..;
..

‘3.
,.

X:

M “ 9:13:
‘3' ’ “$113

‘A 9 A.
32222:.»

. ,. .. _.
6‘2: .3.-

.12K’t',‘e'{f_ 2‘ 4 u...
‘fi; . .-;;-Er.:.‘.3.'y.}.,
“ Uri. .. 2.35;;

, - -~‘r~

in : ~' . i! _ _

WI " h ‘

2.2"“.ch . ‘ . ‘ ‘ 3 .L , 3.2%;

u. I} 3 ‘ ~.\'.-.:<‘:7.f. a
‘ 'l R

N.
#3.,

{2'59" :‘Wn;
I P 49:. ":5’3'31

(-Jv

:33?
dfin

04$“
. 1.-
I :Lu’so.d'-i-\‘§.I.G‘
' 2:31: v- I;
”-1; .. .
41V}-

.21..

:‘ .2

A:~
i‘

'32.

A
‘4

a"’. r"
9| ~u p.
.1? .9!

,. “.54.

4‘.
.‘3.
g

:5:
i...
4.2

w J

.
“2:164 .;-: ‘

69%

’27 v.

4:"? '3’?»

, ..._,.. ,. ..
220 9.75:5

‘r

IE" I;

I.
3

. .9 A...
. .
. ’1", ‘
”Um.-.
an

'1" 2
W n 5.

,m
u
u

1:.

3

["1

{a
J} v
'g‘flf.
«:3 :71

."-

r.--:. 536‘”?
‘u 'V‘lfléf ‘; v'ff‘
no 9“,“ ..
97:5. (iv-fl

Fl"), ..

..

 

3 ‘.

37M;
' L“ . .
‘ :‘Ji Ink

. I‘ ‘
.4 "tr ‘

\l/Hh - .

411

u?“-

. v5 a’. ,. ' 2
"‘ '. . . a '3 'Q
‘ . ' ‘IJV‘I' ”3-3.1.

 

IVE RS TY LIBRARIES

INHUHIHHIIlUIW'IUIH|||1 UH 0lol

3 1293 010152

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled

Physician Self-Efficacy

in the Treatment of Obesity

presented by

Amy Stern Stoffelmayr

O o ‘

has been accepted towards fulfillment
of the requirements for

Doctoral degree in Weathermen:
Counseling, Educational
Psychology and Special Education

    

ajor professor

Date f/Z/fy

MSU i: an Affirmative Action/Equal Opportunity Institution 0- 12771

 

 

LIBRARY
Michigan State
University

 

 

 

PLACE N RETURN BOXtommthbMMflunywmd.
TOAVOIDFINESMumonorbdondatoduo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

359:3 132.99

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU loAn WWW Opportunity Intuition
mm:

 

PHYSICIAN SELF-EFFICACY IN THE TREATMENT OF OBESITY

by

Amy Stern Stoffelmayr

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Counseling, Educational Psychology, and

Special Education

1994

ABSTRACT
PHYSICIAN SELF-EFFICACY IN THE TREATMENT OF OBESITY
BY

Amy Stern Stoffelmayr

Primary care physician level of confidence or "self-
efficacy" in the treatment of obesity varies, on an
anecdotal basis. Patient self-efficacy with regard to
disease and treatment is known to vary. It is well-
researched and is associated with a myriad of positive
health outcomes--both physical and psychological.

The present study surveyed a large and varied group of
physicians about their beliefs about their own obesity
treatment capabilities, their beliefs about the outcomes of
obesity treatment and their actual practice behaviors with
obese patients. A smaller number of physicians were
interviewed to discover what thoughts, feelings, ideas and
experiences they had that fostered their self—efficacy.
They were interviewed about their conceptualizations of the
problem of obesity and its treatment to discover what
supported the development and maintenance of self-efficacy
beliefs.

Measures of obesity treatment self-efficacy and of
obesity treatment outcome expectations were formulated that
had reasonable reliability. The measures correlated
moderately positively with each another. The scores on the

self-efficacy measures and outcome expectations measures

were found to be related to practice behaviors in a way
which is consistent with self-efficacy theory. Physician
self-efficacy beliefs, as measured by the study instruments
predicted self-efficacious physician behaviors.

Self-efficacy factors that predicted efficacious
behaviors were physician beliefs about their own
interpersonal skills and their own professional knowledge
base regarding obesity and its treatment. The "affect
management" factor (how well physicians’ felt they could
deal with any negative feelings they had towards their obese
patients) did not, on its own, predict efficacious
behaviors. I

Outcome expectation factors that were found to be the
most powerful predictors of self-efficacious behaviors were
physician beliefs that obesity was treatable and that it was
a shared responsibility between doctor and patient, that

obesity treatment was effective and that it was important.

I observe the Phisician with the same dilligence, as hee

the disease; I see hee feares, and I feare with him.

John Donne

Devotions, No.6, 1624

ACKNOWLEDGEMENTS

I would like to thank my dissertation committee: Drs.
Nancy Crewe, Deb Feltz, Linda Forrest, Steve Radenbush, and
to especially to thank my two dissertation directors, Linda
Forrest for her assistance at this project’s inception and
Nancy Crewe for her assistance at its completion.

I would like to express my appreciation to Bertram
Stoffelmayr for providing me with a role model of

intellectual integrity.

TABLE OF CONTENTS

List of Tables ----------------------------------------- page vi

Chapter I

Introduction and Overview --------------------------------- page

Chapter II

Literature Review
Study Constructs ------------------------------------- page
Self-Efficacy and Research on Health and Illness ----- page
Teacher Self-Efficacy -------------------------------- page
Primary Care Physicians and Health Behavior Change---page
Obesity-s -------------------------------------------- page
Research Questions and Hypotheses -------------------- page

Chapter III

Method
IntrOduCtion ---------------------------------------- page
DESign ------------------------------------------ page
SUbjeCtS ------------------------------------------ page
Development of the Physician Self-Efficacy Measure--page
Reliability --------------------------------------- page
Validity -------------------------------------------- page
Development of the Self-Efficacy Interview ---------- page
Procedure ------------------------------------------- page
Delimitation and Limitations ------------------------ page
Treatment Of the Data ------------------------------- page
Chapter IV
Results
Quantitative .......................................... page
StUdY sample ---------------------------------------- page
Evaluation of the Study Measures -------------------- page
Evaluation of the Physician Behavior Variables ------ page
Association of Physician Characteristics and
Self-Efficacy Measure Scores ------ page
Association of Physician Behavior Variables and
Self-Efficacy Measure Scores ------ page
Interpretation -------------------------------------- page
Generalizability ------------------------------------ page
Goals Of the Study ---------------------------------- page
Research Questions ---------------------------------- page
Hypotheses ------------------------------------------ page

vi

ii

1

6

16
21
22
25
28

36
36
37
38
4O
41
45
47
48
49

52
52
56
63

65

67
76
77
78
80
81

vii

Qualitative ----------------------------------------- page

Sources of Efficacy Information ------------------- page

How the Problem is Defined ------------------------ page

Interpretation ------------------------------------ page

Attributions which Foster Self-Efficacy ----------- page

Attributions which Undermine Self-Efficacy -------- page
Chapter V
Discussion

Research Question One ----------------------------- page

Study Goal Two ------------------------------------ page

Study Hypotheses ---------------------------------- page

Research Question Two ----------------------------- page

Practical Applications ---------------------------- page

Theoretical Implications -------------------------- page
Appendices

A. Physician Self-Efficacy Measure ---------------- page

B. Physician Outcome Expectations Measure --------- page

C. Physician Practice Measure --------------------- page

D. Self-Efficacy Interview ------------------------ page

References --------------------------------------------- page

83
85
90
95
96
98

100
103
104
105
113
115

119
120
121
122

125

LIST OF TABLES

Table 1. Description of the Study Sample - — — - - page

Table 2. Principal—Components Analysis and
Rotated Factor Matrix (PSEM) - - - - page

Table 3. Principal-Components Analysis and
Rotated Factor Matrix (POEM) - - - - page

Table 4. Correlation Matrix of PSEM and POEM Total

Scores and First and Second PSEM and
POEM Factor Scores - - — - - - - - - page

Table 5. Medical Degree, Practice Size and
PSEM and POEM Scores - - ~ - — - - - page
Table 6. Patient Load and PSEM and POEM Scores - - page
Table 7. Correlation Matrix of Physician Behavior
Scores and PSEM and POEM Scores - - — page
Table 8. Physician Behaviors and PSEM Scores - - — page

Table 9. Physician Behaviors and PSEM Factor
Scores — - - page

Table 10. Physician Behaviors and POEM Scores - — page

Table 11. Physician Behaviors and POEM Factor
Scores — - - page

Table 12a. Physician Behaviors and PSEM and POEM
Scores with a Criterion of .05 - - - page

Table 12b. Physician Behaviors and PSEM and POEM
Scores, with a Criterion of .05 -

page

Table 13. Physician Behaviors and PSEM and POEM
Scores without a Criterion of .05

page

viii

Chapter I

Introduction

Physicians report having different levels of confidence
in their practice of medicine when treating patients with
some types of problems than they do when treating patients
with other types of problems. One such medical problem
treatment area is obesity. On an anecdotal level, some
doctors are more confident than others about their practice
in treating obese patients.

One way of conceptualizing these different levels of
physician confidence is in terms of self-efficacy.
"Perceived self-efficacy" is one’s judgment of one's own
capabilities (Bandura, 1986). Self-efficacy judgments are
specific to an area. In terms of this study, therefore,
self-efficacy obesity treatment beliefs are beliefs that a
physician has about his or her own capability in treating
obese patients.

Knowing more about physician self-efficacy could be
important for several reasons. It would help doctors,
themselves, to practice more confidently, which would be of
direct advantage to them. A sense of mastery and
competence, in the area of ones professional activity, is

sought by practitioners who are in professional practice.

2

It would also be helpful to patients. Patient
self-efficacy with regard to disease and treatment has been
studied in a variety of health and illness areas. Among
these areas are: rheumatoid arthritis, cardiac
rehabilitation, smoking cessation and weight loss. High
self percepts of efficacy are found to be positively
associated with a plethora of different positive physical
and mental health outcomes, across illness area (O'Leary,
1986).

As the physician functions as a kind of medical
"significant other" to the patient in the arena of health
and illness, he or she is an important source of "efficacy
information" for the patient. Therefore, physician
self-efficacy and physician fostering of patient
self-efficacy, could be an important influence on patient
self—efficacy and, consequently, on positive patient health
outcomes.

This relationship, between physician self-efficacy and
patient self-efficacy, could be tested empirically in a
future study. It can only be tested, however, if one knows
more about physician self—efficacy.

The purpose of this study is to develop a measure of
primary care physician self-efficacy in the treatment of
obesity and to provide some evidence of its reliability and
validity. Further, after identifying "high" and "low"

self-efficacy doctors, this study aims to explore how

3
self-efficacy in this treatment area develops and how it is
maintained.

From a counseling practice point of view, knowing more
about what goes into high levels of self-efficacy and how
individuals develop these high levels could assist in
self-efficacy training of new physicians and in enhancing
levels of self-efficacy in practicing physicians. This
information would be useful to counseling psychologists
working in the field of health care in their roles as
teachers and consultants.

Overview of the Study

A first step was to survey a broad group of doctors
about their self-efficacy beliefs and their practice
behaviors with obese patients. The beliefs included those
about treatment outcome. These beliefs and behaviors were,
then, tested against theory.

The beliefs tested were formulated as expressions of
self-efficacy and the physician behaviors tested were
behavioral correlates of self-efficacy beliefs. These were
drawn, specifically, from Bandura’s self—efficacy theory as
applied to obesity treatment of patients in primary care
practice. Using these data, a measure of self-efficacy,
with some evidence of its reliability and validity, was
developed. In addition, these data were used to examine
what background or environmental variables, if any, are

associated with strong self-efficacy beliefs.

4

The next step was to discuss these beliefs with a small
number of doctors, some of whom were self-efficacious and
some of whom were not, using their scores on the study
measures as the criterion for self—efficacy. The purpose of
the interviews was to discover what thoughts, feelings,
ideas and experiences physicians have about themselves in
their treatment of obese patients. It was also to discover
how their obesity treatment beliefs developed and how they
are maintained.

The interview data were used to formulate a "model" of
physician self-efficacy in the treatment of obesity. This
model could be tested, at a later date, with another group
of physicians. The same approach could be applied to self-
efficacy beliefs in other areas of medical practice. It
could be partnered, eventually, with a measure of patient
self-efficacy and patient and physician self-efficacy could
be tested together.

To summarize, the focus of this project is whether
primary care doctors differ in their levels of confidence or
self-efficacy in treating obese patients and, if there is a
difference, to explore what it is based on. It is helpful
to know because, then, one could develop procedures to
enhance physician self-efficacy in this practice area. It
would also provide information about a potentially important
factor, which is currently missing, in the self—efficacy

research in the area of health and illness. More

5
information about physician self-efficacy would help to
illuminate patient self-efficacy and self—efficacy issues in

doctor-patient interactions.

Chapter II

Literature Review

Study Constructs

The following is an introduction to the main
theoretical base of the study, with an elaboration of some
aspects of self-efficacy theory which are salient to
studying this construct in the context of medical practice.

Self-Efficacy. The guiding theoretical construct of
this study is that of self-efficacy. This construct is
central to Albert Bandura’s Social Cognitive Theory.
Self-efficacy is a belief or perception that one has about
oneself: "Perceived self-efficacy is defined as people’s
judgments of their capacity to organize and execute courses
of action required to attain designated types of perfor-
mances." (Bandura, 1986, p. 391). It is a self-made
estimate of one’s own competence in a given area.

