105 with a possible score of 110. The mean
for the scores was 96.12 with a standard deviation of 7.75
and a median of 96.00. The average score for each question
was 4.36 with a range of 1 to 5. The subjects perceived
themselves as complying with their therapeutic regimen more

than 'most of the time".

142

A more precise answer can be given to the
subresearch question regarding compliance behavior by
looking at the dimensions individually.

The scores on the perceived compliance behavior for
diet showed a range of 13 to 25 with a possible of 25. The
mean was 17.64 with a standard deviation of 3.20. The
median was 18.25. The average score for each question was
3.52 with a range of 1 to 5.

The perceived compliance behavior for medications
scores ranged from 24 to 30 with a possible score of 30 and
a mean of 28.64 with a standard deviation of 1.75. The
median was 29.18. The average score for each question was
4.77 with a range of 1 to 5.

The scores on the perceived compliance behavior for
follow-up appointments ranged from 19 to 30 with a possible
of 30 and a mean of 28.60. The average score for each
question was 4.76 with a range of l to 5. The raw scores
indicated that compliance with follow-up appointments and

medication were somewhat higher than compliance with diet.

g]. . J E l

The clinical parameters collected for this study
were (1) height, (2) weight, and (3) the most recent random
blood sugar within the month the questionnaire was
completed. The relationship between these variables and

perceived family functioning and perceived ' compliance

143

behavior was analyzed using Pearson Product-Moment
Correlations.

The evaluation of weight according to height and sex
showed that the weight of only two patients (8%) was within
normal limits. Four patients (16%) were less than 20 pounds
overweight, eight (32%) were 20 to 50 pounds overweight, six
(24%) were 51 to 100 pounds overweight, and five (20%) were
more than 100 pounds overweight. The degree of obesity was
not significantly related to perceived family functioning (r
= .16) or to perceived compliance behavior (r = -.26). The
low negative correlation between weight and perceived
compliance behavior implies a reversed relationship. That
is, the more overweight the patient, the lower the perceived
compliance behavior score.

An examination of the random blood sugar levels
drawn within the month the questionnaire was completed
showed that seven levels (28%) were within normal limits
(70-120 mg/ml), two (8%) were within an acceptable range
(60-69 mg/ml and 121-150 mg /m1), and sixteen (64%) were in
a range indicating lack of control (<60 or >150mg/ml). The
blood sugar levels of the subjects were inversely related to
perceived compliance behavior (r = -.01) and not significant
at the .05 level. The blood sugar levels of the patients
also were inversely related to perceived family functioning
(r = -.32) and not significant at the .05 level. The

negative correlations indicate that the higher the

144

individual's random blood sugar level the lower the
individual's perceived compliance behavior and perceived
family functioning scores.

In summary, the majority (92%) of the subjects in
this sample were overweight in relation to height and sex
and the majority (64%) had random sugar levels that
indicated lack of control of diabetes. In addition, a
significant relationship was not found between perceived
family functioning or perceived compliance behavior and

weight or blood sugar level.

C E i' M . I]

The relationship between the four confounding
variables and perceived family functioning and perceived
compliance behavior was analyzed using Pearson
Product-Moment Correlation. The four confounding variables
examined were: (1) number of years with diabetes; (2)
number of other chronic illnesses; (3) complexity of the
regimen (i.e., number of other medications, number of times
per day taking medication, number of other diet
restrictions, and number of other treatments); and (4)

hospitalization within the last six months.

W
The number of years the participants had diabetes

was neither related to perceived compliance behavior (r =

.26) nor to perceived family functioning ( r = -.l6) at a

145

.05 level of significance. The negative correlation derived
shows that as the number of years with diabetes increases,

the score for perceived family functioning decreases.

W112

The number of other chronic illness the subjects had
was related to perceived compliance behavior (r = -.34) at
the .05 level of significance. The negative relationship
between the two variables indicates that as the number of
chronic illnesses increases the score of the perceived
compliance behavior decreases. No significant relationship
was found between the number of other chronic illnesses and

perceived family functioning (r = .25).

W
The number of medications the patient was taking was

not significantly related to perceived compliance behavior,

I = -0110

g E 3' K . I] 3|
The number of times per day the patient was taking

medication was not significantly related to perceived

compliance behavior, r = .12.

W
The number of diet restrictions was not

significantly related to perceived compliance behavior, r =

.16.

146

C E i' H . I] 31
The number of additional treatments for other

chronic illnesses was not significantly related to percelved

compliance behavior, r = .12.

C E 3' H . I] I

Only one subject in the study had been hospitalized
for diabetes in the six months prior to the study. Thus,
insufficient data were available to compute a correlation
coefficient.

In summary, the number of years with diabetes and
the complexity of the regimen were not significantly related
to perceived compliance behavior in this study. There was a
low correlation between the number of other chronic
illnesses and perceived compliance behavior, significant at

the .05 level.

Summary
In Chapter V data were presented that described the

study population. Reliability analyses were performed on
the perceived family functioning section and the perceived
compliance behavior of the questionnaire utilizing
coefficient alpha. Data were presented and the result of
the Pearson Product-Moment Correlation analysis for the
research hypothesis to test the relationship between
perceived family functioning and perceived compliance

behavior was described. Descriptive statistics (range,

147

mean, and median) were used to answer the subresearch
questions. The Pearson r was also used to analyze a
correlation of the clinical parameters with perceived family
functioning and perceived compliance behavior. In addition,
the Pearson r was also used to analyze a correlation of the
confounding variables with perceived compliance behavior.
Selected confounding variables were correlated with
perceived family functioning scores as well. In chapter VI
the summary, interpretation, and implications of these

findings will be presented.

148

CHAPTER VI

SUMMARY OF FINDINGS, INTERPRETATION,
AND IMPLICATIONS

Dismiss:

In this chapter a summary and interpretation of the
research findings are discussed. Implications for nursing

practice, education, and recommendations for future research

are offered.

5 i I l I Ii E E’ i'
D . I. E' i' E I] SI 1 E J l'

The demographics of this study population include
age, sex, race, and income.. A criterion for participation
in this study was that subjects must be at least 31 years of
fage. The sample ranged from 36 to 70 years old with a mean
age of 52.5 years. Nine (36%) of the subjects were between
thirty-six and forty-seven years of age, nine (36%) were
between forty-eight and fifty-nine years of age, and seven
(28%) were sixty to seventy years of age. Although
incidence of diabetes increases with age (Blevins, 1979:
Sayetta & Murphy, 1979), there was a fairly even
distribution of ages of subjects in this study. Thls may be
due to the age limitation stated in the criteria established
for inclusion in this study.

Thirteen (52%) of the twenty-five subjects were male

149

and twelve (48%) were female. The majority (24) of the
subjects were white; only one was black. The reason for
this distribution, no doubt, is that the sample was drawn
from the patient population of four white physicians whose
clientele is predominantly white.

The source of the subjects, undoubtly, also explains
the large number of middle-class and upper middle-class
patients in the sample. That is, only two subjects (8%) had
an income of less than $10,000. Sixteen (64%) had an income
of greater than $20,000 per year.

According to Blevins (1979), the incidence and
prevalence of diabetes are related to factors of age, sex,
race, and income. That is, those people at high risk to
develop diabetes are the aged, women, the poor, and the
nonwhite. Data from the National Center for Health
Statistics, which show that diabetes increases with
advancing age and occurs more frequently in females and in
blacks (Sayetta &7 Murphy, 1979), support Blevins' (1979)
statement. Thus, the sample for this study was not typical
relative to the incidence of diabetes by age, sex, race, and.
income. The sample's atypicality most probably is due to
the fact that it was drawn from a population of patients who
sought care at private physicians' practices.

Ninety-two percent of the subjects in this study
were overweight. In addition, the majority (64%) of the

subjects had random blood sugar values that indicated that

150

their diabetes was out of control.

A primary concern of this study is family
functioning and, therefore, to be included in the samples
the participants had to be living with at least one other
person. Although traditional definitions of the family
focus on the nuclear unit, such a restricted definition was
not used in this study. Instead, Rakel's (1977) and
Smilkstein's (1978) definition of the family, which
encompasses people living in non-traditional units, was
accepted. According to Rakel (1977) and Smilkstein (1978),
a family can be viewed as two or more people who have a
commitment to nurture each other and who have a significant
effect on one another. Thus, the subjects in this study
lived in families that were conjugal as well as
non-conjugal. For example, twenty-one of the subjects (84%)
lived with their spouse and eleven (52%) of these subjects
still had children who lived at home. By contrast, one
subject was separated and lived with her eight year old son
and three subjects were single. Two of these non-married
subjects lived with a sibling and one with a cousin.

In summary, the subjects in this study were not
typical relative to the incidence of diabetes by age, sex,
race, or income. Moreover, the majority of subjects were
overweight and their random blood sugar indicated their
diabetes was out of control. Finally, all but three of the

subjects lived in a nuclear family.

151

W
There is a relationship between the perceived
family functioning and the percelved compliance
with selected aspects of the prescribed
therapeutic regimen of ambulatory adult onset
diabetic patients.

The relationship between perceived family
functioning and perceived compliance behavior was low ( r =
.10) and not statistically significant at the .05 level.
Thus, there is no relationship between perceived family
functioning and perceived compliance behavior. In other
words, adult onset diabetics who perceive their families as
(l) aiding each other in the time of need (adaptation), (2)
communicating with each other (partnership): (3) encouraging
independence (growth), (4) encouraging emotional expression
(affection), and (5) willing to share time, space, and money
(resolve), do not perceive themselves as complying with
their diabetic regimen concerning their diet, medications,
and follow-up appointments.

Nevertheless, although evidence is lacking Ito
support a relationship between perceived family functioning
and perceived compliance behavior, it still cannot be said
with certainty that no relationship exists between these two
variables. Several factors may have influenced the
findings. For example, the sample size was small and a
convenience sample from limited geographic areas and limited

sources (i.e., the population of patients seeking care from

private physicians). Therefore, the sample cannot be

152

considered representative of the entire population of
diabetic patients. In addition, the sample was not typical
of the adult onset diabetic in terms of age, sex, race, or
income. That is, most of the subjects were white (96%) and
had an income of greater than $20,000 per year (64%). These
factors then, may have influenced the results of this study.

Yet another factor that may have influenced the
findings concerns viewing perceived family functioning in
its totality rather than by individual components. Based on
a review of family literature, the researcher believes the
family functioning section of the instrument encompassed the
major components of family functioning. The study, however,
was not designed to investigate the individual components of
perceived family functioning. That is, the family
functioning section of the questionnaire was not long enough
to evaluate effectively each component separately. If each
component was investigated separately, a relationship may
have been found between some of the individual components of
perceived family functioning and perceived compliance
behavior even though it was not found for the total concept.
For example, there may have been a relationship between
' perceived adaptation and perceived compliance behavior.
Future research is needed to investigate further both the
total concept of perceived family functioning and the
individual components of perceived family functioning as

related to perceived compliance behavior.