Self-efficacy beliefs are hypothesized to influence
one’s actions in diverse ways (Bandura, 1986). Rather than
being defined as stable personality characteristics or
static estimates of future behaviors, efficacy beliefs are
hypothesized to be process variables that function to
influence behavior in a number of different ways. They
influence which behaviors one chooses to do, how much effort

one expends on a behavior and the degree of persistence one

7
maintains in the face of obstacles or of experience which
disconfirm one’s efficacy belief. In this way, the beliefs
influence one’s present behavioral repertoire and the
development of self-efficacy perceptions which, in turn,
affect future behavioral choice, effort and persistence.

Self-Efficacy and Outcome Expectations. Judgments
about one’s efficacy are sometimes confused with, but are
conceptually distinct from, judgments about the outcomes or
consequences of one's behavior: "an outcome expectation is a
judgment of the consequence such behavior will produce."
(Bandura, 1986, p.391). According to Bandura (1986), both
constructs—-perceived self-efficacy and outcome
expectations-—are domain-specific. They are learned and
are, therefore, potentially modifiable.

In the context of this study, the domain is primary
care physician treatment of obesity. Physician
self-efficacy is the physician's belief that he or she can
successfully carry out the necessary behaviors to treat
obese patients. An example of such a behavior is that the
physician believes that he or she has sufficient information
about the treatment of obesity to make treatment recommen-
dations to patients. An outcome expectation, in this
context, is the physician's estimation that his or her
treatment will result in the desired consequences. An

example of such a consequence would be that the patient

8
learns about healthy eating and is able to change his or her
eating behaviors.

The relationship between the constructs of
self-efficacy beliefs and outcome expectations is not simple
one. On a substantive level, it is easy to confuse
judgments about one’s own capabilities and judgments about
the consequences of one’s behavior. Or to confuse, in this
case, physician treatment of an obese patient with patient
actual weight loss. They are not necessarily, of course,
unrelated. "Most outcomes flow from actions" and how one
behaves affects the outcomes one experiences" (Bandura,
1986, p 392). The kinds of outcomes one anticipates, in
one's thoughts, depends on one’s judgments of how well one
will be able to perform in a given situation. "One cannot
sever expected outcomes from the very performance judgments
upon which they are conditional." (Bandura. 1986, p.392).

There is not, however, a perfect correlation between a
belief in one’s own behavior and a belief about its expected
outcome and the relationship, between efficacy belief and
expected outcome is not a uniform one across areas of human
endeavor. It can be thought of as varying on a continuum.
On one end of this continuum, outcomes are inherent in
actions or are tightly linked to actions. In such areas,
outcome expectations closely parallel efficacy beliefs and
knowing about outcome expectations does not add appreciably

to our knowledge, over knowing about efficacy beliefs.

9
(Bandura, 1986, p.392). An example from the field of health
care would be physician self-efficacy beliefs and outcome
expectations in treatment for the surgical removal of a
non-metastatic malignant breast tumor. Knowledge about
physician efficacy beliefs about his or her own behavior are
closely linked to expectations for treatment outcome.

In other areas, beliefs about one’s behavior and the
consequences of one's own behavior are not as closely
linked. This occurs when no action can produce the desired
consequences, when consequences are only loosely linked to
quality or level of performance or when successful outcomes
are linked to a minimum level of performance. Physician
treatment of obesity is closer to this end of the
belief-outcome continuum. Some people lose weight with no
input from their doctors; others do not lose weight even
with conscientious and skillful treatment on their
physicians’ part. In this case, one would expect physician
beliefs about his or her own behavior to be not as closely
linked to expectations of the desired outcome.

To study physician self-efficacy in the treatment of
obesity, one would wish to know not only about self-efficacy
beliefs about treatment, but also about physician outcome
expectations for treatment.

WW
Theory. The relationship between one's own beliefs and

one's own behavior is also not a simple one. Self-efficacy

10
beliefs affect behavior in diverse ways and are not simply
"inert estimates of future action" (Bandura, 1986, p. 393).
They involve decisions involving choice of activity, how
much effort one expends on an activity and the degree of
persistence one maintains in the face of obstacles or
frustration. Self-efficacy beliefs influence emotional
reactions during actual and anticipated performances of
behavior. According to theory, factors which affect the
strength of the relationship between belief and behavior are
presence of requisite sub-skills and resources and whether
there are incentives for the behavior of interest. There
are also "dimensions" of self-efficacy which have
implications for behavior. "Level" refers to one’s estimate
of the degree of difficulty of the performance behavior;
"generality," to one’s estimate of the degree of specificity
of the domain of functioning and "strength," to the degree
of resistance of the belief to disconfirming experiences
(Bandura, 1986, p. 396).

Self—efficacy and Performance: The Relationship in
Practice. In discussing the relationship between beliefs
about behavior and actual behavior, Bandura talks about the
relationship between what he calls "performance efficacy
judgments" (Bandura, 1986, p. 395) and actual performance.
According to theory, "people’s beliefs...function as one set
of proximal determinants of how they behave, their thought

patterns, and the emotional reactions they experience in

ll

taxing situations." (Bandura, 1986 p. 393). "Reasonably
accurate appraisal of one’s capabilities is,..., of
considerable value in successful functioning." (Bandura,
1986 p.393). So a degree of accuracy between how one thinks

one performs and how one actually performs is highly
adaptive. Misjudgments in either direction have negative
consequences: over-estimating one’s abilities sets oneself
up for stress, failure and loss of credibility;
underestimating them sets one up for an unnecessarily
limited sphere of activity or aspiration.

The more significant the consequences are, for
misjudging one’s abilities, the stronger is the incentive to
judge them accurately. Temporal proximity between a
judgment and its corresponding behavior should improve
accuracy, according to theory. Clarity of the goals of
one's behavior and clarity of the level of performance of
one’s own behavior (i.e. completion of a feedback loop)
increase the accuracy of performance efficacy judgments.
When the goals of one’s behavior are ambiguous or when the
quality of one’s performance is unclear (both potentially
characteristic of doctors’ treatment of obese patients),
belief accuracy about behavior is reduced because the
process whereby perceived self-efficacy is used to regulate
effort is interrupted (Bandura, 1986). When judgments are

made in public, social pressure can also affect accuracy.

12
Public appraisals tend to be more conservative than do
private ones (Bandura, 1986).

In a practical application of the theory, such as is
being proposed here, the relationship between beliefs about
behavior and actual behavior is of particular interest. One
cannot necessarily observe people’s actual behavior; data
about behavior are often what people say about their
behavior and believe about their capabilities. Evidence
from empirical studies, where there are data about behavior
as well as about beliefs, suggests that judgments about
efficacy do, fairly accurately, predict subsequent
performance behaviors. This has been documented across a
variety of behavior domains including school achievement
(Bandura & Schunk, 1981), smoking cessation (DiClemente,
1981), physical stamina (Weinstein, Yukelson & Jackson,
1980), treatment of phobias (Bandura & Adams, 1977,)
treatment of arthritis (Lorig, Seleznick, Lubeck, Ung,
Chastian & Holman 1989) and career choices (Betz & Hackett,
1981).

Sources of Efficacy Information. People do not acquire
their perceptions of their own capabilities in a vacuum;
they learn through direct experience with their environments
and through "socially mediated" experiences (Schunk &
Carbonari, 1982, p. 234). The most powerful way people
learn about what they can do is by actually doing it, i.e.

having an authentic experience of mastery. Bandura terms

13
this "enactive," "performance" or "attainment" learning
(Bandura, 1986, p.399.) People also learn by modeling other
people’s mastery, either directly witnessed or imagined
through "vicarious learning." In addition, individuals
judge their abilities from what other people tell them about
themselves, through verbal persuasion, exhortation and
encouragement. In "vicarious" and "exhortative" sources of
efficacy information, the relationship between influencer
and influencee is of importance. One's physiological state,
through autonomic reactivity, also provides one with
information about one’s capabilities are. Somatic arousal
signals doubts about one's capabilities; symptoms of stress
can hamper performance.

A Developmental Perspective on Self-Efficacy. Bandura
gives only very general developmental guidelines in his
"Developmental Analysis of Self-Efficacy" (Bandura, 1986).
Children’s experiences with their environment provide the
initial basis for developing a sense of "causal efficacy"
(Bandura, 1986, p.414). For the infant, the arena of
personal efficacy is the immediate, physical environment.
Her or his experience centers on controlling sensory
stimulation by manipulating objects in the environment,
including the attentiveness of the infant’s care-givers.
These early experiences of personal efficacy are central to
the development of one’s social and cognitive competence.

The infant and the environment are seen as operating

l4
reciprocally, with parental responsiveness increasing infant
competence and infant capabilities eliciting care—giver
responsiveness (Bandura, 1986).

As a child’s environment expands, socially and
psychologically, her or his representational capabilities
develop and peers, within a context of social
comparativeness, assume an increasingly important role in
the development of self-percepts of efficacy. In their
relationship with siblings at home and peers at school,
children form "evaluative habits" (Bandura, 1986, p.416).
These habits influence the choice of "comparative referents
in self-ability evaluations in later life" (Bandura, 1986,
p.416). Self-efficacy appraisals and the verification of
appraisals, particularly in the arena of cognitive
self-efficacy, develop in a school context of social
comparativeness. Children judge their own capabilities with
increasing accuracy as they get older.

Throughout development, each period provides the
individual challenges in "coping efficacy" (Bandura, 1986,
p. 417) with developmental tasks. In late adolescence and
early adulthood, there is a focus on mastery in the areas of
ability to form intimate relationships and make vocational
choices. Self-efficacy is considered an important motivator
for career aspirations and career accomplishments (Bandura,
1986 and Betz & Hackett. 1981). "Coping efficacy" is not

elaborated on further.

15

Middle years are characterized by routines which
stabilize self-percepts of efficacy in the areas of
relationships, especially generative relationships and of
career. It is also the time when one confronts the
limitations of one’s capabilities. Old age is seen as a
time of efficacy reappraisal and a declining sense of
self-efficacy may be instrumental in a process which results
in declining cognitive and behavioral functioning (Bandura,
1986; Rodin, 1986).

These life span guidelines for the development of
self-efficacy are quite vague. In general, Bandura
advocates the use of a "micro-analytic research strategy"
(Bandura, 1977, p. 205) to explore the origins of
self-efficacy beliefs. Specific behaviors, not global
behavioral domains, are targeted for evaluation and measures
are designed that are tailored to these specific behaviors.

The construct of self—efficacy has gone through
refinement and has been tested in a variety of areas since
it was proposed by Bandura (Bandura, 1977). It is related
to other forms of self-referent thought, such as
"self-concept," "self-esteem," "effectance motivation," and
"outcome expectancy" theories (Bandura, 1986). Yet self-
efficacy is distinct from other constructs in that it is not
"global" and "stable," but situation-specific and variable.
Self-efficacy is not invariably tied to self-worth and

viewed as intrinsic. Consequently, self-efficacy can be

16
measured independently from the behavior it affects, but is
not measured independently of context.
Self-Efficacy Theory and Research on Health and Illness

Much current research in clinical and counseling
psychology is concerned with a relatively small number of
questions about the role that basic cognitive processes play
in human behavior. The construct of "self-efficacy" has
become more frequently used since Bandura’s 1977 article in
the Psychological Review: "Self-efficacy: Towards a Unifying
Theory of Behavior Change" (Maddux & Stanley, 1986).

The interdisciplinary field of behavioral medicine has
emerged and grown during approximately this same period of
time. Illness and health are increasingly seen as
multidetermined; social and behavioral factors, as well as
physiological ones, are seen as determining an individual's
health or illness status. In the framework of social
cognitive theory, social, behavioral, environmental and
physiological factors are seen as reciporically interacting
and "health" (or "illness") is seen as an "end state" or
outcome of their mutual influences (O'Leary, 1985).

Behavioral medicine has accrued an empirical research
base that is spread across the knowledge base of a range of
professional disciplines and patient populations (i.e.
diagnostic groups.) Behavioral medicine uses a variety of
research methodologies, hypothesizes a myriad of factors as

predictors and values a plethora of physiological and

l7

psychological factors as health outcomes. Independently
evolving bodies of research have focused on different
aspects of health and illness (Ell, 1985/86, Kaplan, 1990).
Research foci include factors that reduce the risk of
disease or are associated with the promotion of health or
quality of life; and factors that ameliorate the symptoms
(physical and psychological) of illness or of treatment and
are associated with optimal functioning. These latter
factors are of particular significance in the area of
chronic disease or condition. "Self-efficacy" has been
suggested as an integrating construct for this diverse field
by several reviewers (Ell, 1985/86, O’Leary, 1985). There
has developed a body of research in the field of behavioral
medicine that uses self-efficacy as its guiding theory.

Health Promotion. Health promotion or health
enhancement is an area where self-efficacy theory has been
used to guide and evaluate interventions designed to
initiate and maintain health behavior change. In smoking
cessation treatment and smoking cessation relapse, training
to increase self—efficacy to resist smoking increases with
treatment and is positively associated with treatment
effectiveness, at the time of treatment (Condiotte &
Lichtenstein, 1981) and also with the long term
effectiveness of smoking cessation treatment (DiClemente,
1981). Self-efficacy treatment assists smokers to stop

smoking and not to start again.

18

Eating Disorders. Self efficacy theory has been used
to guide the study of eating disorders, the most common of
which is obesity (O'Leary, 1985). The effect of
pre-existing and experimentally manipulated levels of
self-efficacy on weight loss have been examined (Weinberg,
Hughes, Critelli, England & Jacobson, 1984). Cogni-
tive-behavioral principles, such as proximal goal-setting,
manipulation of environmental cues and modeling, are a
common part of behaviorally—based weight loss programs
(O’Leary, 1985).

gaipy_ Self-efficacy theory has also been used to
examine the relationships among social, behavioral and
biochemical aspects of acute and chronic pain. Studies have
focused on the influence of emotional arousal in pain
perception (Craig, 1984); the interpretation and context of
arthritic pain (Lorig, Chastian, Ung, Shoor, Holman, 1989
and Shoor & Holman, 1985); the past history of pain
(Fordyce, Fowler, Lehmann, DeLateur, Sand, & Treischmann,
(1973); attentional focus in pain control (Turk,
Meichenbaum, and Genest, 1983); perceived control over the
pain stimulus (Neufeld & Thomas, 1977 and Reese, 1982); and
affect of ethnic background on pain perception (Weisenberg,
Kriendler, Schachat, & Werboff, 1975). Self-efficacy has
been a useful construct which helps in understanding the

effectiveness of cognitive and non-cognitive coping

19

strategies, how medications and placebos work and the role
of affective arousal in pain.

Cardiac Rehabilitation. Self-efficacy theory has also
been used to guide research on medical patients following a
myocardial infarction who participated in a multifaceted,
interdisciplinarily-administered program, known as "cardiac
rehabilitation" (Ewart, Taylor, Reese & DeBusk, 1984).
"Cardiac efficacy" (the patient’s belief in the capability
of his or her own heart,) was found to be a stronger
predictor of patient recovery than many more conventional
medical parameters (Bandura, in press). An interesting
study also included a measure (and manipulation) of the
patient’s spouse’s perception of the patient’s cardiac
efficacy (Taylor, Bandura, Ewart, Miller & DeBusk, 1985).
The patients’ wives were provided with modeling and mastery
experiences of their own cardiac capability. These
experiences of their own capability increased the wives’
perceptions of their husbands’ cardiac efficacy. The wives’
ratings for their husbands’ cardiac efficacy were more
congruent with husbands’ ratings of their own cardiac
efficacy than were the ratings for wives who did not have
modeling and mastery experiences themselves. Combined
cardiac efficacy ratings (the patient’s and the wife’s) were
the most consistent predictor of patient cardiovascular

functioning.

20

Chronic Disease. Adherence to medical regimens is an
issue of particular importance to patients with chronic
disease, which, by definition, cannot be cured. Chronic
disease can sometimes, however, be "managed" and patient
behavior is often of prime importance of this management.
Kaplan, Atkins & Reinsch (1984) have studied patient
self-efficacy about adherence to exercise programs and
Maddux & Rogers (1985) studied behavioral intentions to stop
smoking. In patient behavior change programs, where there
is a short term loss in return for a long term gain,
focusing on outcome expectations as well as self-efficacy
beliefs is of importance (O’Leary, 1985). If a patient is
to "sacrifice," it must be seen to be worthwhile.

Lorig et al. (1985), at the Stanford Arthritis Center,
have accrued a small body of research on self-efficacy and
arthritis. They have incorporated self-efficacy training
and arthritis self-efficacy measures in to their health
education program for arthritis patients. Using a "modified
grounded theory" approach to data analysis in their
research, they found self-efficacy to be a significant
predictor of study outcomes (Lorig et al., 1989). Using a
modified grounded approach was initiated by their discovery
that they had an educationally-oriented health program,
which clearly helped participants, but whose participants
did not improve significantly on the study’s educational

measures .

21

Self—efficacy has been a useful construct in many
substantive areas in the research on health and illness.
Self-efficacy has been used, primarily, by psychologists
working on problems in the medical arena and has fostered
some conceptual integration of a highly diverse body of
knowledge. This research has been patient (not health
provider)~focused.
Teacher Self-Efficacy

The only developed body of research on self-efficacy of
those who provide a service to others is that of teacher
self-efficacy (Bandura, 1986; personal communication,
10/91). There have been some limitations, however, in the
way that teacher self-efficacy has been studied.
Self-efficacy and outcome expectations have not been
maintained as conceptually distinct. Teacher self-efficacy
has been measured in terms of student outcomes (Fuller,
Wood, Rappoport & Dornbusch, 1982; Dembo, Gibson, 1985; and
Newmann, Rutter, & Smith, 1989). Perceptions of
self-efficacy have been viewed as "stable" and "global,"
rather than variable and domain-specific. A teacher’s
self-efficacy is assumed to be a "fixed" quality of the
person, rather than one which may vary, within an
individual, across situations (Raudenbush, in press).

A recent study (Raudenbush, in press), looked at
teacher self-efficacy independent of student outcome

measures. Using an hierarchical linear modeling design,

22
self-efficacy variation in teachers was examined not only
between subjects but also within subjects across teaching
situations. "Intra" teacher practice variables included
preparation of the teacher and grade level of the students
and academic track of the students taught; "inter" teacher
variables included demographic characteristics and
organizational environment variables.

This way of conceptualizing variation highlights
differences, within individuals, between situations and
takes in to account a complex array of factors that
influence a multidetermined outcome like professional
practice. It focuses on the provider of a service, not on
its recipient.

Primagy Care Physicians and Health Behavior Change

Amongst the different ways of conceptualizing the
practice of medicine is whether it is oriented primarily
around the disease or primarily around the patient. In
recent past history, disease—oriented medical practice has
been in ascendance and has successfully guided Western
medicine: the physician does something "to" the patient,
prescribing substances or performing a procedure (Weiner,
1989). Successes of this approach are obvious, e.g. in
modern surgical techniques and antibiotic treatment for
infection. Weaknesses of this approach are also

increasingly clear in areas that are important in medical

23
practice today: prevention and management of chronic
disease and condition.

Chronic Disease. Chronic disease cannot be cured, but
it can often be controlled. However, usually it cannot be
controlled (or managed) substantially by the physician doing
something "to" the patient. The treatment is a persevering
adherence, by the patient, to a treatment regimen. Patient
behavior, not physician behavior, is of primary importance.
The patient must do something (e.g. take blood pressure
medication) and/or must not do something (e.g. eat foods
high in salt). Preventive health practices are also highly
dependent on patient behavior. Management of chronic
conditions and fostering of preventive health behavior need
patients who are actively motivated and regimen-adhering and
physicians whose practice is oriented to the patient,
particularly, education and engagement of the patient in
health practices, not a practice which is oriented to the
disease (Tresolini, 1992, Weiner, 1989).

Currently, patient health education is considered an
appropriate part of medical practice; doctors are encouraged
through their own professional journals and by
recommendations from government reports to counsel their
patients about making behavioral changes to reduce chronic
disease risk factors (Bartlett, 1987; U.S. Prevention
Services Task Force, 1989). Health promotion material is

increasingly a part of the curricula in medical schools

24
(Bloom, 1989, Sullivan, 1990). Practicing physicians them-
selves consider it to be important (Sobal, Valents, Muncie,
Levine & DeForge, 1985).

Medical Education. However, formal training in the
education and engagement of patients in health behaviors and
behavior change, and opportunities to practice these
behaviors under guidance or to have role models for these
behaviors while training do not appear to be routinely
available to all medical students or even all interested
medical students. Medical students state that they are
interested, have adequate knowledge, but have difficulty
translating their knowledge into practice (Tresolini, 1992).

Current Practice. Practicing physicians also express
discomfort and lack of confidence in their roles of health
educator and engager/initiator of patients in health
behavior change (Becker & Janz, 1990, Orlandi, 1987). In a
survey of primary practitioners, smoking cessation was
considered a "very important" or "important" health behavior
to address with patients (100% of respondents), followed by
taking precautions around hazardous materials, (98%),
avoiding excess caloric intake (97%) and eating a balanced
diet (97%). Limiting alcohol is so considered by 83%,
exercising regularly by 72% and getting adequate sleep by
70% (Sobal et al, 1985). Of all of their patients’ daily
activities which may increase risk for disease or exacerbate

existing disease or condition and which are potentially

25
modifiable, concerns about obesity and adequate diet are
very highly ranked by doctors in practice.

There is a high degree of consensus by medical
educators, medical students and practicing physicians that
patient health behaviors and health behavior change is an
important part of medical practice, for both preventive
health practices and for the management of chronic disease
or condition. It is also an area in which both students and
doctors express discomfort and a lack of confidence.

.Qbeeilix

Obesity is a widespread condition; estimates of its
prevalence range from 15% to 50% (Bray,1976). Mild to
moderate obesity is usually defined as 10% to 20% above
average and gross obesity as 30% above average weight
(Wilson, 1984). Prevalence is inversely proportional to
socioeconomic status, increases with age and is more common
in some ethnic groups (Stunkard, 1975). Obesity has serious
consequences, medically and psychologically. Gross obesity
is clearly shown to have deleterious medical consequences in
the areas of hypertension, hyperlipidemia, diabetes,
pulmonary and renal problems, osteoarthritis and recovery
from surgery (Wilson, 1984). The association of obesity and
cardiovascular disease is usually considered the most
critical area of medical concern.

However, the extent to which deleterious consequences

are directly the result of obesity, as an independent

26
factor, or are the result of a combination of obesity with
other risk factors, is unclear. It is also unclear what the
deleterious consequences are of mild to moderate obesity
(Brownell, 1982). Although there is consensus that mild to
moderate obesity does not have as strong an association with
disease, as does gross obesity, this group still appears to
be at risk for chronic disease related to health habits
(e.g. poor diet). Weight reduction is associated with
reducing these behavioral risk factors (e.g. improving
diet).

Psychological consequences of obesity include social
stigmatization and dieting. Even successful dieting, i.e.
that which is associated with actual weight loss is
associated with undesirable emotional sequelae such as
depression, anxiety, fatigue, irritability and preoccupation
with weight and diet (Wilson, 1984). Motivation for weight
loss is powerfully associated with social concerns and
vanity (Stuart & Jacobson, 1979), rather than health
concerns. Social benefits of weight loss may indirectly
promote improved personal health habits (Wilson, 1984).

Treatment. This common, serious, physical and
psychological health problem is quite refractory to
treatment. Most obese people do not become thin, either in
the short term or the long term (Wilson, 1984; Brownell,
1982). "Obesity has remained resistant to most treatment

methods, and even the most effective procedures have

27
produced relatively modest successes." (Wilson, 1984, p.
667). Attrition rates in treatment programs range from 50%
to 80% Multifaceted treatment based on behavioral
principles appears to be the most effective; moderate weight
loss (10 to 11 lbs.), in the short term, is the result,
across behavioral treatment type (Wilson, 1984). Increasing
physical activity seems to help, for short and long term
weight loss and for reducing the negative emotional sequelae
of both obesity and of dieting.

Very low calorie diets and medication are found to have
some success when used in conjunction with a behavioral
program. Overall, behavioral programs show the most success
with maintenance of weight loss for up to one year. Results
for longer term studies are meager, but in general, do not
show good results over a span of years (Wilson, 1984). A
problem with evaluating the results of long term weight loss
maintenance is the uncontrollability of inevitable confounds
(Brownell, 1982). Even the most effective program may not
be as powerful an influence, as many other factors, on an
individual’s weight over the life span.

Obesity is a chronic health problem. It’s cause is
largely unknown; there is no cure. It has deleterious
physical and psychological consequences. There are a number
of available treatments which are helpful in managing this
chronic problem (i.e. treatments which produce some relief

of symptoms for some people, some of the time.) Although

28
overweight causes physical and social liabilities, so does
dieting.

Obesity is a condition which, for many people, needs
life-long management. Like other chronic health problems,
there needs to be a management plan worked out between the
patient and his or her doctor, to optimize the balance
between the risks and benefits of both the problem and
management of the problem.

Research Questions and Hypotheses

Ideally, in research on self-efficacy, one measures
beliefs and behavior separately at, minimally, two points in
time. Specific behaviors are sampled, from a larger area of
interest and are arranged hierarchically from least to most
difficult (arranged by "level," according to theory).
Measures are taken and, then, an intervention designed to
manipulate the subject’s self-efficacy beliefs is implemen-
ted. Measures of beliefs and behaviors are re-taken
(Bandura, 1986). In field research on health and illness,
this ideal design is not possible in many areas of clinical
interest, for ethical as well as practical reasons.

Although Bandura’s theory of cognitive processes is
"social," self-efficacy research has focused on patient
self-efficacy beliefs alone and not on the beliefs of those
who inhabit the patient’s social environment. Research has
not studied self-efficacy "interactions," among patient

beliefs and those in the patient’s interpersonal

29
environment. Significant health care providers, such as
physicians, have not been studied, either in terms of their
own self-efficacy or its influence on patients.

Recent Studies. There is a single published study that
deals explicitly with physician self—efficacy and
investigates whether physician self-efficacy and physician
outcome beliefs are associated with physician practice
behaviors (Gottlieb, Mullen & McAlister 1987). In this
study, self—efficacy beliefs and outcome expectations were
each measured on a single item, 4-point scale. The study’s
substantive area was various kinds of substance abuse
(nicotine, alcohol and over the counter and illicit drugs);
outcomes were three practice behaviors: history-taking and
counseling, which represented efficacious behaviors and
referral, which represented an inefficacious behavior. More
recently trained doctors (across specialty area) and
internists reported being more likely to engage in the
targeted self-efficacy behaviors with their patients. G.P.s
were the least likely. Family practitioners were most
likely to engage patients in counseling; general
practitioners were least likely.

The method of the Gottleib et al study was a
quantitative analysis done on data from a questionnaire
survey mailed to primary care doctors (general internists,
family practitioners and general practitioners), with a

return rate of approximately 29%. Medical specialty and

30
year of medical school graduation were found to have the
greatest predictive power for practice behaviors.
Self—efficacy beliefs and outcome expectations were also
significant predictors of practice behaviors. An
interesting finding was that self-efficacy expectations were
a stronger predictor than outcome beliefs, across all
subjects.