153

In summary, the research hypothesis was rejected due
to the nonsignificant finding. To substantiate this
finding, further study would be necessary using a larger
sample size and a different type of sample (i.e., clinic
patients versus patients of private physicians). There is
need for future research to investigate the components of

family functioning separately as well as in totality.

thrsssarsh_flnsstions
Question 1: How does the ambulatory adult onset
diabetic perceive her/his family
functioning?

A descriptive analysis of the family functioning
section of the questionnaire was used to explore the degree
to which patients felt their families: (1) gave support in
time of need (adaptation): (2) communicated and made
decisions (partnership): (3) accepted change and promoted
independence (growth): (4) accepted and expressed emotions
(affection); and (5) shared time, space, and money
(resolve). The scores for perceived family functioning
ranged from 73 to 102 with a possible score of 110, thus
showing little variation in response among subjects. The
mean score of the total index was 87.60. The average score
for the questions in the perceived family functioning
section of the questionnaire was 3.98 with the greatest
possible score being 5. The scores indicated that in this
sample, on the average, the subjects' perception of family

functioning was positive. No studies reviewed reported how

154

a person perceives her/his family functioning, thus a
comparison of results cannot be made.

The coefficient alpha for the total percelved family
functioning section of the questionnaire was .84 and showed
interrelatedness of the items on the total perceived family
functioning section of the questionnaire. This finding,
however, does not mean that the subjects percelved each of
the components equally. That is, each of the components of
family functioning (adaptation, partnership, growth,
affection, and resolve) were not viewed separately, and
therefore, it cannot be implied that the subjects perceive
each of the components equally.

For example, subjects may be positive in their
perception of the family's ability to give each other
support in times of need, in their ability to communicate
and make decisions, and in their ability to accept change
and to promote independence: however, they may not feel as
positive about the family's ability to give and receive
affection, and/or their ability to share time, space, and
money with each other. Thus, in an effort to ascertain how
the subjects in this sample viewed the separate components
of family functioning, the raw scores were analyzed. That
is, an estimate of how, on the average, each component was
viewed by the subjects was obtained. With a possible score
of S, the average score for the questions were: adaptation

3.64, partnership 3.89, growth 3.95, affection 4.01, and

155

resolve 4.02. The variance between the components of family
functioning was only slight. For example, the subjects in
this sample were almost as positive in their perception of
the family's ability to give support in times of need as
they were in their perception of their family's ability to
share time, space, and money.

In summary, the subjects in this study, on the
average, veiwed their family functioning positively.
Examination of the raw scores for each component of family
functioning showed, however, only slight variance between
the separate components of family functioning. Further
research is necessary to study the interrelationship of each

component (adaptation, partnership, growth, affeCtion, and

resolve) in more detail.

Question 2: How does the ambulatory adult onset
diabetic perceive her/his compliance
with the therapeutic regimen?

The subjects indicated, via the perceived compliance
section of the questionnaire, the degree to which they felt
they complied with the diet, medications, and follow-up
appointments for diabetes prescribed by their health care
providers. The scores of perceived compliance behavior
ranged from 75 to 105 with a possible score of 110, again
showing little variation in response among subjects. The
mean score for perceived compliance behavior was 96.12. The

average score for the questions on the whole was 4.36.

Concerning the positively stated items "all the time” was

156

scored as 5, "most of the time“ as 4, “half of the time" as
3, "seldom" as 2, and ”never“ as l. Negatively stated items
were scored in the reverse. The scores suggest that the
subjects in this sample, on the average, perceive themselves
to be following their diet, taking their medication, and
keeping their follow-up appointments more than "most of the
time".

The coefficient alpha for the total compliance
behavior section of the questionnaire was .88 and showed
interrelatedness of items. However, because of the
possibility that the compliance score might be somewhat high
due to loss of precision by combining the compliance scores
concerning diet, medication, and follow-up appointments,
each dimension was viewed separately.

Perceived compliance scores for follow-up
appointments showed a range of 19 to 30 with a possible of
30. The mean for the group of subjects was 28.60. The
average score for the questions was 4.76, suggesting that
the subjects perceived themselves as complying with
follow-up appointments and having their blood sugar drawn
almost ”all of the time". An audit of the clinical records
to ascertain number of appointments the patients missed in
the 6 months prior to completing the questionnaire,
supported this finding. That is, none of the subjects had
missed appointments during the specified six-month period.

In addition, all of the subjects had had a blood sugar drawn

157

prior to or at the time of each follow-up appointment for
diabetes. It appears, then, that self-reported compliance
concerning keeping follow-up appointments was consistent
with actual behavior. Moreover, the findings of this study
support previous research which has shown that patients who
experience continuity of care and receive care from private
physicians tend to keep follow-up appointments (Haynes et
al., 1979).

Perceived compliance behavior for taking medications
ranged from 24 to 30 with a possible of 30. The mean score‘
was 28.64. The average score for each question concerning
medication was 4.47. This score suggests that the subjects
in this sample perceive themselves as complying with taking
medications for their diabetes midway between "most of. the
time" and "all of the time". An audit of the clinical
record to ascertain the patients random blood sugar within
the month the questionnaire was completed, did not support
this finding. That is, the subjects in this sample perceive
themselves as complying with taking medication very well.
Yet, the majority (64%) had random blood sugar levels that
indicated the patients' diabetes was not under control.
(For further discussion concerning how the patient's blood
sugar relates to perceived compliance behavior, see the
clinical parameters section below).

All 25 subjects in this study were on a calculated

diabetic diet according to their clinical records. However,

158

when the researcher was administering the questionnaire four
of the subjects commented that they did not count calories
or use the exchange list. Instead they indicated, they
limited the amount they ate and avoided sweets. Such
noncompliance was reflected in the scores derived on the
perceived compliance behavior for diet. That is, the scores
ranged from 13 to 25 with a possible of 25 and a mean score
of 18.25. The average score for the diet questions was
3.52, suggesting that, on the average, the subjects perceive
themselves as complying with their diet midway between "half
of the time“ and "most of the time".

The chart audit concerning weight, as related to
height and sex, indicated that the subjects in this sample
probably overestimated their compliance with following their
diet. That is, 92 percent of the subjects in this sample
were overweight and 44 percent were more than 50 pounds
overweight. It appears, then, that higher rates of
compliance may have been reported than if compliance was
based on more objective measures such as weight. This is
consistent with Haynes and his associates' (1979) findings,
i.e., higher rates of compliance are reported when based on
self-report by patients than when based on more objective
measures such as clinical outcomes. (For further discussion
concerning the relationship between weight and self-reported
compliance with diet, see the clinical parameter section

below).

159

Although the findings in this study concerning
perceived compliance with diet suggest that patients may
have overestimated their compliance, the findings do show
that the subjects do not perceive themselves as complying
with their diet as well as they perceived themselves
complying with_ taking medication or keeping follow-up
appointments. This finding supports other research in which
it was found that it is more difficult to comply with
restrictions on behavior or changes in personal habits
(i.e., changes in dietary habits) than to comply with taking
medication (Davis, 1966). Although the overall compliance
scores were high, the fact that the subjects in this sample
did perceive themselves as complying with their diets
considerably less than they perceived themselves complying
with taking medication and keeping follow—up appointments,
might suggest a relationship between actual and perceived
compliance. That is, although the subjects may have
overestimated their compliance behavior, they may have
overestimated each dimension (i.e., diet, medication, and
follow-up appointments) to the same degree.

In summary, it was found that the subjects were
positive concerning perceived family functioning. Moreover,
the subjects perceived themselves as complying with their
therapeutic regimen concerning diet, medications, and
follow-up appointment more than "most of the time”.

However, when the three dimensions were examined

160

individually, differences were found in the subjects'
perceptions. That is, subjects perceived themselves as (l)
complying with their diet midway between 'half of the time”
and "most of the time”, (2) complying with the medication
regimen midway between ”most of the time” and 'all of the
time“, and (3) complying with follow-up appointments almost
"all of the time".

The chart audit supported some, but not all of these
patients' perceptions. That is, it appears that patients'
perceived compliance with follow-up appointments was
consistent with actual compliance. In contrast, it appears
that patients may have overestimated their compliance with
taking medications and compliance with their diabetic diet.
Nevertheless, although the patients did appear to
overestimate compliance, it appears they may have
overestimated each dimension to the same degree thus showing
a relationship between perceived compliance behavior and
actual compliance behavior. In the next section the
clinical parameters, blood sugar levels and weights, will be

discussed in relation to compliance behavior.

91' . J E l
Each subject had had a follow-up appointment for her/his
diabetes within the month the questionnaire was completed.
All clinical parameters, therefore, reflect patients' health
status very close to the time they were interviewed. A

comparison of the patients weight, height, and sex with the

161

Metropolitan Life Insurance height, weight chart showed that
92 percent of the subjects were overweight. . Of the
twenty—three patients who were overweight, four (18%) were
less than twenty pounds overweight, eight (34%) were twenty
to fifty pounds overweight, six (26%) were fifty to one
hundred pounds overweight, and five (22%) were more than one
hundred pounds overweight. Degree of obesity, however, was
not significantly related to perceived family functioning
( r = .16) or perceived compliance behavior ( r = -.25).
This finding is not surprising since a
characteristic of the adult onset diabetic is obesity.
Moreover, the adult onset diabetic who has her/his weight
under control usually does not need to be on an oral
hypoglycemic or insulin to control her/his blood sugar.
And, of course, one criterion for inclusion in the study was
taking an oral hypoglycemic or insulin. Given the fact that
92 percent of the subjects in the sample were overwelght,
the subjects' perception of their compliance with diet seems
an overestimation. This finding suggests that patients may
have difficulty recognizing noncompliance concerning diet.
Or, this finding may reflect the fact that patients answered
questions in a way they thought the researcher expected.
Random blood sugar levels also were examined to
explore a possible congruence between perceived compliance
and actual behavior, the assumption being made that patients

in control would more likely comply with treatment and have

162

a more positive perception of their compliance than patients
not in control. This assumption was not supported by the
findings. The blood sugar levels were not significantly
related to perceived compliance 'behavior (r = -.01).
Although the blood sugar levels of seven patients (28%) were
within normal limits (70-120 mg/dl) and those of two
additional patients (8%) were within an acceptable range
(60-150 mg/dl), sixteen (64%) of the patients were not in
control ( <60 or >150 mg/dl). A random blood sugar, of
course, gives only a fragmentary glimpse of the patient's
metabolic status (Bondy & Felig, 1971).. Yet, this one time
measure suggests that the majority (64%) of the subjects
were out of control, even though they felt they were
complying with a prescription whose intent was to help them
achieve control.