Weaknesses of this study include possibly overly-simple
instrumentation (a single question may be a sufficient
measure of a physician’s self—efficacy in a practice area,
but this was not tested); and an extremely diverse
substantive area of practice. Comparing knowledge and
skills, (not to mention attitudes) for treating cigarette
addiction and heroin addiction are like comparing apples and
oranges. The focus of the Gottleib study illuminated
differences in treatment for different substances within the
broad practice area of drug abuse treatment rather than
differences among individual physicians in their
self-efficacy beliefs. Analyses were done on doctors as a
group, not on individual doctors, across substance abuse
areas.

A study about the relationship between physician
decision making and practice behaviors (Gruppen, Wolfe &
Stross, 1990) asked about physician "confidence" in treating
four common medical problems seen in primary practice.

Vignettes depicting treatment and management of rheumatoid

31
arthritis, asthma, chronic obstructive pulmonary disease and
fever of unknown origin were rated as to physician
preference for treating the patient themselves, pursuing
consultation before treating themselves or referring the
patient to another doctor for treatment. The relationship
between targeted self-efficacy behaviors and board
certification status and administrative practice
characteristics of individual doctors was examined. Those
who were board certified were more likely to seek
consultation via the literature and then treat themselves
and wished, in general, to retain a higher degree of primary
care control of their patients. Those in group practice
sought informal consultation with colleagues and then
treated themselves. Doctors in solo practice relied on
formal consultation. Physicians were mail-surveyed and data
were quantitatively analyzed on the questionnaires which
were returned and useable (18%).

The cognitive processes of individual physicians were
not studied, but as in the Gottlieb et al. study, doctors
were studied, as a group, within the larger administrative
contexts of their practices and their specialties. They
were studied more from a sociological point of view than a
psychological one. Differences in physician practices
between diagnostic patient groups were not highlighted as in
Gottlieb et al. This study emphasized differences in

administrative practice arrangements and board certification

32
status and how they affect practice behaviors (or vice
versa). An alternative way of conceptualizing practice
contexts are as sources of efficacy information for doctors.
They are the settings where enactment, modeling and
exhortative and somatic experiences and behaviors take
place.

Tresolini and Stritter (1992) used case study
methodology to examine self-efficacy in "patient health
education for health promotion" for final year medical
students intending to practice in primary care areas. Using
a semi-structured format, researchers interviewed students
about their educational experiences during medical school
(both formal and informal), in gaining knowledge about 1.)
the benefits of health promoting patient behavior, 2.)
conveying this information to their patients and 3.)
engaging patients in the educational process. 28 students
(one third of their class) were interviewed. 6 faculty
members involved in health promotion-related teaching were
also interviewed and a documentary review was conducted on
materials such as course syllabi and curriculum guides.

Interviews and other materials were submitted to a
qualitative analysis, using grounded theory and case study
methodology. A framework from self-efficacy theory was used
to focus on relevant data, organize categories and define
alternative explanations as well as to provide a descriptive

framework for "cases". Content areas from theory which

33
interviews covered were: sources of efficacy information,
appraisal of this information and "many aspects of student
perspectives and learning experiences" (Tresolini &
Stritter, 1992). A 9-item questionnaire, measuring
self—efficacy in each of three health promotion areas
(smoking cessation, "heart-healthy" nutrition, and exercise
and fitness) was given to students following the interview.
Questionnaire items, within area (3 items each), were graded
along a continuum of difficulty (or self-efficacy "level"),
from easiest (e.g. believing that one has the health promo-
tion information oneself), to harder (believing that one can
transmit it to patients) to hardest (believing one can
engage patients in the educational process).

Three "patterns of medical student experiences" or
"cases" were formulated on the basis of the qualitative
analysis; descriptive statistics on the self-efficacy
measure showed a difference associated with each pattern, in
the expected direction. Students with more health promotion
experience and exposure to models scored higher on the
self-efficacy measure. For students in two of the three
"patterns", self-efficacy scores also varied in the expected
direction for behaviors at the different difficulty levels;
self-efficacy was higher for behaviors the researchers
considered "easier". For the students in the third group
(those with the most experience, most exposure to models and

highest interest), there was no difference in self-efficacy

34
scores. Theirs were uniformly high across level of behavior
difficulty, possibly showing a "ceiling effect" of the
measure.

Conclusion. Because the psychological context of
medical practice is complex and not well-studied, and
because research in the area of physician self-efficacy is
very sparse and that which exists does not focus on
psychological variables, this study begins at a basic level.

The study focuses on identifying physicians who
consider themselves to be efficacious in the treatment of
obese patients and exploring the attributions these doctors

make for their efficacy in this practice area.

Goals

A first goal of the study is to develop an instrument
to measure physician self-efficacy in the treatment of
obesity, with evidence of its reliability and validity. A
second goal is to examine whether there are demographic
characteristics or practice characteristics of physicians
that are associated with strong and weak self-efficacy in

the treatment of obesity.

Research Questions

1.) Assuming that doctors vary in their self-efficacy in

treating obese patients, can this difference be measured

35
reliably and validly? A reliability coefficient of .70 is
sought.
2.) What attributions do physicians make for the development

of their own self-efficacy in this treatment area?

Hypotheses

Scores on the self-efficacy measure are associated with
physician practice behaviors, in a manner which is
consistent with self-efficacy theory. Specifically, that
doctors with stronger self efficacy beliefs in the treatment
of obesity will:

Hypothesis 1. have a higher proportion of obese
patients in their practices,

Hypothesis 2. discuss obesity with a higher proportion
of their obese patients,

Hypothesis 3. initiate discussion of obesity with a
higher proportion of their patients,

Hypothesis 4. have a higher proportion of their
patients on some kind of weight control program,

Hypothesis 5. consult more with their medical
colleagues and the medical literature about their obese
patients,

Hypothesis 6. refer fewer of their obese patients to
other doctors for primary care,

than will doctors with weaker self-efficacy beliefs.

Chapter III

Method

Introduction
A combination of quantitative and qualitative

methodologies was used for studying physician self-efficacy.
Doctors with strong and weak self-efficacy were identified
using quantitative methods; the study of each group was by
quantitative and qualitative methodology. The qualitative
base for this study was twofold. "Modified grounded theory"
methodology (Glaser & Strauss, 1967) was used. Rather than
initiating data collection and analysis with an atheoretical
eye, collection and analysis were done through a conceptual
lens from self-efficacy theory. Self—efficacy constructs
were grounded in data from the doctors under study. Case
study methodology (Yin, 1989) was also used to study a small
sample of doctors. Statistical and measurement principles
from Mehrens and Lehmann (1984) and Glass and Hopkins (1984)
provided the basis for the quantitative methodology used in
the study.
Design

The study had two stages. In the first stage, an
instrument was developed that measured physician
self-efficacy in the treatment of obesity. It was given to

a large group of doctors and a distribution of scores was

36

37
obtained. Demographic and practice information from the
doctors was also gathered. Analyses were done on the scores
to evaluate the self-efficacy instrument and to provide
empirical evidence of the anecdotal difference among doctors
in their level of confidence in treating obese patients.

Self-efficacy scores were examined to see if physician
beliefs predicted physician practice behaviors. Demographic
and practice variables were tested for their association
with self-efficacy and outcome expectations scores.

In the second part of the study, the self-efficacy
beliefs of a small sample of individuals (doctors with high
and low self-efficacy scores) were examined through
interviewing, using partly structured and partly open-ended
questions.

Subjects

The subjects for this study were primary care
physicians practicing in the Lansing, Michigan area and in
the Rapid City, South Dakota area, who were listed on
professional and institutional rolls and other public
listings, e.g. the telephone book and newspaper
advertisements. The South Dakota cohort included doctors in
Nebraska, Wyoming, Montana and North Dakota, in addition to
those in western South Dakota. All doctors, with M.D. or
D.O. degrees who specialize in family practice or general
internal medicine were recruited as study subjects. In

addition, general practitioners in the Rapid City area were

38
recruited. This group numbered approximately 365. Attempts
were made to recruit a sample which was as representative as
possible, of primary care doctors practicing in the states
of Michigan and South Dakota. It included doctors
practicing in cities, in the surrounding suburban and rural
areas, in small towns and, in the South Dakota cohort,
doctors practicing under the auspices of the Veteran’s
Administration, the Indian Health Service and the U.S. Air
Force. Doctors were included with a high variation of level
of university-affiliation (including none) and a high
variation of type of practice arrangement, ranging from solo
private practice and small group private practice to
university clinic based practice, HMO practice, and U.S
government practice.

For the first part of the study (testing the
instrument,) the study sample was the group of every
physician who could be readily identified in the two areas.
For the second part of the study (studying individual high
and low self-efficacy doctors,) a small sample was drawn
from the physicians who had returned their questionnaires.
The sample size for the qualitative part the study was
determined, as the data was collected and analyzed,
according to principles of qualitative data analysis
(Gorden, 1980).

DeveIOpment of the Physician Self-Efficacy Measure

The Physician Self-Efficacy Measure sought to answer

39
the question: are there differences between doctors in their
self-efficacy in treating obese patients? In addition to
measuring self-efficacy, its purpose was to designate
individuals for further study. Items for the Physician
Self-Efficacy Measure were of two kinds and were given as
two separate questionnaires. One focused on self-efficacy
beliefs and used as its question stem: "How confident are
you that you could:" The other, the Physician Outcome
Expectations Measure, focused on outcome expectations and
used as its question stem: "Indicate your agreement with the
following statements:".

Items for each measure were drawn from principles of
self—efficacy theory as applied to practice beliefs and
outcome expectations in the area of obesity treatment. The
relationship between the measures was examined but the two
were maintained as two separate measures throughout the
analyses.

A brief measure of physician demographic information
and self-reported behavior, the Physician Practice Measure,
was also given. This asked the doctors about themselves and
their actual practice behaviors with patients. Questions
were single—item measures that pertain to substantive
practices that illustrate self-efficacy principles of
behavioral choice, effort and persistence. Copies of the

three measures are Appendices A, B and C.

40
Evaluation of the Self-Efficacy Measure. Outcome
Expectations Measure and Physician Practice Measure

Reliability. Test reliability is the consistency
between two measures of the same thing. Estimates of a
test’s reliability are made by evaluating its stability,
equivalence and internal consistency. (Mehrens & Lehmann,
1984.) The Physician Self-Efficacy Measure and Physician
Outcome Expectations Measure were "criterion-referenced"
measures, with response repertories that were specified by
theory (i.e. beliefs about one’s confidence in treating
obese patients and beliefs about the outcome of obesity
treatment). The measures tested the correspondence between
what an individual stated (the doctor’s self—reported
efficacy and outcome beliefs) and an underlying theoretical
continuum of efficacy ("perfectly efficacious" beliefs and
"optimally strong" outcome expectations.) The score of a
"criterion-referenced" measure is interpretable according to
"specified performance standards" (Mehrens & Lehmann 1984)
which are, in this case, beliefs about particular behaviors
and the outcomes of these behaviors.

The measures were also "norm-referenced," in the sense
that scores were interpreted compared to a distribution of
scores, not only compared to a self-efficacy criterion.
Inferences made on the basis of the measures, however, were
limited to a narrow subdomain, defined by the criterion

(i.e. primary care physician treatment of obesity), rather

41
than a broad domain, as they would have been in a primarily
norm-referenced test (Mehrens & Lehmann, 1984).

Evidence for test stability is not as essential for the
reliability of criterion-referenced measures as it is for
norm-referenced tests, (Mehrens & Lehmann, 1984). Test-
retest reliability was, therefore, not evaluated. Evidence
for a criterion-referenced test’s internal consistency is
important (Mehrens & Lehmann, 1984) and this can be
estimated, on a single test administration, by calculating a
consistency coefficient (e.g. Cronbach's alpha,) which is an
index of the homogeneity of items, or the degree of
correlation between single test items and the total test
score (Mehrens & Lehmann, 1984.) An internal consistency
coefficient was calculated for the self-efficacy and the
outcome expectations measure to evaluate the reliability of
each.

The Physician Practice Measure is a series of single
items and, therefore, no test for reliability was
appropriate for them.

Validity. Test "validity" is the extent to which the
test measures what it is supposed to measure. The kinds of
inferences that can be made about an individual from his or
her test score depends on the test’s validity. If one
wishes to make inferences about a person's performance other
than that which is measured ("predicting" performance, in a

sense) criterion—related validity evidence is of particular

42
importance (Mehrens & Lehmann, 1984). In this instance, one
was trying to "predict" physician behavior from a test score
based on items culled from self-efficacy theory.

Content-Related Validity. To evaluate the content
validity of a measure, one examines how well the test items
serve as a sample of the domain of the behavior of interest.
In test development, the task is to define the content
domain and universe of situations as specifically as
possible for a finite set of behavioral objectives, then to
devise a sampling process of this finite set to choose test
items. These items are then inspected by experts for
adequacy, precision, comprehensiveness and clarity (Mehrens
& Lehman, 1984). The preliminary drafts of the measures for
this study were evaluated by both content and methodological
experts. They were Deborah Feltz Ph.D. (a self-efficacy
researcher and committee member); Bertram Stoffelmayr, Ph.D.
(a self—efficacy researcher and medical educator); Ruth
Hoppe, M.D., (a general internist and medical educator), and
John Strandmark, M.D. (a general internist with a practice
specializing in the treatment of obesity).

Criterion-Related Validity. Criterion-related validity
is evidenced by empirically comparing scores on the measure
of interest to another, independent measure (i.e. a
criterion) of the thing one is interested in measuring.
Self-efficacy scores on a second, independent measure of

physician self-efficacy or observations of physician

43
behavior would provide criterion-related validity evidence
for study measures. In the absence of having an established
instrument which has accrued validity evidence or being able
to observe physicians in practice, I proposed the following
criterion—related validity estimates.

Physician Practice Measure items 10a. through 10d. and
11a. and b. provided evidence of criterion-related validity
for the measures. These items asked physicians about
experience with actual patients in practice. Reported
behaviors from the recent past were considered to be
independent, external indicators of physician beliefs.

Construct-Related Validity. Construct-related validity
evidence refers to the degree one is able to infer
constructs in psychological theory based on a test score.

If a measure has construct validity, there is evidence that
scores will vary as the underlying theory predicts (Mehrens
& Lehmann, 1984.) Multiple indices of validity and
reliability evidence are needed to provide information about
a measure's construct validity. These multiple sources of
evidence, ideally, provide convergent information which
supports the operationalization of an underlying
psychological construct.

Heterogeneity of subjects and of item difficulty
increase construct validity. Heterogeneity of subjects with
regard to geographic area of residence, medical degree and

specialty, practice type, setting and size; and with regard

44
to age, gender, and medical school graduation date was part
of the sampling procedure.

Self-efficacy theory, itself, served as the foundation
for validity of the measures which identify doctors with
strong and weak self-efficacy. Constructs from the theory
were not independent criteria because they provided the
conceptual basis from which items were drawn determining
measure score. But the larger underlying constructs of
self-efficacy and of outcome expectations served as criteria
for the subparts of the theory and vice versa. Each test
item which dealt with a single, specific aspect of
efficacious behavior, (e.g. confidence in diagnosing
patients whose obesity puts them at health risk) provided
validity evidence for the test as a measure of the larger,
underlying construct of self-efficacy in treating obese
patients. Beliefs about specific treatment outcomes, (e.g.
patients who try, can lose weight) provided validity
evidence for the test as a measure of the larger underlying
construct of outcome expectations in treating obese
patients.

This relationship, between items pertaining to specific
behaviors and a larger, underlying, construct is also part
of the evidence for the test’s content validity and content
reliability. Content validity and reliability are related
to "generalizability" in a broad sense (Mehrens & Lehmann,

1984.) One is, in each case, trying to generalize: from a

45
specific behavior to behavior in a diversity of situations,
from a test item to a total test score and finally, from a
paper and pencil indicator to behavior in a real life
situation. As different types of validity evidence are
interrelated with each other and with reliability evidence,
conceptually as well as operationally (Mehrens & Lehman,
1984), the relationship between the subparts of a theory to
the theory as a whole, and between individual test items and
to the larger, underlying construct make up the links in a
logical chain.

Variability in scores is also a part of the validity
evidence for the measures. Scores indicating efficacious
and inefficacious doctors serve as evidence for construct
validity for each other, based on self—efficacy theory.
Strong self-efficacy beliefs and outcome expectations serve
as a criterion (by way of contrast) for weak self-efficacy
beliefs and expectations.

The Development of the Self-Efficacy Interview

The qualitative part of the study addressed the
question of how self-efficacy beliefs develop and are
maintained by focusing on the doctors’ own perceptions of
themselves. The interview was highly structured in its
underlying form although not all questions were asked in a
highly structured way.

The underlying structure of the interview was based on

self-efficacy theory. It dealt with sources of efficacy

46
information (i.e. how beliefs are formed) by asking
physicians about their experiences of "enactment,"
"modeling," "exhortation" and affective response, during
their training and in their current practice. It also
sought to discover spontaneous sources of self-efficacy
information which are not a formal part of the theory. The
interview dealt with specific aspects of self-efficacy
theory as applied to practice in this area: behavioral
choice, effort and persistence; incentives for behavior; and
dimensions of behavior, such as level of difficulty of
behavior, generalizability of behavior and strength of
behavior.

Some questions were open-ended, to learn how physicians
understand these beliefs themselves. Data were gathered on
spontaneously-generated sources of efficacy information and
on aspects of self-efficacy in a professional practice area,
which were not a formal part of the theory, but which were
logically associated with it. These included motivation for
the development and maintenance of self-efficacious
behaviors, criteria by which one evaluates one’s own
behavior and goal-setting (and goal adjustment) which
supports self-efficacy beliefs. Special attention was paid
to interview material that did not support theory, that
disconfirmed theory or that theory neglected.

The analysis of data from the interviews was qualita-

tive and quantitative. Tabulations and groupings were

47

formed to present clear summaries of the data. Some
material is presented in the form of "patterns of practice."

Interview material is quoted to illustrate, with
examples, what interviewees said about themselves and their
practice with obese patients. Guidelines for the self—
efficacy "training" or "enhancement" of physicians are made
on the basis of the results of quantitative and qualitative
analyses.
Evaluation of the Self-Efficacy Interview

Validity. Content validity of the interview was
evaluated similarly to the self-efficacy measures, with
examination of the questions by self-efficacy experts.
Construct-related validity was tested by comparing high and
low self-efficacy doctors to each other, on variables of
interest and on test items and scores, as is described for
evaluating the self-efficacy measures.
Procedure

Physicians’ names were garnered from the phone book,
from the rolls of MSU’s Faculty Group Practice Directory,
from Health Central and PHP and from a physicians’ directory
for the Rapid City area. Guidelines for mail surveys were
followed from Dillman (1978) and Maheux, Legault & Lambert
(1989), with special attention paid to techniques which
increase response rate. These included personalizing the

cover letters with physician names and addresses, signing

48
them and including a handwritten note at the end of each
letter. There were no follow-up letters or phone calls.
Delimitations and Limitations of the Study

The sample chosen for the study was relatively broad.
It included doctors in two geographic areas, one rural and
western, one industrialized and mid-western. The physicians
practiced in cities, suburbs, small towns, VA hospitals,
Indian reservations and a military base. They were of a
wide range of ages and trained in a wide range of medical
schools. Their practice areas included all of the common
primary care practice designations: family practice, general
practice and general internal medicine.

The sample included doctors of both sexes and of
diverse practice affiliations. Practice affiliations ranged
from solo private practice to group private practice, health
maintenance organizations and university clinics. It
included doctors in government practice through the VA, the
USAF and the Indian Health Service. From the point of view
of the study sample, generalizability of findings of the
study is fairly broad.

Limitations of the study include a possible bias in the
sample, due to only efficacious doctors (or some other
subgroup of the sample) choosing to participate. To
encourage inefficacious doctors to participate, subjects had

the option of returning questionnaires anonymously.

49

There are also limitations inherent in using a self-
report measure. Both sources of information about study
doctors, (those about their beliefs and those about their
behaviors) are in the form of "self—reports." The estimates
that study subjects provide and the relationship between the
estimates may be, therefore, inflated. There is past self-
efficacy research evidence, however, which indicates that
individuals do report their self-efficacy beliefs accurately
(Bandura & Shunk, 1981).
Treatment of the Data

After the first phase of data collection,

descriptive statistics were used to examine the
characteristics of the sample on demographic variables,
practice behavior variables and on scores from the
self-efficacy measures. The measures were evaluated by
calculating reliability coefficients on each. A factor-
analysis was done on each measure to test for the structure
of the underlying constructs of the measures. Reliability
coefficients were also calculated for measure factor scores,
as well as total measure scores. A scatterplot graph and
Pearson Product Moment Correlation Coefficient was
calculated to examine the relationship between scores on the
physician self—efficacy instrument and the physician outcome
expectations instrument. Correlation coefficients were also

calculated on measure factor scores.

50

Regression analyses were done, using the self-efficacy
(PSEM) and the outcome expectations (POEM) scores as
predictors and physician practice behaviors as outcomes
(Physician Practice Measure items 10a. through 10d. and 11a.
and b.). This step in the analysis tested whether physician
beliefs and outcome expectations predicted physician
behavior. It was a test of the construct validity of the
study measures.

Three items on the outcome expectations measure (#1, #2
and #13) provided information for a possible confound in the
association between behavior and beliefs (and expectations.)
These addressed the issue of the degree of importance the
physician placed on the problem of obesity. They helped to
evaluate whether existing choice, effort and persistence may
be undermined, if a physician considers obesity treatment
unimportant.

A second analysis tested whether physician demographic
characteristics or practice characteristics (Practice
Measure items 1. through 7.) predicted self-efficacy (PSEM)
or outcome expectations POEM) scores.

The second phase of the data analysis was qualitative
in nature. The self-efficacy beliefs of a small sample of
individuals were examined via a semi-structured interview.
Individuals were chosen on the basis of self-efficacy
scores, so that efficacious and inefficacious doctors were

included among the interviewees.

51

Techniques of qualitative analysis were applied. Data
were sorted and organized; they were examined to identify
and generate categories and, then, themes and patterns based
on sources of efficacy information, physician attribution
about behavioral choice, effort and persistence, physician
goal-setting and physician performance self-evaluation. The
"constant comparative" method was used, which entails
coding, then grouping and regrouping data, working towards
an integration of data. (Strauss & Corbin, 1990.)

Emergent theories were evaluated against the data and
data were scrutinized to identify alternate explanations
(Marshall & Rossman, 1989). Attention was paid to spontan-
eously generated sources of self-efficacy information, to
strategies for appraisal and modification of beliefs and
behavior and to supports for self-efficacy development and
maintenance. "Patterns of practice" are presented, with
comparisons made of their relative strengths and weaknesses,
in supporting or undermining obesity treatment self-

efficacy.

Chapter IV

Results

Quantitative

Study Sample

Three hundred and sixty-four questionnaire packets were
sent out; 199 (55%) to doctors in Michigan and 165 (45%) to
doctors in the South Dakota area. In this original sample,
81 doctors (24%) had a Doctor of Osteopathic Medicine (D.O.)
degree and 278 (76%) had a Doctor of Allopathic Medicine
(M.D.) degree.

Altogether, 178 questionnaires were returned, a return
rate of 49%. The return rate from Michigan doctors was 47%,
and from South Dakota area doctors it was 48%. Five
questionnaires (3%) were returned anonymously. The return
rate for M.D.’s was 46%, and for D.O.’s it was 54%.

The make up of the final study sample is as follows:
94 study respondents (53%) are from Michigan; 79 (44%) are
from the South Dakota area; 133 (74%) are M.D.’s and 40
(23%) are D.O.’s. The medical degree from 5 doctors (3%) is
unknown.

The sample of doctors who returned their questionnaires
appears to be quite representative of the study target

sample with regard to geographic area of residence and

52

53
medical degree. The extent of their representativeness on
other study characteristics is unknown.

Description of the Sample. Besides type of medical
degree and area of residence, demographic information
collected on study doctors included age, sex, year of
medical school graduation, medical specialty, type of
practice arrangement, size of practice, approximate number
of patients seen each week, percent of patients seen who are
female and whether there are obese patients in the practice.

Age; The study doctors ranged in age from 30 to 76;
the median age was 43 years. 36% of the doctors were aged
40 or younger, 60% were 45 or younger and 80% were 55 or
younger.

Sex. Year of Medical School Graduation. Seventy-eight
percent of the study sample were males and 22% were females.
The year of medical school graduation ranged from 1943 to
1992 and the median year was 1974. Approximately 70%
graduated between 1966 and 1986.

Medical Specialty. Eighty-eight percent of the study
sample specialized in family practice (61%) or general
internal medicine (28%). The remaining 12% were general
practitioners, all of whom resided in the South Dakota area.

Practice Type. Sixty percent of the doctors were in
private practice and 25% worked primarily in a university-

affiliated clinic. Of the remaining 15%, 6% were employed

54
by the Veterans’ Administration, 5% by the Indian Health
Service, 3% by the US Air Force and 1% by HMO’s.

Practice Size. The size of the practices in which the
doctors work were as follows: 17% were in solo practice,
28% were in a practice of 2 to 3 doctors, 24% were in
practice with 4—6 doctors and 32% worked in a practice of 6
or more doctors.

Patient Load. The size of the doctors’ weekly patient
load was as follows: 11% saw fewer than 20 patients per
week, 30% saw approximately 50, 30% saw about 100 per week
and 23% saw more then 100 patients per week.

Percent Female. The percent of patients in a practice
that was female ranged from 1% to 100%. The average percent
of female patients in a practice was 60%.

Ninety-nine percent of study doctors stated that they
had obese patients in their practices.

The typical study subject was a male doctor,
specializing in family practice; he was in his early
forties, worked in a group practice of 2 to 6 doctors and
saw between 50 and 100 patients per week, 60% of whom were
women and some of whom were obese. Table 1. summarizes

information about the study sample.

55
Table 1. Description of the Study Sample

Tarqet Sample Study Sample

Qualitative

 

 

 

 

n = 364 n = 178 Sample n = 10
State
MI 199 (55%) 94 (53%) 10(100%)
SD 164 (45%) 79 (49%)
Medical
Degree
MD 278 (76%) 133 (74%) 8 (80%)
DO 81 (24%) 4o (23%) 2 (20%)
Aqeplmedian) 43.0 yrs. 43.5 yrs.
§§§
M 137 (78%) 8 (80%)
F 40 (22%) 2 (20%)
Med.sch.
grad.year
(median) 1974 1976
Medical
Specialty
PP 108 (61%) 5 (50%)
IM 49 (28%) 5 (50%)
GP 21 (12%)
Practice
Type
private 106 (60%) 2 (20%)
university 45 (25%) 8 (80%)
VA. 10 (6%)
IHS 9 (5%)
USAF 6 (3%)
HMO 2 (1%)
Practice Size
solo 30 (17%)
2-3 drs. 49 (28%) 1 (10%)
4-6 drs. 43 (24%) 1 (10%)
>6 drs. 56 (32%) 8 (80%)
Patient Load
per Week
<20 23 (11%) 1 (10%)
20-75 55 (30%) 6 (60%)
75-125 55 (30%) 3 (30%)
>125 45 (23%)
% Practice
who are Female (mean) 60% 55%

 

56
Evaluation of the Study Measures

Two separate measures were formulated for the study.
This reflects the conceptual distinction between self-
efficacy beliefs and outcome expectations and tests this
distinction, empirically. Analyses are done, separately,
with each measure.