This finding concerning the lack of relationship
between blood sugar levels and perceived compliance
behavior, may suggest several things. First, the findings
may mean there is little or no relationship between
compliance with taking medication and staying on the
diabetic diet and the outcome measure, control of diabetes
as measured by blood sugar levels. This is consistent with
the findings of Williams and associates (1967) who found no
association between compliance, as measured by direct
observation and self-report, and control of diabetes. This

would imply that appropriate outcome measures to indicate

163

diabetic control are difficult to identify (Malone 1976).

Second, the subjects in this sample may have
overestimated their compliance with taking medication and
following their diet. For example, Haynes and associates
(1979) suggest self-reporting by patients usually results in
findings of higher rates of compliance. This may be so
because the patients may want to answer the questions in the
way they think the researcher expects them to answer the
questions. Or, the patients may be afraid to admit their
noncompliance for fear their physician may become aware of
their noncompliance.

Third, the findings regarding the chart audit may
mean that indeed there is no relationship between perceived
compliance behavior and actual compliance behavior. Further
study is necessary to verify the relationship between
perceived compliance and actual compliance behavior.

In viewing the blood sugar and weight of each
patient, it was noted the random blood sugar was not
directly related to the degree of obesity i.e., of the
subjects whose ramdom blood sugar was within normal limits
one was less than twenty pounds overweight, three were
twenty to fifty pounds overweight, one was fifty to one
hundred pounds overweight, and two were greater than one
hundred pounds overweight. This finding is difficult to
explain since there also was no relationship between the

degree of obesity and the amount of diabetic medication the

164

patient was taking. For example, the patient who was less
than 20 pounds overweight was taking 22 units of insulin
while the two patients that were more than 100 pounds
overweight were taking only oral hypoglycemics to aid in
maintaining diabetic control.

Another finding that is worth noting concerns blood
sugar levels and perceived family functioning. The
correlation between perceived family functioning and blood
sugar levels was r = -.32 approaching significance (p =
<.058). Had the sample size been larger the correlation
might have been higher and a significant relationship may
have been found. The negative correlation implies that the
higher the individual's blood sugar the lower the
individual's perceived family functioning score. The
inverse relationship between these two variables might be
explained by the fact that low perceived family functioning
may cause stress and/or overeating and stress and/or
overeating causes the blood sugar to rise (Guthrie . &
Guthrie, 1977).

The finding of an inverse relationship between
perceived family functioning and perceived compliance
behavior is consistent with the findings of Williams and
associates (1967), who, found in a study of insulin
dependent diabetics, that the greater the degree of stress,
the poorer the control of diabetes. Thus, there is a need

for further research using a more uniform means of obtaining

165

blood sugar i.e., fasting blood sugar rather than random
blood sugar. A significant relationship may have been found
had fasting blood sugar levels been evaluated.

This tendency toward an inverse relationship between
perceived family functioning and blood sugar levels has
implications for nursing i.e., a need exists to assess
perceived family functioning and to deal with the deficits.
Solving perceived family functioning problems may result in
improved diabetic control as indicated by the blood sugar
levels.

In summary, those who perceived family functioning
and/or compliance behavior as high were as likely to be
overweight as those who perceived their family functioning
and/or compliance behavior to be somewhat lower. In
addition, 92% of the subjects were overweight. Yet, they
reported they complied with their diet midway between " half
of the time“ and ”most of the time". Second, no significant
relationship was found between random blood sugar and
perceived compliance behavior. Moreover, 64 percent of the
patients had random blood sugar levels that indicated their
diabetes was not controlled. Third, there was no direct
relationship between random blood sugar and degree of
obesity. And finally, there was a tendency toward an
inverse relationship between random blood sugar and
perceived family functioning. These findings suggest to the

nurse the importance of assessing percelved family

166

functioning and dealing with any deficits. Furthermore, the
nurse should keep in 'mind that patients may tend to
overestimate compliance with their diet.

In the next section, other variables that may
influence perceived family functioning and percieved

compliance behavior will be discussed.

g E 2. fl . l]

The review of the literature indicated a host of
variables that vinfluence compliance behavior. For the
purposes of this study, four were considered as confounding
variables and their relationship to perceived compliance
behavior explored. These .four confounding variables
included: (1) the number of years with diabetes: (2) the
number of chronic illnesses other than diabetes; (3) the
complexity of the regimen: and (4) hospitalization in the
last six months.

Number.9f_¥sars_with_diahetes. Analysis of the data
_showed that the number of years the subject has had diabetes
was not significantly related at the .05 level to perceived
compliance behavior (r = .26). This finding supports the
results of other studies that show that the length of time
the patient had an illness is not associated with compliance
behavior (Haynes et al., 1979). .Yet, previous research also
has shown that the continuation of therapy over time is
negatively related to compliance behavior, i.e., the longer

the patient is on a treatment regimen, the less he/she

167

complies (Haynes et al., 1979). Thus, the assumption was
made in this study that the duration of the regimen was
equal to the number of years the subject had diabetes. This
assumption of equivalence was based on the fact that it is
necessary for diabetic patients to comply with a diet from
the time their diabetes is diagnosed. This means that since
the diabetic diet is basic to all diabetic regimens, it is
the first treatment prescribed to control diabetes.

The results showed, however, that this assumption
may not have been warranted. The lack of a significant
relationship between number of years with diabetes and
perceived compliance behavior, then, may be due to several
factors. First, the data concerning compliance behavior was
elicited by self-report. Thus, patients may have
overestimated their compliance and the degree of
overestimation may not be influenced by the number of years
the patient has had diabetes. Second, patient may not have
accurately reported the number of years they had diabetes.
That is, patients who have had diabetes for an extended
period of time may have forgotten the actual number of years
with diabetes and either underestimated or overestimated the
length of time, thus resulting in the lack of relationship
between the variables.

Analysis of the data also showed that the number of
years the subject had diabetes was not related to percieved

family functioning (r = -.l6). Nevertheless, although the

168

results were not significant at the .05 level, the analysis
showed an inverse relationship existed between the two
variables. That is, those subjects who had diabetes the
fewest number of years tended to perceive their family
functioning more positively than those who had had the
disease longer.

The reasons for this finding are difficult to
explain. For example, Rakel (1977) stated that illness of
one family member affects all other family members and can
affect the functioning of her/his family as well. Moreover,
he stated that the more prolonged and complicated the
illness, the greater its effect on the functioning of the
family. Since diabetes is a chronic and complex illness, it
might be expected that the presence of a diabetic in a
family would affect the functioning among its members. For
example, in diabetes the patient must eat certain foods in
specific amounts, and at specified times throughout her/his
life time. Since the alteration in diet and eating habits
must be permanent, there also may be a tendency to require
the rest of the family to eat the type of food and at the
times necessary for the diabetic patient. Thus, the family
may be willing at the outset to cooperate with these
changes, but as time goes on become less tolerant of the
need for change.

There are several possible reasons for the lack of a

significant relationship between perceived family

169

functioning and the number of years with diabetes. First,
the diabetic patient is not generally debilitated and the
not dependent on the rest of the family for their care.
Thus, the length of time the patient has had diabetes, since
diabetes in itself is not generally debilitating, may not
affect the functioning of the family. Second, the diet and
possibly the cost of the regimen may be the only aspect of
the regimen that affects the rest of the family. Thus, the
family may have adapted to the cost and changes in diet and
eating habits and therefore its functioning was not
affected. Or, possibly, the rest of the family members were
not required to make a change in their diets or eating
habits and the increased cost may not have affeCted the
family since most of the patients (64%) had adequate incomes
(i.e., greater than $20,000 per year). Third, to
participate in this study the subjects must have had
diabetes at least one year. Thus, the members of the
families may have perceived themselves as having adapted to
the situation and the disease may no longer be impacting the
family. Finally, the sample size was small. If the sample
size was larger, a relationship may have been found. Each
of these factors may have contributed to the lack of
significant relationship found.
WWW. It
was found that the number of other chronic illnesses

patients had was significantly related to perceived

170

compliance behavior (r = -.34, p < .05). The correlation
between the variables, however, was low, indicating only a
slight relationship. Moreover, it was found that the
relationship between number of chronic illnesses and
perceived compliance behavior was inverse. That is, as the
number of chronic illnesses increased, the score of the
perceived compliance behavior decreased. What this means is
that patients with multiple diseases tended to percieve
themselves as complying less than those who have fewer
chronic illnesses. This is reasonable since as the number
of chronic illnesses increases the complexity of the regimen
increases also. For example, patient with additional
chronic diseases usually have more medications prescribed to
be taken at varied times throughout the day and have
increased dietary restrictions.

No relationship was found, however, between the
number of other chronic illnesses patients had and perceived
family functioning (r = .25). This finding was surprising
since illnesses of a complex and debilitating nature have a
disruptive effect on family functioning (Rakel (1977) and an
increasing number of chronic illnesses may tend to make
illness, complex and, sometimes debilitating. It might have
been expected then, that an increasing number of chronic
illness would create problems within the family that were
reflected in patients' perceptions of their family

functioning.

171

There are several possible reasons why no
significant relationship was found between the number of
other chronic illnesses patients had and their perceptions
of the functioning of their families. First, the other
chronic illnesses they had may not have been debilitating.
The most frequently mentioned other chronic disease was
hypertension. Sixty percent of the subjects had
hypertension. Hypertension in itself is not debilitating
but may lead to other debilitating diseases such as cerebral
vascular accident, kidney disease, and cardiac disease.
Four subjects (16%) indicated they were taking medication
for a heart problem but mentioned no other treatments or
restrictions. Peripheral vascular disease and arthritis can
be debilitating and six subjects (24%) indicated having one
of these diseases. However, no subjects expressed being
incapacitated. An additional possibility for no significant
relationship between the number of other chronic illnesses
and perceived family functioning is that the family may .be
functioning to help or the family may have adjusted to the

chronic illnesses.

WW. Analysis of the data

also revealed that the number of medications the subjects
were taking was not significantly related to perceived
compliance behavior (r = -.ll). This finding was at
variance with previous research findings. For example,

Haynes and his colleagues (1979), in a review of the

172

literature, found consistent evidence which suggested that
multiple drugs reduced compliance behavior. In six of the
eleven studies they reviewed, compliance fell off sharply
when the number of medications reached three or more. In
this study, however, only seven subjects (28%) were taking
three or more medications. The reason for the lack of a
significant relationship between number of medications the
patient was taking and compliance, then, may be due to the
fact that compliance behavior was elicited by self-report
and the subjects were answering to the best of their
knowledge. That is, they perceived themselves to be
complying with their physicians' prescriptions when indeed
they may not have been following their recommendations.