The Physician Self-Efficacy Measure (PSEM). The PSEM
is an ll-item measure. For this initial sample, the range
of scores is between 26 and 55; the mean is 42.11 and the
standard deviation is 5.80. The distribution of scores is
roughly normal.

To evaluate the reliability of this measure, a
coefficient of internal consistency was calculated. The
Cronbach’s alpha for the PSEM is .84. The standardized item
alpha is .85, indicating that the item variances are
uniform. The correlations between individual items and the
scale range from .36 to .66 and the reliabilities of the
scales, if each item is deleted from the scale, in turn,
range from .82 to .85. This measure has a high degree of
internal consistency and each item contributes equivalently
to its reliability.

The Physician Outcome Expectations Measure (POEM). The
POEM has 14 items. The range is 37 to 65, on this
administration; the mean is 47.83 and the standard deviation
is 5.25. On a sample of 178 scores, the distribution

appears relatively normal, but is positively skewed.

57

The coefficient of internal consistency for this
measure is a Cronbach’s alpha of .62. The standardized item
alpha is .62, indicating uniform item variances. The
correlations between individual items and the scale range
from -.03 to .54 and the reliabilities of resulting scales,
if each item is deleted, range from .56 to .64. This
measure has a lower degree of internal consistency than the
PSEM .

Factor—Analysis (PSEM). The principal-components
analysis for the PSEM indicates that this is a measure with
three factors. Three uncorrelated linear combinations of
item values have eigenvalues that are greater then 1.0.
These three factors account for 67% of the total variance in
PSEM scores. On varimax rotation (which minimizes the
number of items that have high loadings on a factor), the
three factors that result are consistent with the
substantive areas which were identified as salient. The
three areas are obesity knowledge unique to medical
practice, interpersonal skills and dealing with one’s own
negative feelings about obesity. The first factor, which
loads on items 4,7,8 and 10, embodies the construct of
physician self-efficacy in interpersonal relationship skills
with patients. These items ask if doctors believe they can
"help" patients, "engage" patients, "adjust" their advice to
patients. This factor accounts for approximately 40% of the

variance in scores.

58

The second factor (which loads in items 1,2,5,6 and 9)
encompasses physician self-efficacy in knowledge of and
teaching about, obesity and treatment, which is unique to
medical practice. These items ask doctors about their
confidence in diagnosing obesity with health risks and in
selecting, explaining and rationalizing a treatment plan to
and with patients. These two factors, together, account for
55% of the variance in PSEM scores.

The third factor (loading on items 3 and 11) accounts
for 12% of the variance in scores and represents the area of
self-efficacy in managing negative affect. Third factor
items ask how capable physicians believe themselves to be in
managing their own negative feelings about obese persons and
in maintaining good relationships with obese patients who do
not lose weight. Table 2. summarizes information about the

factor structure of the PSEM.

59

Table gnyrincipal-Components Analysis and Rotated Factor
Matrix - PSEM

Factor Eigenvalue Pct. of Var. Cronbach’s a (adj.a)

 

 

1 4.43513 40.3 .84 (.84)
2 1.60108 14.6 .79 (.80)
3 1.30626 11.9 .70 (.71)
Factor 1 Factor 2 Factor 3
PSEM 4 .86601 .02984 .14565
PSEM 8 .85073 .16313 .00694
PSEM 10 .83050 .16057 .11185
PSEM 7 .56188 .37766 .35070
PSEM 5 .11992 .82422 .09928
PSEM 9 .15513 .78904 .12873
PSEM 1 .00641 .77426 .07817
PSEM 6 .54742 .56757 .11559
PSEM 2 .41496 .55082 .05502
PSEM 3 .13917 .06849 .86523
PSEM 11 .08339 .16705 .84103

Factor—Analysis (POEM). The principal-components
analysis for the outcome expectations measure produces five
factors with eigenvalues greater than 1.0. On varimax
rotation, the five resulting factors reflect areas which
were judged to be salient aspects of obesity treatment
outcome expectations: whether or not obesity treatment is
important, whether or not obesity is treatable and, if so,
with whom does responsibility for treatment lie--the
physician, the patient or both. These five together account
for 63% of the variation in POEM scores.

The first factor, which loads on items 5,6,7 and 8 and
accounts for 20% of the variation in scores, encompasses the

idea that obesity is treatable and that treatment is the

60
responsibility of both patients and doctors ("patients who
try, can lose weight," "doctors have influence on their
patients," "doctors can be effective counselors.")

The second factor, which accounts for 15% of score
variation and which loads on items 1, 2 and 3 (obese
patients are at risk, losing weight makes a positive
difference, patients who restrict calories lose weight)
illustrates the idea that obesity is detrimental to health
and happiness, that treatment for it is important and that
treatment is effective. Inherently, this factor subsumes
that obesity is treatable.

The third factor (loading on items 4,9 and 10 "patients
do what they want", " patients attempt, but do not lose
weight") is the idea that obese patients are non-compliant
and obesity is untreatable. This factor accounts for 11%
of POEM variance.

The fourth factor accounts for 10% of score variation
and encompasses the idea that obesity is treatable, but that
weight loss is often temporary, not permanent. This factor
loads on items 11 and 12.

The fifth factor (loading on items 13 and 14) and
accounting for 8% of total score variance is the idea that
obesity is not treatable ("whatever patients or doctors do,
weight loss is only temporary.") Table 3. summarizes

information about the factor structure of the POEM.

61

 

 

 

 

 

 

 

Table 3. Principal-Components Analysis and Rotated Factor
Matrix — POEM

Factor Eigenvalue Pct. of Var. Cronbach's a (adj.a)

1 2.82247 20.2 .70 (.71)

2 2.07753 14.8 .66 (.68)

3 1.49493 10.7 .51 (.51)

4 1.32669 9.5 .43 (.43)

5 1.10836 7.9 .46 (.47)

Factor 1 Factor 2 Factor 3 Factor 4 Factgr 5
POEM 7 .79476 .11266 .01176 .04813 .09821
POEM 6 .79109 .13539 .14929 .05970 .09577
POEM 5 .65423 .17780 -.17080 .18513 .22271
POEM 8 .56831 .01169 .10672 .27056 -.22959
POEM 2 .12708 .85444 -.02816 .06424 .04658
POEM 1 .00294 .77519 —.28953 .12448 -.09966
POEM 3 .22066 .65138 .11892 .17402 .00966
POEM 9 .15430 .02288 .70620 .00723 .30283
POEM 10 .20896 .07071 .70247 .01286 .04575
POEM 4 .00374 .31528 .63728 .08723 -.29909
POEM 11 .04105 .04792 .04905 .85514 .11589
POEM 12 .42893 .11543 -.07439 .58912 -.16516
POEM 14 .14622 .10938 .00225 .09099 .81605
POEM 13 .02166 .16482 .26096 .45020 .65701
Relationship Between the PSEM and the POEM. When

 

plotted against each other and correlated, the total PSEM

and POEM scores show a moderately strong, positive

relationship (; = .46) that is linear. Doctors whose self-

efficacy beliefs about treating obesity are strong are also
more likely to have high outcome expectations for obesity
This correlation coefficient suggests that

treatment .

though they are not independent of one another, self-

62
efficacy beliefs and outcome expectations are empirically as
well as conceptually distinct.
The first two factors of each measure also correlate
with each other and with each total score. A correlation
matrix of the total scores and the first two factor scores

for each measure follows (Table 4.).

Table 4. Correlation Matrix of PSEM and POEM Total Scores
and First Two PSEM and POEM Factor Scores

Correlations PSEM POEM PSEFACl PSEFACZ POEMFACl POEMFACZ

PSEM 1.0 .459** .695** .588** .432** .316**
POEM 1.0 .397** .254** .647** .413**
PSEFACl 1.0 -.006 .468** .094
PSEFAC2 1.0 .186 .365**
POEMFACl 1.0 .094
POEMFAC2 1.0

* signif at .05 ** signif at .01

The interrelationships between PSEM and POEM factor scores,
with each other as well as with both total scores, is
further evidence of the relatedness of the constructs of
self-efficacy and outcome expectations.

When placed on the same metric, to compare the scores
to each other, the mean of the PSEM is 3.83 (egg; = .38) and

the mean of the POEM is 3.40 (s.d. = .53). PSEM scores are

63
significantly higher (p = -11.45, p < .01) than POEM scores.
In general, doctors’ beliefs about their own capability in
treating obese patients are greater than their beliefs about
the outcomes of obesity treatment.
Evaluation of the Physician Behavior Variables

Items 10 a.,b.,c. and d., items 11.a. and b. of the
Physicians Practice Measure asked physicians about their
behavior with patients in the past month and the past year.
These first six items related directly to study hypotheses.
The final two questionnaire items were more general. Item
12. attempted to determine whether knowing when to refer a
patient to an obesity expert is part of primary care obesity
treatment self-efficacy.

The final question (whether a physician would be
willing to be interviewed about obesity treatment) sought to
test whether stronger self-efficacy doctors would be more
willing to be interviewed than would weaker self-efficacy
doctors. The answer to this item also determined the pool
of study subjects who were approached for an interview.

The spread of scores for the reported percent of
patients seen in the past month who were obese (item 10a.)
ranged from 0% to 100%. The mean for this variable is 28%
(gygy = 18%.) The distribution of scores is steep and is
positively skewed. More doctors see a low percent of obese
patients than a high percent. This variable is not normally

distributed.

64

The distribution of scores for the reported percent of
obese patients with whom the physician discussed weight loss
(10b.) ranges from 0% to 100%. The mean is 52% (s.d.= 33%).
The distribution of scores is not normal; it is steep and
irregular and the numbers do not coalesce in to a smooth
shape, but instead in to a jagged, irregular "mountain
range" shape. Doctors discuss obesity with many different
percents of their patients.

The distribution of scores for item 10.c. also has a
steep, jagged, irregular shape. Its values range from 0% to
100% with a mean of 50% (sygy = 38%). Doctors initiate a
discussion of obesity with many different percents their
patients.

The distribution of scores for the reported percent of
patients on a weight loss program (10d.) ranged from 0% to
100%, the mean was 30% (sygy = 28%). The distribution of
scores is steep and is positively skewed. Lower, rather than
higher, percents of study doctors’ patients are on weight
loss programs.

The scores on item 11.a. (number of patients for whom
doctors sought consultation) are very abnormally
distributed. Scores range from 0 to 4,000. The
distribution of scores is very steep and very positively
skewed. Some subjects responded to this item
idiosyncratically (e.g. "all of my patients" ) and many

responded by naming "dietician" or "nutritionist" as the

65
‘person.to whom patients were most often "referred out". The
majority of respondents sought consultation on none of their
obese patients. Due to the irregularity of item responses,
this item was not used in the study analyses.

The scores on item 11.b. (the number of patients
referred out of a practice for primary care) are also
extremely abnormally distributed. It is extremely
positively skewed and steep. It is rare for a patient to be
referred elsewhere for primary care, among study doctors.
This behavior has a low base rate and, therefore, was not
used in study analyses.

The responses for the final two Physician Practice
Measure questions are dichotomous. For question #12., 77
doctors (43%) have a patient currently in their practice who
they would like to refer to an obesity specialist and 96
(54%) do not. Four doctors did not answer this question.
For the final question, "Are you willing to be interviewed?"
76 doctors (43%) agreed to an interview and 101 (57%) did
not.

Association of Physician Characteristics with Self-Efficacy
Scores

When self-efficacy (PSEM) and outcome expectations
(POEM) scores are examined for differences on the basis of
medical specialty, sex, practice type, practice size,
patient load per week, state of residence and medical

degree, few differences emerge. After critical values are

66
adjusted for experimentwise error rate, using Tukey multiple
comparison procedures, there are three differences.

Type of medical degree is associated with a difference
in POEM scores. D.O.'s have significantly higher outcome
expectations than do M.D.’s for their obesity treatment
behaviors (Table 5.). One type of medical practice is
associated with a difference in both PSEM and POEM scores.
Doctors in solo practice score significantly higher on both

measures (Table 5.).

Table 5. Medical Deqree. Practice Size and PSEM and POEM
Scores

 

 

 

 

Medical Degree Means and SD E
PSEM POEM PSEM POEM

entire pop. 42.11 (5.8) 47.83 (5.25)

M.D. 42.08 (5.3) 47.31 (4.9)

D.O. 42.48 (6.9) 49.88 (6.1) .15 7.60**
Practice Size PSEM POEM

solo 44.67 (5.5) 50.13 (5.4) 2.89* 3.38*

2-3 doctors 42.2 (6.6) 47.98 (4.5)

4-6 doctors 41.19 (5.3) 47.55 (6.3)

6+ doctors 41.18 (5.2) 46.47 (4.4)

* signif at .05 ** signif at .01

There is also one patient load size which is associated
with a difference in scores. Doctors who see fewer than 20
patients per week score lower on the self-efficacy measure
than doctors seeing more than 175 patients each week. When
patient load size is regressed on self-efficacy scores, this
number predicts both the PSEM and the POEM, in the expected

direction. The more patients seen per week, the higher are

67
higher are the self-efficacy beliefs and the outcome

expectations (Table 6.).

Table 6. Patient Load and PSEM and POEM Scores

 

Patients Means and SD E R-sgpare
per week PSEM POEM PSEM POEM PSEM POEM
<20 38.68 (5.0) 45.83 (3.7) 3.91* 6.49** .04 .04

20-75 42.25 (5.6) 47.21 (4.9)
75-125 42.58 (5.3) 47.88 (4.4)
125-175 42.73 (6.1) 48.63 (6.5)
175+ 42.80 (7.6) 50.0 (6.0)

* signif at .05 ** signif at .01

With a few exceptions, demographic and practice
characteristics are not associated with differences in self-
efficacy beliefs or in outcome expectations. No hypotheses
were made about these relationships.