The results also showed no statistically significant
relationship between the number of times per day the
subjects were taking medication and perceived compliance
behavior (r = .12). This finding is not surprising given
the fact that past research also has produced no clear
consensus linking frequent dosage with low compliance
(Haynes et al., 1979). It appears, than, that the number of
times per day a patient takes medication does not influence
compliance behavior as perceived by the patient.

In a similar vein, no statistically significant
relationship was found between perceived compliance behavior
and the number of other dietary restriction (r = .16).

Again, this finding may be due to the fact that compliance

173

was measured by subjective self-reports and the subjects may
not have been following all the dietary restrictions as well
as they indicated they were complying. Or, possibly, when
patients are already on a complex diet, such as a diabetic
diet, they may not perceive added diet restrictions as
increasing the complexity of the regimen.

Furthermore, the complexity of the regimen is
increased by other recommendations that are prescribed in
addition to the recommendations necessary for the diabetic
regimen. Studies have revealed that compliance was lower
for patients for whom both drugs and other recommendations
were made (Haynes et al., 1979). In this study, the
majority (22) had no other treatments prescribed. Three
subjects mentioned exercise as another treatment. There
were too few other treatments indicated by the subjects to
be of any statistical significance.

In summary, none of the measures of complexity of
regimen showed a relationship to compliance behavior.
Perhaps if the complexity of the regimen was viewed totally,
a significant relationship may have been found. Or,
perhaps, when a patient already has a complex regimen, which
is common to diabetes, the addition of further dietary
restrictions and medication, may not be refleCted in the
patient's perception, i.e., he/she may not perceive

herself/himself as complying to any lesser degree.

174

H i! J' i . I] J I . I] E i' l I .

Hospitalization in the last six months for diabetes may
involve reinforcement teaching concerning diabetes.
Accordingly, it was postulated that such reinforcement
teaching might act as a cue for compliance and temporarily
improve compliance behavior. This relationship, however,
could not be explored since only one subject had been

hospitalized in the six months prior to the study.

H . I J' I' E E I.

Advances in medicine, have increased life
expectancy. One outcome of this increased life expectancy
is chronic illness. Individuals with such chronic diseases,
however, can reach a maximum potential for daily living.
But compliance with a therapeutic regimen is necessary if
they are to function at their optimum level. And, the need
for such compliance has important implications for nursing
practice.

A primary responsibility of the nurse is to assist
patients to reach their maximum potential for daily living.
Accordingly, the nurse must recognize those factors that
individuals perceive as motivating or as hindering the
achievement of this goal. One factor that individuals may
perceive as motivating or hindering to compliance behavior
is the functioning of their family. This is so because the
family is the core of a person's environment. According to

King (1971), man and his environment is the central focus

175

for the framework of nursing.

Given this focus, King's (1971) theory of nursing
seems particularly relevant for this study. King states
that man functions in a social system which includes the
family unit. Moreover, she states that man functions
through interpersonal relationships which could refer to
family functioning. Further, King states man functions in
terms of her/his perception, i.e., in terms of her/his
conscious awareness. All of King's concepts (social system,
interpersonal relationship, and perceptions) influence the
concepts of life and health. In other words, an
individual's conscious awareness of functioning within the
family through interactions with family members influences
her/his life and health. Thus, since according to King, an
individual's perception of the functioning within the family
influences her/his life and health, and since compliance
behavior in chronic illness also influences life and health,
it seemed reasonable to expect there may be a relationship
between perceived family functioning and stated compliance
behavior.

This study was an effort to assess the relationship
between the perceived family functioning and perceived
compliance behavior of adult onset diabetics. This effort
was not successful. That is, no relationship was found
between perceived family functioning and perceived

compliance behavior yet, as suggested above, the negative

176

findings of this study do not necessarily negate the fact
that the family and the behavior of its members to one
another influence compliance behavior. For example, it was
found in this study that an inverse relationship, albeit
low, existed between perceived family functioning and the
random blood sugar, i.e., as the random blood sugar elevated
the perceived family functioning score declined. Thls
finding suggests that the nurse has an important role to
play in fostering the family functioning and compliance of
patients with diabetes.

For example, the professional role of nursing in
ambulatory care settings involves the systematic assessment
of patients' health states as well as systematic assessment
and management of compliance behavior. Traditionally,
nurses have been instrumental in making the prescribed
regimen understandable and possibly even helping the patient
feel competent to manage the regimen. However, transmitting
information alone is not enough to overcome noncompliance.
It is also necessary to acknowledge that the therapeutic
regimen is only a part of the life regimen the patient must
manage. Thus, one aspect of the nurse's role is to promote
patient participation in identifying barriers to compliance
and helping the individual resolve difficulties that are
interfering with compliance.

To perform this aspect of her/his role the nurse

must first systematically assess the patient's problem.

177

He/she must ascertain whether or not the. patient has
achieved the goal of treatment (in the case of diabetes,
control of blood sugar levels). By reviewing the patient's
record for blood sugar levels, weight, urine glucose levels,
and alpha hemoglobin (if one is available), as well as
asking the patient questions regarding her/his health state,
the nurse can ascertain which patients have not reached
their treatment goals. Next, she must determine whether or
not the regimen prescribed is vigorous enough to achieve the
treatment goal and if the patient has adequate information
concerning the therapeutic regimen. If the answer to these
are affirmative and the patient has not reached her/his
treatment goal, then the nurse may consider compliance to be
a problem.

Although the results of this study suggest that
patients tend to overestimate compliance, most particularly
compliance with diet, Haynes and his associates (1979)
indicate that the most useful approach to elicit
noncompliance among patients in the clinical setting is
simply to ask the patients in a nonthreatening fashion,
whether they are or are not compliant. It is likely that
the patients who will admit to low compliance will be the
ones who will show the greatest response to compliance
improving strategies. In other words, those patients who do
not perceive themselves as noncompliant will not see a need

to change their compliance behavior. A questionnaire

178

similar to the one used in this study may be used to assess
systematically the patients' subjective view of compliance
behavior.. If it is ascertained that compliance with the
therapeutic regimen is a problem, it is appropriate to
investigate variables that may be. interfering with
. compliance.

Based on findings from this study, as well as
previous research, variables that the nurse should
investigate that may be interfering with compliance are:
(1) number of years with diabetes; (2) complexity of the
regimen: (3) number of chronic illnesses other than
diabetes; and (4) the separate components of perceived
family functioning as well as the total percelved family
functioning. For example, although this study did not find
significant relationships between compliance behavior and
the complexity of the regimen, and the number of years with
diabetes, previous research has shown a relationship between
these variables. Until further research has been conducted
to substantiate this lack of evidence of a relationship
between compliance behavior and complexity of the regimen
and the number of years with diabetes, it would be wise for
the nurse to continue to investigate if these variables may
be interfering with compliance behavior because the size of
the sample for this study was small and was not
representative of all diabetic patients and, since previous

studies found a relationship, these variables may in fact

179

influence compliance behavior.

Moreover, the significant inverse relationship found
between the number of other illnesses and perceived
compliance behavior indicates that as diabetes is
complicated by an increased number of other chronic
illnesses, perceived compliance behavior decreases. The
nurse, then, should give particular attention to the
assessment of patients who have multiple chronic illnesses
in addition to their diabetes to ascertain if such patients
are having difficulty with compliance behavior. These
patients may need added assistance in planning and
implementing strategies to improve compliance behavior.

In addition, although no significant relationship
was found between total perceived family functioning and
perceived compliance behavior in this study, there is strong
reason to believe that the behavior of family members toward
the patient affects her/his compliance behavior (Baekeland
and Lundwell, 1973; Heinzelmann and Bagley, 1970: Litman,
1966: Oakes et al., 1970: West et al., 1977). That is,
the separate components of family functioning (adaptation,
partnership, growth, affection, resolve) may have an
influence on the degree to which a patient follows one or
more aspects of the treatment prescribed (e.g., diet,
medication, follow-up appointments). Therefore, it is
appropriate to investigate if any of the components of

family functioning may interfere with compliance behavior to

180

ascertain if nursing intervention concerning the patient's
family functioning is warranted.

And finally, the finding of an inverse relationship
between perceived family functioning and random blood sugar
(i.e., declines in perceived family functioning score with
elevation in random blood sugars) suggests the need for
systematic assessment of perceived family functioning. This
is so because how the patient interacts with her/his family
may have an influence on the outcome, diabetic control.

It would appear that a tool similar to Smilkstein's
Family APGAR screening tool might be useful for such a
systematic assessment. The Family APGAR was developed for
clinical use, is easily administered in the clinical
setting, and has been shown to have construct validity (Good
et al., 1979). The instrument can be used to identify areas
in which deficits in family functioning exist (i.e.,
adaptation, partnership, growth, affection, or resolve). By
means of a questionnaire the patient and the nurse can
together identify problem areas in family functioning.
Further discussion of the problem areas can be initiated by
using open- ended questions suggested by Smilkstein (Table
2, Appendix I). Open-ended questions may help to more
specifically identify the problems.

Once the nurse and the patient determine the way in
which the functioning of the family adversely affects

her/his ability to follow the treatment regimen, it is

181

appropriate to implement the next step in the nursing
process. That is, the nurse, tOgether with the patient,
plan strategies to overcome some of the family functioning
problems and compliance behavior problems as well. Such
strategies may require the involvement of the whole family
or only of certain members. One strategy may be to provide
guidance and counseling to family members. If the family
problems are very complex, another strategy may be to refer
the family to a mental health clinic for family therapy.
Still another strategy may be to educate the family as well
as the patient concerning diabetes, including a discussion
of the effect that diabetes may have on the family. For
example, the alteration of diet and eating habits such as
eating at specified times may require the cooperation of the
whole family. In addition, it may be helpful to discuss
with the family the effect that the family may have on the
compliance behavior of the individual with diabetes as well
as the effect of perceived family functioning on the random
blood sugar. Such educating of the patient and the family
may be achieved by the nurse educating the individuals
herself/himself or it may be achieved by referring the
patient and the family to diabetic education programs.
Following the planning of strategies, the third step
of the nursing process involves implemention of the
strategies by the patient and the family, that were planned

by the nurse and the patient, to improve the patient's

182

family functioning and compliance behavior. The nurse can
offer support to the patient and family as they implement
the strategies.

After implementation of the strategies, the final
step of the nursing process is to evaluate. The nurse can
evaluate whether the patient's perception of the functioning
of her/his family has improved. This evaluation can be
achieved through an interview. It is more important that
the individual perceives her/his family functioning to have
improved rather than the nurse perceiving the individual's
family functioning to have improved. This is so because
according to King (1979) the patients perceptions are
her/his image of reality. And, the patients image of
reality will do more to influence the patient behavior than
the influence of the nurses image of reality on the patients
behavior.

The nurse should then evaluate if the improved
perception of family functioning and the improved stated
compliance behavior has had an effect on the outcome. The
patient and the nurse can subjectively evaluate if there was
an improvement in how the patient appears and her/his
general feeling of well being. The records can be evaluated
to ascertain if there was an increase in weight lose if the
patient was overweight. Also, the nurse can ascertain if
the improved perceived family functioning resulted in

improved diabetic control as indicated by random blood

183

sugar. By these means the nurse and patient could evaluate
if there has .been an improvement in the patient's health
state and if the individual has reached her/his maximum
potential for daily living.

In summary, the purpose of this study was to analyze
the relationship between perceived family .functioning and
perceived compliance behavior. Although this study did not
produce evidence to support a relationship between perceived
family functioning and perceived compliance behavior, until
further research substantiates this finding, the nurse
should realize that chronically ill patients usually live
within a family and the family may influence compliance
behavior as well as the health state. Thus, when assessing
a patient with chronic illness, the nurse should assess the
individual's family functioning and her/his compliance
behavior in addition to the patients health state. If a
problem is identified, the nurse should continue with the
subsequent steps of the nursing process in an effort to

alleviate the problem.

I 1' l' E H . 81 I.

The results of this study have several implications
for nursing education. Although the curriculum of nursing
schools usually include family content, there is a need to
further extend education in the area of family in the man-
agement of chronic illness including diabetes. The tendency

toward a slight inverse relationship (r = .-32 p = < .058)

184

between perceived family functioning and random blood sugar
suggests the importance of the role of the family of
patients with diabetes. That is, it appears that as the
perceived family functioning declines the blood sugar of the
diabetic increases. This would suggest the importance of
family interaction on diabetic control. King (1971) implied
that nursing should be concerned with the patient as well as
the environment in which the individual functions. The
environment includes the patient's family and/or social
system. King's theory should be included at all levels of
nursing education, undergraduate as well as. graduate. A
course on the family and other social systems that influence
the life and health of patients should be included in the
nursing curricula.

The concept of family functioning should be taught
as well as a means to assess family functioning.
Introduction of a family functioning assessment tool such as
Smilkstein's Family APGAR may be useful to professional
nursing students. Nursing students should be taught the
importance of teaching and working with families of patients
with chronic illness as well as the patients themselves.

The teaching of the concept of family functioning
should not be limited to only present students of nursing
but should be extended to professional nurses in practice
through continuing education classes as well as through

staff development classes at various health care facilities.

185

A family functioning assessment tool should be introduced to
professional nurses in practice as well. Practicing
professional nurses should also be made aware of the
importance of teaching and working with the families of
patients with chronic illness.

Nursing students and nurses in practice should also
be taught the importance of assessing compliance behavior,
i.e., compliance behavior being conducive to achievement of
control in diabetes. They should be made aware that
compliance behavior elicited by self-report, may be
overestimated and, therefore, they should seek means whereby
compliance behavior can more effectively be assessed, i.e.,

several assessments of alpha hemoglobin.

1W

As nurses, we need to recognize our lack of
understanding of compliance behavior of patients with
chronic illness. This lack of understanding is especially
prevalent in the area concerning the influence of family
functioning on compliance behavior. Continuing research is
needed to overcome this lack of understanding. This study
has investigated the relationship between perceived family
functioning and percieved compliance behavior of patients
with the chronic illness, diabetes. No relationship was
found between the two variables. One study was located that
investigated the relationship between family functioning, as

determined by a family therapist, and compliance behavior

186

(Steidl et al., 1980). Like this study, no relationship was
found between total family functioning and compliance
behavior. However, in Steidl and associates' study (1980) a
relationship was found between some of the components of
family functioning and compliance behavior. More
descriptive studies concerning family functioning and
compliance behavior as well as some experimental studies
should be completed to explore further the possibility that
there is no relationship between total perceived family
functioning and adherence to a regimen by patients who are
chronically ill. In addition, further investigation of the
individual components of family functioning and their
relationship to compliance behavior, in other chronic
illnesses as well as with diabetes, may prove beneficial.
The confounding variables (number of years with
diabetes, complexity of regimen, and hospitalization in the
last six months) were not statistically significant with
perceived compliance behavior in this study. But, the
number in this sample was small and compliance behavior was
assessed by self— report. A larger sample or a more
objective means of measuring compliance behavior of the
diabetic may have produced, significant relationships
between these variables. By contrast, sample size and
methodology did not appear to have an effect on the
influence of other chronic diseases on perceived compliance

behavior. That is, a statistically significant negative

187

correlation between perceived compliance behavior and the
number of other chronic illnesses in addition to diabetes
was found. All of these findings need to be reevaluated in
future research to substantiate the findings.

It is'difficult to find a totally objective means of
measuring compliance behavior of the diabetic that from a
one time measurement can give sufficient evidence to draw a
conclusion concerning the compliance behavior of the
individual. A longitudinal study would be necessary for
objective measures of compliance behavior in the diabetic to
be of value. However, the one time indirect measure of
weight as related to compliance behavior suggests that
patients do not recognize noncompliance with diet, i.e., 92
percent of the subjects were overweight. Yet, the subjects,
on the average perceived themselves as complying with their
diets midway between “half of the time“ and "most of the
time“. More studies should be completed to investigate the
relationship between percieved compliance behavior and
actual compliance behavior to test the veracity of
self-report.

As stated previously, conducting a study similar to
this study might provide further substantiation of its
findings. Further refinement of the instrument as well as
increasing the number of questions for each component may be
beneficial. By increasing the number of questions for each

component, the individual components of perceived family

188

functioning could be evaluated adequately. Measuring
compliance behavior by more objective means rather than
self-report may give a more accurate estimate and more
convincing evidence of compliance behavior. However, more
than a one time measurement of objective parameters are
usually necessary to evaluate compliance behavior in the
diabetic. In addition, using more stringent clinical
control may add veracity to the findings. Replicating a
study similar to this study but altering certain variables
may add credibility to the findings of this study. Using a
larger sample size would give a truer picture of the
population than a smaller sample size because the larger the
sample the smaller the chance of error. Replicating the
study in other settings, in different geographical areas,
and with patients with other chronic illnesses such as
hypertension, cardiovascular disease, and arthritis may
either substantiate the findings of this study or show a
relationship between perceived family functioning and
perceived compliance behavior.

If further investigations find a correlation between
certain components of family functioning and compliance
behavior, further study should be done to ascertain the
direction of the relationship. For example, do the
components of family functioning influence compliance
behavior or does compliance behavior influence the

components of family functioning? Possibly they influence

189

each other in a circular fashion. Further studies could
also test the predictive value of family functioning and
evaluate whether nursing interventions with families that
have areas of dysfunction can promote adherence to the
therapeutic regimen.

There continues to be a need for systematic
assessment of the family. A few family assessments tools
have been developed. Pless and Satterwhite (1973) have
developed the Family Functioning Index which was shown to be
valid and reliable for family's of children with chronic
illness. Smilkstein's (1978) Family APGAR is still in the
process of being tested. As of now Smilkstein's and Pless
and Satterwhite's tools need further testing to substantiate
their value for systematically assessing family functioning
in other types of family situations. Developing valid and
reliable means to assess family functioning are important to
practice, education, and research.

Nurses in the clinical setting can contribute Ito
research. As the nurse treats the patients holistically,
including assessing the family unit and its effect on the
patient's health state, the nurse should document the
outcome in the patient's record concerning family
functioning and its influence on compliance behavior.
Publishing descriptive reports of their findings would be a
benefical contribution to the knowledge base of nursing.

In summary, suggestions for future research include:

190

(l) investigating the individual components of .family
functioning and their relationship to other chronic
illnesses as well as to diabetes: (2) conducting a
longitudinal study using objective means of measuring
compliance behavior rather than self-report and more
stringent clinical control; (3) investigating the
relationship between self-reported compliance behavior and
actual compliance behavior: (4) using a larger sample size,
- in other settings, and in other geographical areas: (5)
conducting experimental intervention studies to investigate
the predictive value of family functioning for compliance
behavior and blood sugar levels: and (6) continuing to
study the effect of confounding variables on perceived
family functioning and perceived or actual compliance

behavior in chronic illness.

Summary

Chapter VI provided a summary and intepretation of
the research findings in relation to the hypothesis and
confounding variables. Nursing implications for practice,

education, and research were discussed.

191

APPENDIX A

CONFIDENTIAL PATIENT

DATA SHEET

(l-2) Patient No. ___ ___
(3-5) Date _J_

MICHIGAN STATE UNIVERSITY
SCHOOL OF NURSING
Master's Thesis

C E'i l' J E I. I D I S] I

Setting

Please complete. If you have any questions concerning this

feel
free to ask for assistance.

 

 

 

 

 

 

 

 

 

 

 

 

Age
Sex: Male Female
1 2
Race: Black White Other_____
1 2 3
Income: Less than $10,000/year $10,000-19,999/year
2
$20,000 or greater/year
3
Marital status: Single Married 2 Separated
Divorced Widow
4 5

Please indicate the relationship, sex, and age of the people
with whom you live.

Sex Age Relationship
1.
2.

3.
4.

192

Sex ‘ Ass Relationship (Cont'd)

7.
Number of years you have had diabetes.

1-5 6-10 11-15 16-20 more than 20_____
1 2 3 4 5

 

 

 

 

Medications you are taking for your diabetes.

P111
1 2

Insulin

 

 

Number of years you have been following the present course

of treatment for your diabetes.

 

What other illnesses are you being treated for other than
diabetes?

No other illnesses Asthma or Emphysema

 

 

Arthritis
3 4

High Blood Pressure

 

 

Poor.circulation in legs Thyroid Problems

 

 

6

Female Problems
7 8

Heart Problems

 

 

Indicate any other illnesses

 

What other diet restrictions are you on other than your
diabetic diet?

Low Salt Low Cholesterol
1 2 3

None

 

 

 

 

Indicate any other

193

What other medications are you taking other than your
insulin or oral hypoglycemics?

 

 

How many times each day do you take medication?

 

Besides your diet and medications what other treatments do
you follow?

 

Have you been hospitalized for your diabetes in the last six
months?

Yes No
l 2

 

 

194

APPENDIX B

QUESTIONNAIRE

FAMILY FUNCTIONING SECTION

Patient Questionnaire

Below is a series of statements that describe patient's
attitudes concerning their families. Patients vary in the
way they view their family. I am interested in how you view
your family.

Directions: Read each statement carefully and then circle
the most appropriate category that represents how you feel
about each statement made concerning your family.. Please
circle only one answer per question. Do not skip any
questions.
There are no right or wrong answers to these questions.
1. I receive help from my family when something is
troubling me.
Strongly Strongly
Agree Agree Undecided" Disagree Disagree

2. My family spends time together for family activities.

Strongly Strongly
Agree Agree Undecided Disagree Disagree

3. In my family we are not encouraged to express our

opinions.
Strongly Strongly
Agree Agree Undecided Disagree Disagree

4. Our family encourages each member to be independent.

Strongly Strongly
Agree Agree Undecided Disagree Disagree

5. Our family does not work well together toward a common

goal.
Strongly Strongly
Agree Agree Undecided Disagree Disagree

195

10.