Association of Physician Behaviors and Self-Efficacy Scpres

Physician-reported practice behaviors are of particular
significance for this study because they provide evidence
for the validity of study measures. To explore physician
self-efficacy, as exemplified by behavioral choice, effort
and persistence, Physician Practice Measure items 10
a.,b.,c. and d. and 11 a. and b. asked for quantitative
estimates of recent practice behaviors. These behaviors

were considered, for this study, to be independent evidence

68
of self—efficacy beliefs in obesity treatment. The
behavioral estimates were regressed on the self-efficacy
scores to test if self-efficacy scores predict self-
efficacious behaviors.
First a correlation matrix was calculated for self-
efficacy measure scores and physician behavior variable

scores (Table 7.).

Table 7. Correlation Matrix of Physician Behavior Scores and

PSEM and POEM Scores

Correlations PSEM POEM PPMlOa. PPMlOb. PPM10C PPMlOd.

PSEM ‘ 1.000 .459** .194 .320** .292** .343**
POEM 1.000 .096 .252* .252* .237*
PPMlOa. 1.000 .151 .190 .032
PPMlOb. 1.000 .807** .512**
PPMlOc. 1.000 .423**
PPMlOd. 1.000

* signif at .05 ** signif at .01

When self-efficacy scores (PSEM) are used in a
univariate regression analysis, they predict the percent of
patients seen in a doctor’s practice who are obese, the
percent of obese patients with whom the physician discusses
weight loss, the percent of patients with whom the physician
initiates such a discussion and the percent of patients who

are on a weight control program. The higher a doctor scores

69
on the self-efficacy measure, the higher a proportion of
efficacious behaviors of choice, effort and persistence in

obesity treatment he or she reports (Table 8.).

Table 8. Physician Behaviors and PSEM Scores
Physician Behaviors B SE B Beta E R-s are

 

Proportion of
patients who are

obese .007283 .002337 .237101 9.71* .06

Proportion of
patients dis-

cussed weight
loss with .020286 .004177 .355863 23.58** .13

Proportion of
patients ini-

tiated dis-
cussion with .020290 .004796 .325274 19.16** .11

Proportion of
patients on a

weight loss
program .017997 .003577 .373548 25.30** .14

* signif at .05 ** signif at .01

A similar pattern emerges when physician behaviors are
regressed on PSEM factor scores. The first factor predicts
the latter three behaviors; it does not predict the first.
The second factor scores (obesity knowledge) predicts
behaviors in an identical pattern to the total PSEM score;
it predicts behaviors 10a. through 10d. The third factor
scores (affect management) predicts no behaviors. The
reliability of the third factor (a = .70) is less than those
for the first and second factors (a = .84 and .79,

respectively), but not very much less, so it is unlikely

70

that lack of reliability accounts,

predictive power of the third factor.

summarized in Table 9.

Table 9.

Physician Behaviors

Proportion of
patients who
are obese
PSEFAC 1
PSEFAC 2
PSEFAC 3

Proportion of
patients dis-

cussed weight
loss with
PSEFAC 1
PSEFAC 2
PSEFAC 3

Proportion of
patients ini-
tiated dis-
cussion with
PSEFAC 1
PSEFAC 2
PSEFAC 3

Proportion of
patients on a

weight loss
program
PSEFAC 1
PSEFAC 2
PSEFAC 3

E

.016359
.044140
.009869

.094711
.073116
.020577

.092166
.082227
.018359

.093654
.055723
.009420

* signif at

.013849
.013236
.014363

.024937
.025143
.026889

.028580
.028536
.030666

.021096
.021524
.022980

.05

totally,

for the lack of

This information is

Beta

.092128
.252731
.053739

.283540
.220797
.059251

.245611
.220806
.046694

.334911
.202961
.032802

E

** signif at

Physician Behaviors and PSEM Factor Scores
§E_B

R- SQ pare
.40 .01
.12** .07
.47 .01
.42** .08
.46** .05
.58 .00
.40** .06
.30** .05
.36 .00
.71** .11
.70* .041
.16 .001
.01

The outcome expectations scores predict physician

behaviors almost as well as do the self-efficacy scores.

Outcome scores predict the prOportion of patients with whom

a physician discusses weight loss,

the proportion with whom

he or she initiates such a discussion and the proportion of

71
his or her patients who are on a weight control program. It
does not predict proportion of a physician’s patients who

are obese (Table 10.).

Table 10. Physician Behaviors and POEM Scores
Physician Behaviors B SE B Beta F R-s are

Proportion of

patients

who are obese .004132 .002663 .120986 2.4 .01
Proportion of

patients dis-

cussed weight

loss with .017360 .004785 .272588 13.16** .07
Proportion of

patients ini-

tiated dis-
cussion with .019830 .005436 .276300 13.31** .08

Proportion of
patients on a

weight loss
program .044137 .004106 .266533 11.85** .07

* signif at .05 ** signif at .01

When physician behaviors are regressed on POEM factor
scores, only the first factor (obesity is treatable and
treatment is a shared responsibility) has good predictive
power. It predicts all of the physician behaviors. Factor
two (obesity is treatable and treatment is important)
predicts two behaviors and factor three (patients are non-
compliant and do not try) predicts one behavior. Only the
first two POEM factors have respectable reliability
coefficients (a = .70 and .66, respectively). A summary of

POEM factor scores is contained in Table 11.

Table 11. Physician Behaviors and POEM Factor Scores

 

72

Physician Behaviors E SE B Beta 3 R-sgpare
Proportion of

patients who

are obese

POEFAC 1 .026132 .013686 .153047 3.65* .02
POEFAC 2 .009352 .013913 .054442 .45 .00
POEFAC 3 .004618 .014556 .025726 .10 .00
POEFAC 4 .033493 .014789 .180666 .52 .00
POEFAC 5 .009468 .014171 .054111 .45 .00
Proportion of

patients dis-

cussed weight

loss with

POEFAC 1 .080292 .025594 .246600 9.84** .06
POEFAC 2 .061865 .026092 .188852 5.62* .04
POEFAC 3 .026023 .027686 .076018 .88 .01
POEFAC 4 .088247 .028636 .051615 .41 .00
POEFAC 5 .021078 .027009 .063173 .61 .00
Proportion of

patients ini-

tiated dis-

cussion with

POEFAC 1 .072698 .029547 .195707 6.05* .05
POEFAC 2 .059190 .029931 .158378 3.91* .04
POEFAC 3 .028782 .031624 .058334 .52 .00
POEFAC 4 .016523 .032686 .040967 .26 .00
POEFAC 5 .000573 .030875 .001508 .00 .00
Proportion of

patients on a

weight loss

program

POEFAC 1 .063613 .021371 .234692 8.86** .06
POEFAC 2 .034391 .021942 .126111 2.45 .02
POEFAC 3 .050765 .022745 .178140 4.98* .03
POEFAC 4 -.025364 .023782 -.086185 1.14 .01
POEFAC 5 -.021477 .022462 -.077325 .91 .01

* signif at .05 ** signif at .01
The self-efficacy and outcome expectations measures
were also used together in a multiple regression analysis.

When behaviors are regressed on both scores in a stepwise

73

fashion using a criterion of p < .05, PSEM scores, alone,

predict physician behavior

(Table 12a.).

Table 12a. Physician Behaviors and PSEM and POEM Scores with

a Criterion of

Physician Behavior

Proportion of
patients who

are obese
Proportion of
patients dis-
cussed weight
loss with
Proportion of
patients ini-
tiated dis-
cussion with
Proportion of
patients on a
weight loss
program

.05.

.006886

.019707

.020468

.017759

* signif at .05

 

 

SE B Beta E R-sguare
.002356 .223829 2.9* .05

.004215 .342919 4.7** .12

.004846 .315853 4.2** .10

.003623 .366383 24.02** .13

** signif at .01

POEM scores do not meet the criterion of .05 and are

dropped out of the equation

(Table 12b.).

74

Table 12b. Physician Behaviors and PSEM and POEM Scores with
a Criterion of .05

 

Physician Behaviors Beta In Partial Min toler 1

Proportion of
patients who

are obese .021034 .019003 .781860 .24
Proportion of

patients dis-

cussed weight

loss with .139769 .130365 .767658 1.7
Proportion of

patient ini-

tiated dis-

cussion with .161290 .148801 .766227 1.9
Proportion of

patients on a

weight loss

program .124419 .118756 .788746 1.5

* signif at .05 ** signif at .01

However, when physician behaviors are regressed on PSEM
and POEM scores with no criterion, both show some predictive

power (Table 13.).

Table 13.

75

Physician Behaviors and PSEM and POEM Scores

 

 

 

 

 

 

 

 

without Criterion of .05
Physician Behaviors B SE B Beta B R-s are
Proportion of
patients who PSEM .006583 .002672 .214005 4.28 .05
are obese POEM .000718 .002966 .021034
Proportion of
patients dis-
cussed weight PSEM .015835 .004785 .275547 12.46** .13
loss with POEM .008901 .005303 .139769
Proportion of
patients ini-
tiated dis- PSEM .015414 .005491 .237870 10.88** .12
cussion with POEM .011576 .006082 .161290
PrOportion of
patients on a
weight loss PSEM .14987 .004063 .309198 13.31** .15
program POEM .006599 .004464 .124419
* signif at .05 ** signif at .01
Self—efficacy (PSEM) scores predict all physician-

reported behavioral estimates.

This tests the validity of

the measure. The higher a doctor scores on the self-

efficacy measure, the more often he or she reports
efficacious behaviors of choice, effort and persistence in
obesity treatment behaviors.

Outcome expectations scores (POEM) predict physician
behaviors also, but they are a less powerful predictor. The
reliabilities of the two predictor variables are quite
different: the PSEM has a reliability coefficient of a =.84
and the POEM as a coefficient of a = .62. The lesser
predictive power of the POEM may be, at least, partly

accounted for by this lesser reliability.

76
Relationship Between Other Study Variables and Scores

The self-efficacy measure score was tested for its
association with whether doctors had a patient currently in
practice whom they would like to refer to an obesity
specialist. There is no significant difference in scores
for those who have such a patient and those who do not.
Knowing when to refer to an obesity expert is not associated
with a higher self-efficacy score.

When scores were tested for their association with
willingness to be interviewed, there was found to be a
positive association (p = 2.23, p < .05.) Those scoring
higher, more often agreed to be interviewed.

The outcome expectations measure was similarly tested.
There was not found to be a difference in scores for those
who had an obese patient they would like to refer to an
expert and those who did not. Higher outcome expectations
are not associated with wishing to refer a patient to a
specialist. There was also no difference found on this
measure between those willing to be interviewed and those
who were not. Those with higher outcome expectations were

not more willing to be interviewed.

Interpretation (QuantitativeL
Study Sample. The original group of 365 doctors was
every primary care physician who could be readily identified

in the two areas. It was assumed to be broad-based on the

77
study demographic and practice variables because it sought
to include all primary care physicians, practicing under a
variety of auspices, in two different geographic areas.

The return rate for questionnaires, about 50%, was
good, when compared to other studies of primary care
physicians. The sample of doctors who returned their
questionnaires was representative of the study target
sample, with regard to medical degree and geographic area of
residence. On these variables, the doctors who participated
and those who did not were similar.

The similarity, between those who returned their
questionnaires and those who did not, on other study
variables is not known. However, there is wide variation,
on all variables--demographic, practice and behavioral--
among the doctors who participated in the study.

There is also wide variation in the scores on the self-
efficacy measures of those who participated. Self-efficacy
in treating obesity appears to be roughly normally
distributed among study subjects. The scores on the outcome
expectations measure are not as normally distributed; they
are positively skewed. More doctors have low outcome
expectations for obesity treatment than high outcome
expectations.

Generalizability. The high degree of variation on
study variables and measure scores is supporting evidence

that the actual study sample is similarly as broad-based, as

78
was the study sample sought. The sample appears to be
representative of the study population. The
generalizability of study findings, therefore, is fairly
strong.

The return rate for D.O.’s was better than it was for
M.D.’s (54% as opposed to 46%). D.O.’s are slightly over—
represented in the study sample, as compared to the general
population of doctors. However, overall, many more M.D.’s
than D.O.’s participated in the study. This reflects the
reality of more M.D.’s than D.O.’s being in medical practice
in America.

Goals of the Study. Two measures were developed and
tested. Self-efficacy beliefs were the primary focus of the
study, but outcome expectations were measured also. Outcome
expectations are of interest in themselves because they are
a closely-related construct. They are also a potential
confound, for doctors reporting their self-efficacy beliefs.

The scale developed to measure self-efficacy is a
highly reliable one. The PSEM’s items add weight to it’s
reliability uniformly. The distribution of scores is
roughly normal, providing supporting evidence that self-
efficacy, like many other human abilities and
characteristics, is normally distributed throughout the
population and that the study sample is an adequate one.

The three areas of importance in measuring self-

efficacy in the treatment of obesity--obesity knowledge

79

unique to medical practice, interpersonal skills and
management of negative affect--are reflected in the analysis
of factors for the measure. These three factors, together,
account for 67% of the variation in the scores. A doctor’s
interpersonal skills with patients, that accounts for 40% of
the variation in scores, is clearly the most important
factor in physician self-efficacy, as measured here.
Obesity knowledge (15%) and management of negative affect
(12%) are the other factors that contribute to self-efficacy
in treating obesity.

The scale developed to measure outcome expectations is
a less reliable scale and the distribution of scores is less
normal; it is positively skewed. Doctors, therefore,
whether they have strong or weak self—efficacy beliefs, tend
to have low outcome expectations for obesity treatment.