11.

12.

13.

My family accepts my wishes to take on new activities.

Strongly . Strongly
Agree Agree Undecided Disagree Disagree

We encourage family members to take care of their own
problems.

Strongly Strongly
Agree Agree Undecided Disagree Disagree.

Our family does not have areas of common interest.
Strongly Strongly
Agree Agree Undecided Disagree Disagree
My family gets upset when I spend time away from home.
Strongly ‘ Strongly

Agree Agree Undecided Disagree Disagree

In my family we have difficulty expressing affection
to each other.

Strongly Strongly
Agree Agree Undecided Disagree Disagree
I feel love for my family members.

Strongly Strongly
Agree Agree Undecided Disagree Disagree

My household chores do not get done if I am unable

to do them.

Strongly Strongly

Agree Agree Undecided Disagree Disagree

Expression of anger is not easy for my family to
accept.

Strongly Strongly
Agree Agree Undecided Disagree Disagree

196

‘14.

15.

16.

17.

18.

19.

20.

21.

My family discusses problems such as allowances,
clothing, and household responsibilities.

Strongly Strongly
Agree Agree Undecided Disagree Disagree

Our family does not enjoy being together.

Strongly Strongly
Agree Agree Undecided Disagree Disagree

Our family would not ask for help from community
agencies.

Strongly ' Strongly
Agree Agree Undecided Disagree Disagree
My family responds to my feelings of love.

Strongly Strongly
Agree Agree Undecided Disagree Disagree

The amount of family's money I am allowed for personal
needs is fair.

Strongly Strongly
Agree Agree Undecided Disagree Disagree

It is not easy for my family to deal with change
within the family unit.

Strongly Strongly
Agree Agree Undecided Disagree Disagree

I accept it when other members of the family spend time
doing things differently from the things I do.

Strongly Strongly
Agree Agree Undecided Disagree Disagree
The amount of responsibility I have in our family is

fair.

Strongly Strongly
Agree Agree Undecided Disagree Disagree

197

22.

23.

24.

25.

26.

Our family is willing to accept help or assistance

from friends.

Strongly

Agree Agree Undecided Disagree

Responsibility is not divided fairly among our

family members.

Strongly

Agree Agree Undecided Disagree

Change within the family is not easy for me to

with.

Strongly

Agree Agree Undecided Disagree

I feel my family cares what happens to me.

Strongly

Agree Agree Undecided Disagree

The amount of space provided for may personal
belongs is fair.

Strongly

Agree Agree Undecided Disagree

198

Strongly
Disagree

Strongly
Disagree

deal
Strongly

Disagree

Strongly
Disagree

Strongly
Disagree

APPENDIX C

QUESTIONNAIRE

COMPLIANCE SECTION

Below is a series of statements that describe certain things
that patients do concerning diet, medication, and keeping of
health care appointments. Patients vary as to the way that
diabetes affects their lives. I am interested in how you
respond to what the doctor, nurse, or dietician has told you
to do because of your diabetes.

Directions: Read each statement carefully and then circle
the meet__epprepriete_eetegery that represents the extent to
which you have done what the statement indicates during the
past tge_weeke. Please circle only ene answer per question.
Do not skip any questions.

1. I have eaten snacks that are not on my diet.
Never Seldom Half of Most of All the
the time the time time
2. I have been known to forget to take my medication.
Never Seldom Half of Most of All the
the time the time time
3. I have eaten more than the number of calories ordered
by the doctor.
Never Seldom Half of Most of All the
the time the time time
4. I have tried to at least maintain my present weight.
Never Seldom Half of Most of All the
the time the time time
5. I have taken my medicine in the amounts ordered.
Never Seldom Half of Most of All the
the time the time time
6. I have eaten the foods I was told to avoid.
Never Seldom Half of Most of All the
the time the time time

199

10.

11.

12.

13.

14.

I have not ignored my diet when I have eaten away from
home.

Never Seldom Half of Most of All the
the time the time time

I have taken my medicine only when I have felt like I
needed it.

Never Seldom Half of Most of All the
the time the time time

I have eaten only the amounts of food suggested on the
diet ordered by the doctor.

Never Seldom Half of Most of All the
the time the time time

I have gotten mixed up and have taken my medicine
wrong.

Never Seldom Half of Most of All the
the time the time time

I have forgotten to take my medicine when I have
been away from home.

Never Seldom Half of Most of All the
the time the time time

I have followed my diet as ordered by the doctor.

Never Seldom Half of Most of All the
the time the time time

I have watched for the negative (bad) effects of my
medicine.

Never Seldom Half of Most of All the
the time the time time

I have gotten my prescriptions refilled before or as
soon as I have run out.

Never Seldom Half of Most of All the
the time the time time

200

15. Following my diet has interferred with my normal
daily activities.

Never Seldom Half of Most of All the
the time the time time

16. I have taken my medicine at the times ordered by my
doctor. -

Never Seldom Half of Most of All the
the time the time time

Directions: Read each statement carefully and then circle
the mee;_eppreprie;e_eetegery that represents the extent to
which you have done what the statement indicates during the

past eix__men;he. Please circle only one answer per
question. Do not skip any questions.

17. I have been known to forget my appointments.

Never Seldom Half of Most of All the
the time the time time

18. I have had my blood sugar checked whenever the doctor
suggested I should.
Never Seldom Half of Most of All the

the time the time time

19. I have not gone for my follow-up appointments when I
have felt good.

Never Seldom Half of Most of All the
the time the time time

20. I have reported reactions from myIdiabetic medication
to the doctor.

Never Seldom Half of Most of All the
the time the time time

201

21.

22.

23.

24.

I have kept my appointments to see the doctor or
nurse.

Never Seldom Half of Most of All the
the time the time time

I have avoided having my blood sugar drawn whenever
possible. -

Never Seldom Half of Most of All the
the time the time time

I have gone to the doctor only when I have felt bad.

Never Seldom Half of Most of All the
the time the time time

If something has occurred making it impossible for me
to keep my appointment I have rescheduled my appoint-
ment for as soon as possible.

Never Seldom Half of Most of All the
the time the time time

202

APPENDIX D

CHART AUDIT WORK SHEET

CHART AUDIT WORK SHEET

Patient No.

 

Most recent physical measurements: Height Weight

 

 

Age at onset of diabetes:

Number of years with diagnosed diabetes:

Type of calculated diabetic diet:

 

Follow-up appointments for diabetes scheduled in the last
six months. (Circle if did not keep appointment).

Date:

 

 

 

Weight each visit:

 

 

 

Blood Sugar results each visit:

 

 

 

Type of test:

Normal lab value:

 

 

 

Patient lab value:

 

 

 

Urine glucose each visit:

 

Type of test:

 

Glucose:

 

 

 

 

 

Ketones:

 

203

APPENDIX E

LETTER TO PHYSICIAN

10751 Barnes Rd.
Durand, Mi. 48429
March 17, 1981

Dear Dr.

I am Barbara Sipes R.N., a M.S.U. graduate student,
conducting research for my Master's in nursing thesis. I
would like your permission to use some diabetic patients
from your private practice. A brief outline of my plans are
as follows.

I plan to begin my, study ”The Relationship Between Family
Functioning and Compliance Behavior of the Adult Onset
Diabetic,” in March. I will collect data until 25 subjects
have been included in the study. It is necessary that
either you or your nurse ask your diabetic patients if they
would consider participating in a research project and give
them the introductory letter explaining the nature of the
study. If an individual agrees to participate he/she must
sign the consent form. The consent form will also allow me
access to the following information from the patient's
record concerning the specified criteria necessary to be
part of this study.
- A known diabetic as indicated in the records for at
least one year prior to the time of selection
- Onset after age 29
- Between 31-70 years of age
- On a calculated diabetic diet
- On an oral hypoglycemic or less than 60 units insulin
per single dose
- Seen at least two times per year in an ambulatory
care setting

In addition to the above mentioned specified criteria the
subject must live with at least one other person, must not
be pregnant, and must be able to fill out the questionnaire
without assistance. I will contact each subject who meets
the specified criteria by phone and will arrange a time with
the individual when I can administer the questionnaire.

I will need a member of your staff to identify prospective
subjects when they come in for appointments, give the
subjects the introductory letter and consent form, and refer
them to me. It should involve 5-10 minutes of their time.
I will be the only person involved in the data collection
besides the staff member who identifies the diabetic
patients. I will need additional information from the chart
concerning blood sugar, urine glucose, weight, and number of
follow-up appointments in the last six months.

204

Patients' rights, confidentiality, and anonymity will be
safeguarded as follows:
- Patients will not be solicited by the researcher
- Patients will be assured as to their right to refuse
or withdraw without it affecting their care in any
way
- As much privacy will be provided as possible give the
physical restraints of the situation
- The researcher will administer the questionnaire
- The questionnaire will be coded with no name on the
questionnaire
- The consent form will be separated from the question-
nalre
- No one from the ambulatory care setting will be per-
mitted to read the questionnaire following completion
- Data will be accessible only to the researcher and
thesis committee

If this is agreeable please sign the attached consent form.

Sincerely,

Barbara Sipes R.N.

I grant permission for Barbara Sipes R.N. access to clients
from my private practice to conduct research for her
Masters'in Nursing thesis concerning "The relationship
between family functioning and compliance behavior of the
adult onset diabetic”.

 

Signature

 

Date

205

APPENDIX F

CHECK LIST OF PATIENTS

ADMITTED TO THE STUDY

Date:

 

W

Obtain from patient record:

A known diabetic as indicated in the records
for at least one year prior to the time of
selection

 

Onset of diabetes after age 29

 

Age at time of admission to the study -no more
than 70 no less than 31

 

On a calculated diabetic diet

 

On an oral hypoglycemic or less than 60 units
insulin per single dose

 

Seen at least two times per year in an ambulatory
care setting

 

Patient reports:

Lives with at least one other person

 

Is able to read and fill our the questionnaire
without assistance

 

Is not pregnant

 

206

APPENDIX G

INTRODUCTORY LETTER

MICHIGAN STATE UNIVERSITY
SCHOOL OF NURSING

Barbara Sipes R.N., a Michigan state graduate nursing
student, is conducting a study for her thesis which is a
part of the requirements for a master's degree in nursing.
She is looking for diabetic patients who live with at least
one other person to participate in her study. The study is
seeking to determine the relationship between family
functioning and its influence on the adult diabetic
following the treatment plan.

The study will involve completing one questionnaire
with two parts, one concerning your family and one
concerning following your diabetic treatment plan. The
questionnaire should take about 20-30 minutes to complete.
If you are willing to participate, the following consent
form will need to be signed. The consent form will indicate
you agree to participate and it will also grant her
permission to obtain the following information from your
medical records:

A known diabetic for a least one year prior to the time of
selection

- Onset after age 29
- Between 31-70 years of age
- On a calculated diabetic diet

- On an oral hypoglycemic or less than 60 units insulin per
single dose

- Information concerning weight, blood sugar, urine glucose,
and appointments

If your records indicate you meet this criteria she
will contact you by telephone to arrange a time when you
could fill out the questionnaire. She will arrange to meet
you at your physician's office or bring the questionnaire to
your home, whichever you prefer.

Thank you so much for your time and cooperation.

207

APPENDIX H

CONSENT FORM

MICHIGAN STATE UNIVERSITY
SCHOOL OF NURSING
CONSENT FORM

Investigator: Barbara Sipes, R.N., B.S.N.
Graduate Student, Family Nurse Clinician
School of Nursing
Michigan State University
(313) 266-4490

The study in which you are about to participate is
seeking to determine the relationship between family
functioning and its influence on adult diabetic following
the treatment plan. This study is being conducted by myself
as part of the requirements for a master's degree in
nursing. The study may not benefit you personally, but we
hope the results of this study will help us improve health
care delivery for all diabetic patients.

This study will involve completing one questionnaire
with two parts, one concerning your family and one
concerning following your diabetic treatment plan. The
questionnaire should take about 20-30 minutes to complete
and will require you to respond to the questions as honestly
and accurately as possible. Additional information such as
age, sex, race, income, relationship of family members, as
well as specific questions about your diabetes and other
illnesses will also be obtained. It will be necessary to
obtain information from your medical record. Your answers
will be kept in complete anonymity and no attempt will be
made to identify you in any manner. You are free to
withdraw at any time. Withdrawal from the study will in no
way affect the care you are now receiving.

As a result of participation in this study you may
become more aware of your disease and/or your family
relationships. If this occurs, I am willing to speak with
you or refer you to a counselor (at your expense) for
further assistance. If you have any questions, feel free to
ask me immediately or call me. Thank you for your time and
cooperation.

Barbara Sipes
Family Nurse Clinician Graduate Student
Telephone: (313) 266—4490

 

I voluntarily consent to participate in this study. I

have had an opportunity to ask questions. I may withdraw
from the study at any time without penalty if I choose to.

 

 

Signature of Subject Date

 

 

Signature of Investigator Date

208

APPENDIX I

FAMILY APGAR TABLES

FAMILY APGAR TABLES*

TABLE 1. Definitions of Family APGAR Components

 

Component

Definition

 

Adaptation

Partnership

Growth

Affection

Resolve

Adaptation is the utilization of intra and
extrafamilial resources for problem solving
when family equilibrium is stressed during
a crisis.

Partnership is the sharing of decision mak-
ing and nurturing responsibilities by family
members.

Growth is the physical and emotional matur-
ation and self-fulfillment that is achieved
by family members through mutual support
and guidance.

Affection is the caring or loving relation-
ship that exists among family members.

Resolve is the commitment to devote time to
other members of the family for physical and
emotional nurturing. It also usually invol-
ves a decision to share wealth and space.

 

*Smilkstein, Gabriel. The Family APGAR: A Proposal for a
Family Function Test and Its Use by Physicians, The_aeurnal
of Family Practice, Vol. 6, No. 6, 1978.

209

TABLE 2. Open-ended Requests for Family Function Information

 

Component

Relavant Open-Ended Questions

 

Adaptation

Partnership

Affection

.Besolve

How have family members aided each other in
time of need?

In what way have family members received
help or assistance from friends and com-
munity agencies?

How do family members communicate with each
other about such matters as vacations, fi-
nances, medical care, large purchases, and
personal problems?

How have family members changed during the
past years? How has this change been accept-
ed by family members? In what ways have fam-
ily members aided each other in growing or
developing indepenent life-styles? How have
family members reacted to your desires for
change?

How have family members responded to emo-
tional expressions such as affection, love,
sorrow, or anger?

How do members of your family share time,
space and money?

 

210

TABLE 3. Family APGAR Questionnaire

 

Almost Some of Hardly
always the time ever

 

I am satisfied with the
help that I receive from
my family* when some-
thing is troubling me.

I am satisfied with the
way my family* discusses
items of common interest
and shares problem solv-
ing with me.

I find that my family*
accepts my wishes to
take on new activities
or make changes in my
life-style.

I am satisfied with the
way my family* expresses
affection and responds
to my feelings such as
anger, sorrow, and love.

I am satisfied with the
amount of time my family*
and I spend together. _____. _____. _____

 

Scoring: The patient checks one of three choices which are
scored as follows: 'Almost always' (2points), 'Some of the
time' (1 point), or 'Hardly ever' (0). The scores for each
of the five questions are then totaled. A score of 7 to 10
suggests a highly functional family. A score of 4 to 6 sug-
gests a moderately dysfunctional family. A score of 0 to 3
suggests a severely dysfunctional family.

*According to which member of the family is being inter-

viewed the physician may substitute the word 'family' for
either spouse, significant other, parents, or children.

211

TABLE 4. What is Measured by the Family APGAR?

 

Component

 

Adaptation How resources are shared, or the degree to
which a member is satisfied with the assis-
tance received when family resources are
needed.

Partnership How decisions are shared, or the member's
satisfaction with mutuality in family
communication and problem solving.

Growth How nurturing is shared, or the member's
satisfaction with the freedom available
within the family to change roles and
attain physical and emotional growth or
maturation.

Affection How emotional experiences are shared, or
the member's satisfaction with the intimacy
and emotional interaction that exists in a
family.

Resolve How time (and space and money*) is shared,
or the member's satisfaction with the time
commitment that has been made to the family
by its members.

 

*Besides sharing time, family members usually have a com-

mitment to share space and money. Because of its primacy,
time was the only item included in the Family APGAR; how-
ever, the physician who is concerned with family function

will enlarge her/his understanding of the family's resolve
if he/she inquires about family member's satisfaction with
shared space and money.

 

212

APPENDIX J

QUESTIONS USED FOR

DATA ANALYSIS

QUESTIONS USED FOR DATA ANALYSIS
Eamilwnotionino

Adaptation

12.

My household chores do not get done if I am unable to
do them.

16. Our family would not ask for help from community
agencies.
22. Our family is willing to accept help or assistance from
friends.
Partnership
3. In my family we are not encouraged to express our
opinions.
8. Our family does not have areas of common interest.
14. My family discusses problems such as allowances, cloth-
ing, and household responsibilities.
21. The amount of responsibility I have in our family is
fairs.
23. Responsibility is not divided fairly among our family
members.
Growth
4. Our family encourages each member to be independent.

6. My family accepts my wishes to take on new activities.
9. My family gets upset when I spend time away from home.
20. I accept it when other members of the family spend time
doing things differently from the things I do.

24. Change within the family is not easy for me to deal

with.

213

Affection

10. In my family we have difficulty expressing affection to
each other.
11. I feel love for my family members.
13. Expression of anger is not easy for my family to
accept.
17. My family responds to my feelings of love.
25. I feel my family cares what happens to me.
Basolxo
2. My family spends time together for family activities.
5. Our family does not work well together toward a common
goal.
15. Our family does not enjoy being together.
18. The amount of our family's money I am allowed for per-
sonal needs is fair.
26. The amount of space provided for my personal belongings
is fair.
Somalianoa
Diet
1. I have eaten snacks that are not on my diet.
3. I have eaten more than the number of calories ordered
by the doctor.
4. I have tried to at least maintain my present weight.
6. I have eaten the foods I was told to avoid.
9. I have eaten only the amounts of food suggested on the
diet ordered by the doctor.
12. I have followed my diet as ordered by the doctor.
15. Following my diet has interferred with my normal daily

activities.

214

M 3' I.

I have been known to forget to take my medication.

2.
5. I have taken my medicine in the amounts ordered.
8. I have taken my medicine only when I have felt like I
needed it.
10. I have gotton mixed up and have taken my medicine
wrong.
11. I have forgotton to take my medicine when I have
been away from.
13. I have watched for the negative (bad) effects of my
medicine.
14. I have gotton my prescriptions refilled before or as
soon as I have run out.
W
18. I have had my blood sugar checked whenever the doctor
suggested I should.
19. I have not gone for my follow-up appointments when I
have felt good.
20. I have reported reactions from my diabetic medication
to the doctor.
21. I have kept my appointments to see the doctor or nurse.
22. I have avoided having my blood sugar drawn whenever
possible.
23. I have gone to the doctor only when I have felt bad.
24. If something has occurred making it impossible for me

to keep my appointment, I have rescheduled my appoint-
ment for as soon as possible.

215

BIBLIOGRAPHY

BIBLIOGRAPHY

Abram, H. S. The psychology of chronic illness. Jenrnal
of_§hronio_Disoasas, 1972 25. 659-664-

Ary, D., Jacobs. L., Razavieh, A. Introdnotion_to__Rasoaroh
in__,§deeetien. New York: Holt, Rinehart, and
Winston, Inc., 1972.

Baekeland, F., & Lundwall, L. Dropping out of treatment: a

critical review. Esxohologioal_8nllatin.,1975, 82,
738-783.

Barsky, A. J. Patient heal thyself: activating the

ambulatory medical patient. Jon;nal__of__§hronio
Dieeeeee, 1976, 22, 585- 597.

Becker, M. H. Understanding patient compliance: the
contributions of attitudes and other psychosocial
factors. In Cohen, S. J. ' '

' . Lexington: D. C. Health 5
Co., 1979.

Becker, M., Drachman, R., Kirscht, J. Predicting mothers'
compliance with pediatric medical regimens. Journal
of_£odiatrios,1972, 81, 843- 853.

Becker, M. Drachman, R., Kirscht, J. A new approach to
explaining sick-role behavior in low 1 income

Populations. Amerioan__Jonrna1__of__£nhlio__flealth,
1974, 54, 2059216.

Bille, D. A. The role of body images in patient compliance
and education. Heart_Lnno, 1977, 6, 143-148.

Blackburn, S. L. Dietary compliance of chronic
hem°dialysis patients. Journal_of_Amarioan_Diatatio
Asaooiation, 1977, 10, 31-37.

Blackwell, B. Drug therapy: patient compliance. New
England_Jonrnal_of_Madioino, 1973 282, 249- 252.

Blevins, D. Tha__Diahetio__and__Nnrsins__£are. New York:
McGraw-Hill Book Co., 1979.

Blood, R., & Wolfe, D.

. New York: The Free Press of
Glencoe, 1960.

216

Bondy, P., & Felig, P. Relation of diabetic control to
development of vascular complications. Medieal
£11n198_o£_Notth_Amati£a 1971 55(1), 839- 897-

Bowen, R. G., Rich, R., Schlotfeldt, R. M. Effects of
organized instruction for patients with the

diagnosis of diabetes mellitus. Nursing__Beeeareh,
1961, 10, 151-159.

Capell, P. T.. Case. D. B. Amhnlator¥__£ara__Mannal__for
Nursa__.£raotitionars. Philadelphia: J- 3-
Lippincott Co., 1976. ‘

Caplan, R., Robinson, E., French, J., Caldwell, J., Shinn,
M. -
' ' ' . Ann Arbor:

Institute for Social Research, 1976.

Cavan, R., & Ranck, K. Tha__Eamil¥__and__tha__nanrassion.
Chicago: University of Chicago Press, 1938.

Cerkoney, K. A., Hart, L. K. The relationship between the
health belief model and compliance of persons with
diabetes mellitus. D1ehetee__§ere, 1980, 1(5),
594-598.

Charney, E., Bynum, R., Eldredge, D., Frank, D., MacWhinney,
J., McNabb, N., Scheiner, A., Sumpter, E., Iker, H.
How well do patients take oral penicillin? A
collaborative study in private paractice.

Pediatxiee, 1967, 50, 188-195.

Cohen, E., & Etzwiler, D. Diebetee___Mennel. Health
Education for Living Program, Inc., 1976.

Crano, w., 8 Brewer, M. Prinoio1as_of__Basaaroh__in__Sooial
Peyehelegy. New York: McGraw-Hill Book Co., 1973.

Davis, M. Variations in patients' compliance with doctors'

orders: analysis of congruence between survey
responses and results of empirical investigations.
Jotrnal_of__naoioa1__£duoation, 1966, 41, 1037-1048.

Davis, M. Physiologic, psychological, and demographic
factors in patient compliance with doctors' orders.
' 19683, 6,, 115-122.

Davis, M. Variations in compliance with doctor's advice:
an empirical analysis of patterns of communications.

lournal__of__£nhlio__Haalth, 1968b, 58,
274-288. ‘

217

Diamond, M. D., Weiss, A. J., Grynbaum, B. The
unmotivated patient. ' '

_ahahilitation, 1968, 42, 281- 284.

Donabedian, A., 8 Rosenberg, L. Follow-up study of
chronically ill patients discharged from hospitals.
, 1964,11, 847- 862.

Dunbar, J. Issues in assessment. In Cohen, S. J. (ed)
' ' . Lexington:
D. C. Health 8 Co., 1979.

Etzwiler, D. D., 8 Robb, J. R. Evaluation‘ of programmed
education among juvenile diabetics and their
families. Diebetee, 1972, 21, 967-971.

Ferguson, K., 8 Bole, G. G. Family support, health
beliefs, and therapeutic compIiance in patients with
rheumatoid arthritis.

Haalth.£dooation,1979,1(3), 101-104-

Francis, V., Korsch, B. M., Morris, M. J. Gaps in

doctor-patient communication. Nen__England_Jonrnal
of_fladioino, 1969, 280, 535-540-

Freeman, R. 9ommunitx____Haalth____Hursino____Eraotioa,
Philadelphia: W. B. Saunders Co., 1970.

Geismar, L. L., LaSorte, M., Ayres, B. Measuring family

disorganization. Marriage__£emily_Liying, 1962, 25,
51.

Geyman, J. The family as the object of care in family
practice. 1onrnal__of__£amilx_£raotioa, 1977, 5(4),

Gillium, R., 8 Barsky, A. Diagnosis and management of
patient noncompliance. JAMA, 1974, 228(12),
1563-1565.

Good, M. J., Smilkstein, M., Good, B., Shaffer, T., Arons,
T. The family APGAR index: a study of construct

validity. Jonrnal__of__£amil¥__£raotioa. 1979. 8,
577-582 0

Gordis, L. Conceptual and methodologic problems in
measuring patient compliance. In Haynes, B.,
Taylor, D., Sackett, D. (eds) '
flare. Baltimore: John Hopkins University, 1979.

218

Green, L. W., Levine, D. M., Wolle, J., Deeds, S.
Development of randomized patient education
experiments with urban poor hypertensives. Patient

, 1979, 1(3),
106-111.

Guthrie, D., 8 Guthrie, R. '
.Mellitne. Saint Louis: C. V. Mosby Co., 1977.

Haggerty, R. J. Family diagnosis: research methods and
their reliability for studies of medical-social
unit, the family. ' '
Health, 1965, 55, 1521—1533.

Haynes R. B., Sackett, D. L., Taylor, E. W. How to
detect and manage low patient compliance in chronic

illness. Getiattiee,, 1980, 35(1), 91- 93, 96- 97.

Haynes, R. B. Taylor, D. W., Sackett, D. L. Comolianoa
in_Health_§axe. Baltimore: John Hopkins University
Press, 1979.

Heinzelmann, R. Factors in prophylaxis behaviors in
treating rheumatic fever: exploratory study.

Journal_of_Hnman_Bahaxior, 1962 3, 73- 81.

Heinzelmann, F., Bagley, R. Response to physical activity
programs and their effects on health behavior.

Pnhlio_Haalth_Rasnort, 1970, 85, 905-911-

Hershey, J. C., Morton, B. G., Davis, J. B., Reichgott,
M. J., Patient compliance with antihypertensive

medication. Amarioan__Jonrnal__of__Pnhlio___Haalth,
1980, 19(10), 1081-1088.

Hulka, B., Kupper, L., Cassel, J., Mayo, F. Doctor-patient
communication and outcomes among diabetic patients.

Journal_of_§ommunit¥_Haalth.1975, l, 15- 27.

King, Imogene. Toward_a__Thoor¥__for__Nnrsino. New York:
John Wiley and Sons, 1971.

King, R. : .
Berkeley, California: The Glendessary Press, 1969.

Kirscht, J. P., Rosenstock, I. M. Patient adherence to
antihypertensive medical regimens. Journal___9£

Communit¥_Haalth 1977 3(2). 115-124

Leslie, R. How quickly can haemoglobin Al increase?

British_Madioal_Journal, 1979, 1, 19.

219

Litman, T. J. The family as a basic unit in health and
medical care: a social- -behavioral overview. Sneial

Ssionos_and_nsdioine,1974.8, 495- 519.

Litman, T. J. The family and physical rehabilitation.
loorna1_of_£hronio.Disaasas,1966,19, 211.

Malone, J., Hellrung, J., Malphus, E., Rossenbloom, A.,
Grigic, A., Weber, F. Good diabetic control--a

study in mass delusion. Jonrna1__of__£adiatrios.
1976 p 13' 943-947 D

Marston, Mary-vesta. Compliance with medical regimens: - a

review of the literature. ungaing_Reeeateh, 1970,
12, 312-321.

Mechanic, D. The concept of illness behavior. Journal__of
Qhronio_Diaoasos,1962,15, 109-194

Nye, F. I., Bernardo, F. M. Emergino___9onoaotnal
Eramaworts__in___amilx__Ana1¥sis. New York: The
MacMillian Co., 1966.

Oakes, T., Ward, J., Gray, R., Klauber, M., Moody, P.
Family expectations and arthritis patient compliance
to a hand resting splint regimen. Jon;nal__of

Qhronio_Disoase.197o,22, 757- 764.

Peretz, D. Reactions to loss. In B. Schoenberg, et al.
(Eds. )r ' .
Medieal_£taetiee. New York: Columbia University
Press, 1972.

Pless, I. B., Satterwhite, B. A measurement of family

functioning and its application. Seeial__Seienoe
”fidiflinfir 1973' ll 613- '

Podell, R. N., Gary, L. R. Compliance: a problem in

medical management. Amstioan_£amil¥_kraotioe,1976,
1.3.. 74-800

Rakel, R. Principles of Family Medicine. Philadelphia: W.
B. Saunders, 1977. A

Sayetta, R. B., Murphy, R. S. Summary of current
diabetes-related data from the national center for
health statistics. Diabetes__£ars, 1979, 2(2),
105-119 D

Schmidt, D. The family is the unit of medical care.

aoornal_of_Eamil¥_£raotioai_19181_1(2), 303- 313.

220

Skillman, T. G., Tzagournis, M. Diabetee_ne111tne, Upjohn
Co., 1977.

Skyler, J. S., Skyler, D. L., Seigler, D. E., O' Sullivan,
M. J. Diah£t£5_£are, 1981,1(2), 311— 318.

Smilkstein, G. The family in trouble--how to tell. Journal
of_£amilx_2raotioa. 1975, 2. 19-24.

Smilkstein, G. The family APGAR: a proposal for a family
functioning test and its use by physicians. Journal

of__amilx_£raotioe., 1978, 8. 1231-1234.

Steidl, J. H., Finkelstein, F. O., Wexler, P., Feigenbaum
H., Kitsen, J., Kliger, A. S., Quinlan, D. M.
Medical condition, adherence to treatment regimens,
and family functioning. '

W! 1980' 3.1: 1025-1027.

Tagliacozzo, D. 8 Ima, K. Knowledge of illness as a
predictor of patient behavior. Journal_of_£hronio
Disease, 1970, 22, 765- 775.

Taglizcozzo, D., Luskin, D., Lashof, J., Ima, K. Nurse
intervention and patient behavior. Amerioan_lonrnal

of_P.uhliLHaalth, 1974,14, 596- 603.

Tattersall, R., Walford, S., Peacock, J., Gale, E.,
Allision, S. A critical evaluation of methods of
monitoring diabetic control. Diabetee__Care, 1980,
3,(1), 150-154.

Taylor, D. A test of the health belief model in
hypertension. In Haynes, B., Taylor, D., Sackett,

D. (eds) Wealth—Cara Balitimore:
John Hopkins University, 1979. 4 '

Van Ormer, E. B., 8 Williams, C. D. W
' . Ann

Arbor: Edwards Brothers, Inc., 1941.

Vincent, P. Factors influencing patient noncompliance: a
theoretical approach. Nure1n9__Reeearoh, 1971, 20,
509- 516.

Watkins, J., Williams, T., Martin, D., Hogan, M., Anderson,
E. A study of diabetic patients at home. Amerioan

WW,1967,51, 452- 459.

221

West, D., Graham, 8., Swanson, M., Wilkenson, G. Five year
follow-up of a smoking withdrawal clinic population.

Journal__of__£nhlio__Hea1th, 1977, 61,
536-544.

Williams, J. R., Alling, K. A program for normal glycemic
control of insulin-dependent diabetes.

Care, 1980, 1(1), 160-162.

Williams, M., Martin, M., Hogan, M., Watkins, J., Ellis, E.
The clinical picture of diabetic control, studied in

four settings. Amer1oan__Janinal_9£_2nhli£—HfiilihI

1967, 51, 441- 451.

222