Interpreting the pattern of factors that emerges from
the analysis for the outcome expectations measure is
difficult. The important factors for treatment outcome--
whether a physician thinks that obesity can be treated,
whether she or he thinks it is important to treat obesity
and with whom the physician thinks responsibility for
treatment lies--with the doctor or with the patient--are
present. They do not as clearly distinguish between
separate items, as expected. The issue of the
"treatability" of obesity (or its untreatability) is

explicitly or implicitly a part of each item.

80

When the scores (self-efficacy and outcome
expectations) are compared to each other within individuals,
outcome expectations scores are lower than self-efficacy
scores. Doctors are more self-efficacious about their own
treatment behavior than about treatment outcome.

Maintenance of normal weight, as a health "endpoint,"
especially over a life span, is notoriously multidetermined.
It seems reasonable that physicians are more sure of
something over which they have personal control (their own
knowledge and skills,) rather the than other factors which
influence that endpoint.

Research Questions

There are few differences in self-efficacy scores on
the basis of age, sex, medical specialty, year of medical
school graduation, practice type or size, number of patients
normally seen per week, percent of patients seen who are
female, area of residence or medical degree. There are a
few exceptions.

Doctors in solo practice score significantly higher on
both measures. This is an uncommon practice type,
particularly in Michigan and this group is most likely
highly self-selected. One assumption that might be made
about this group is that solo practitioners are different
from doctors who practice in a group, on a number of

personality variables. The high degree of autonomy and

81
self—sufficiency that solo practice necessitates may "over-
determine" self-efficacy in any area.

Self—efficacy scores increase with number of patients
seen per week. This increase is gradual, but significant
and the biggest difference is between doctors seeing fewer
than 20 patients each week and doctors seeing more than 175
patients each week. "Performance enactment", or doing the
required behavior, is the most powerful source of efficacy
information, according to Bandura’s theory. Consistent with
theory, doctors who see more patients and, therefore, are
most likely do more of the requisite obesity treatment
behaviors are more self-efficacious.

One other demographic difference in scores emerges--
those with D.O. degrees score significantly higher on the
outcome measure than M.D.’s do. Osteopathic physicians have
higher outcome expectations than do M.D.’s. It may be that
the philosophical underpinnings of osteopathy contribute to
that optimism.

Study Hypotheses

The self efficacy measure predicts 4 out of 6 self-
efficacious behaviors (Hypotheses 1 through 4.) Higher
self-efficacy scores predict a higher percent of patients
seen in a practice who are obese, a higher percent with whom
obesity is discussed, a higher percent with whom the doctor
initiates such a discussion and that a higher percent of

patients in the practice will be on a weight loss program.

82
These behavioral reports represent choice, effort and
persistence, i.e. efficacious obesity treatment behavior.

Hypotheses 5 and 6 dealt with scores that predicted
behaviors that happen only rarely. Therefore, these
hypotheses were not formally tested. Interestingly, whether
doctors feel themselves to be efficacious or not, they keep
obese patients in their practices and do not refer them to
another doctor for primary care.

When each of the three PSEM factors is used as a
separate scale, the predictive power of each of the first
two factors (interpersonal skills and obesity knowledge)
compares to that of the total scale. Individually,
knowledge and interpersonal skills, each, predict physician
behaviors.

The affective component of the scale, when examined on
its own, has poor predictive power. From a counseling point
of view, this is surprising. It was expected that the way a
doctor feels about obese patients, particularly obese
patients who do not lose weight, would be important to her
or his self—efficacy in obesity treatment. This does not
appear to be the case. Interpersonal skills and knowledge
about obesity are far more important in predicting
efficacious behaviors.

The affective scale is a two-item scale. Though its
reliability is less than that of the other two factor

scales, (by approximately .10,) it still respectable (a =

83
.70,) so it seems unlikely that lack of reliability, alone,
accounts for its lack of predictive power. The lack of a
relationship between negative feelings about obesity and
self-efficacious obesity treatment behaviors appears to be a
bona fide finding and not just a statistical artifact.

Poem scores also predict physician behaviors, in the
expected direction. Higher outcome expectations scores
predict that a doctor will discuss obesity with higher
percent of patients, will initiate this discussion with a
higher percent and will have a higher percent of patients on
a weight loss program than do doctors with lower scores.
Outcome expectations scores do not predict that a higher
proportion of patients seen in practice will be obese.

Neither self-efficacy nor outcome scores were
associated with knowing when to refer a patient to an
obesity expert. Self-efficacy is associated with a
willingness to be interviewed. The interviewed group,

however, is a diverse one.

Qualitative
Introduction
The qualitative part of the study was designed to

address the question of what attributions physicians make,

84
themselves, for the development and maintenance of their
self-efficacy in obesity treatment.

The results of the qualitative data analysis are
presented according to principles of qualitative analysis:
they are sorted and examined to generate categories;
patterns are identified, that are based on self-efficacy
principles of efficacy information sources and self-efficacy
dimensions. Practice themes emerged from the interplay of
these self-efficacy principles and the manner in which the
problem of obesity and its treatment was conceptualized by
primary care physicians. Physician attributions, within
these conceptualizations, were identified which support or
which weaken the link between self-efficacy beliefs and
behavior.

Qualitative Sample. A total of 10 primary care
physicians were interviewed for the qualitative part of the
study. Of the 10, two were females, two were D.O.'s and two
were African-American. Five were family practitioners and
five were internists; two were community practitioners and
eight practiced primarily in university clinics. Their ages
ranged from 38 to 57, their weights from normal to
moderately obese. They all resided in Michigan.

By the criteria of the study measures, six scored above
average in self-efficacy and four below average. Five
scored above average in outcome expectations and five scored

below. Seven subjects had "consistent" scores, four had

 

~xu~

85
above average scores on both measures, three had below
average scores on both. Of the three who were
"inconsistent" scorers, two were high on self—efficacy and
low on outcome expectations and, for one subject this was
reversed.

Although the initial recruitment of interview subjects
was done on the basis of study scores, many dOCtors refused
the request for an interview. Those interviewed were
selected solely on the basis of their agreeing to an
interview, not on the basis of measure scores. It was,
therefore, partly fortuitous that the interviewees were a
varied group. A doctor’s score was not noted immediately
prior to the interview to provide some control for
interviewer bias.

The qualitative sample was quite varied on all study
variables including the study measures. There was no strong
evidence that this group was not fairly representative of
the study sample as a whole or, indeed, of the study
population. University-affiliation, however, is over-
represented in the study sample (25%) and even more in the
qualitative sub-sample (80%).

Sources of Efficacy Information

Behavioral Enactment. All 10 doctors mentioned that
experience in practice treating obese patients has been
their primary means of their learning how to treat obesity,

how to assess their competence and how to improve their

86
skills. All subjects used phrases like "I’ve gotten more
comfortable in addressing the problem over time," "I’ve
become more effective with my patients over time" and "my
patients have been my teacher and their response is my
guide." Seeing patients, particularly the same patients
over time, has provided a feedback loop, so that treatment
approach, it’s outcome (positive or negative) and possible
modification developed along with and as a part of the
doctor-patient relationship. All the doctors brought up the
"tempering" of their outcome expectations over time: "I’m
more humble now in having limited goals with patients;" "I
see that ’holding the line’ is a success for some obese
patients."

Role Modeling, Vicarious learning through role-models,
in a positive sense, was mentioned by only two doctors.

This occurred because of specific qualities of a training
site, for one doctor, and of "mentoring" individuals, for
the other doctor.

More common was role-modeling in a negative sense.
Doctors in training saw management of obesity being ignored;
it was not specifically addressed as a training issue: "It
wasn’t discouraged, but it wasn’t part of the program."

Two doctors mentioned the importance of their own life
experience and use themselves as role models for their

patients: "I've learned through my own successful and

87
unsuccessful experiences with diet and exercise" and "I’m an
exercise nut."

Persuasion. Exhortative encouragement during training,
during medical school, internship and residency, was
lconsidered by few doctors as a source of efficacy
information. It was not a meaningfully remembered part of
the medical education for eight doctors. Four specifically
remembered that it had no place in their medical educations;
these doctors’ training occurred historically earlier than
that of other subjects. Three were not sure if it had been
a part of medical education or not, but knew that it had not
been stressed.

One doctor remembered an occasional lecture about the
causes of obesity, but nothing about the management of
obesity. The two doctors whose training site or supervision
had specifically included an obesity orientation did report
that their self-efficacy was fostered by encouragement
during training.

In general, doctors who trained more than 15 years ago
had not had encouragement to consider obesity treatment to
be a significant part of physician-provided medical care.

Physiological Cues. Emotional reactivity, as reported
by subjects, represents cues about self-efficacy in the form
of autonomic reactivity. All study doctors considered
obesity treatment to be difficult; all study doctors stated

that they have felt more comfortable over time, as they have

88
gained experience with treating obesity. Enactment
behaviors have increased self-efficacy.

Four of the 10 doctors stated, unequivocally, that they
enjoy treating obese patients, get satisfaction out of it
and have a strong (though tempered) sense of their own
effectiveness. Two are self—described "true-believers," who
see themselves as proselytizers for healthy eating and
exercise behaviors. One has a social mission and sees
providing competent and compassionate obesity treatment as
treating an under-served group. One doctor has a deeply
felt sensitivity to and respect for patients whose sense of
well-being is dependent on their being heavy.

Two other doctors describe treating obese patients with
enthusiasm. Their outcome expectations are frankly low and
their feeling of mastery is unusually independent of
outcome. For one, early training experiences in obesity
treatment were prominent. This doctor describes feeling
"resigned," not demoralized about treating obesity and sees
it as just an ordinary "problem of living."

For the other, obesity treatment is called "futile" but
not "frustrating." The focus of obesity treatment is on
alleviation of a specific medical condition which would be
ameliorated by a loss of weight. Focus is at a very
concrete level and on limited goals, e.g. a change in a lab

value. These goals are shared with the patient.

Fh.

PL;

T.

Pb

FL

a.»

*4

89

Four doctors expressed neutral or negative feelings
about treating obese patients. Reasons given include
patients not being interested; patients only wanting to talk
and not make behavior changes; lack of clarity for
themselves (and for their patients) about whether obesity
without concomitant medical problems is really a medical
problem; lack of a clear medical treatment alternative, if
patient behavior does not result in desired outcomes or
results in outcomes which are temporary and, lastly, fear of
offending patients.

Other Sources of Efficacy Information. The importance
of protecting and maintaining a good doctor-patient
relationship was spontaneously mentioned by most study
doctors. Expressed positive patient response to physician
treatment approach appears to be a primary criterion by
which self-efficacy is judged.

An obese patient who is not well~engaged in making
eating or exercise behavior changes or, who is engaged but
is not making clear progress but who continues to stay in
treatment and at least be monitored or who is not gaining
weight or whose hypertension is not badly out of control,
may still engender feelings of self-efficacy in his or her
physician.

Physicians find patients’ liking their approach to be

rewarding in and of itself. Success in the doctor-patient

9O
relationship is a desired outcome and can be construed as
independent from any other outcome.
How the Problem of Obesity is Defined

One of the conclusions drawn from the interviews is
that how the problem of treating obesity is defined,
determines what goals are set for treatment, how they are
arrived at, how progress toward them is measured and how
evidence which potentially disconfirms treatment self-
efficacy is handled.

Pattern One. Four study doctors consider obesity
treatment to be necessary primarily when obesity is
accompanied by, or is a partial cause of, medical problems
which would be ameliorated with a loss of weight and an
increase in activity. Problems most commonly mentioned in
this regard are hypertension, diabetes and osteoarthritis.
Goals for obesity treatment and measuring progress toward
them flow in an orderly manner from treating this "other"
medical problem (e.g. hypertension).

Dietary control of hypertension without medication is
an example. The patient’s blood pressure values, their
placement in a healthy or pathological range and the change
in values over time are the treatment focus. These may be
shared in specific detail with the patient. Control of the
"other" health problem (hypertension) is measure of
treatment success, not control of weight. If dietary

control is unsuccessful for any reason, medication treatment

91
of hypertension is implemented, with similar specific
monitoring of blood pressure values.

Successful treatment is, therefore, possible under a
plethora of doctor and patient treatment behaviors and
outcomes. This view of obesity treatment was characteristic
for doctors with “inconsistent" self-efficacy/outcome
expectations scores.

Pattern Two. Four doctors consider obesity treatment
to be an important part of on-going, preventive medical care
and, as such, important for any patient who is obese. The
primary measure for treatment success, in this case, is
weight loss, though increase in activity level is also a
measure. Increased activity level is not as verifiable and
is, therefore, not as powerful a shared, "socially mediated"
treatment indicator as is weight loss. The onus is
necessarily on patient behavior. The physician’s role is
one of educator, facilitator, or even coach.

The problem comes, for the physician, with how to
handle evidence that disconfirms his or her self-efficacy,
i.e. when a patient does not verifiably change eating
behaviors or increase activity level (does not lose weight).
For a patient who is not very obese, (perhaps only 20% over
recommended weight,) who has no current medical sequelae to
his or her obesity and whose level of subjective distress is
not high, the physician is left with self-doubt, when the

patient does not lose weight.

92

Both physician and patient can get lost in the "maze"
of health maintenance: for both patient and physician, in
their hearts they know it is right, but the patient "fails"
in terms of progress on the criteria to which both the
patient and the doctor subscribe. The patient’s failure
then becomes the physician’s and they are both left feeling
lost and inefficacious.

Pattern 2a. One way out of the "maze" is to focus on
the "bigger picture." Goals, then, become attached to
something more general than those which are part of a
particular practice with a particular patient. An example
of such a transcendent framework is a belief that exercise
is good and that one is a crusader for that good. An
obesity treatment "true believer" can integrate failure, in
a manner that protects self-efficacy, in the same way that a
missionary can. Two study doctors, both with consistent and
high self-efficacy and outcome expectations scores
illustrated this approach to obesity treatment.

Pattern 2b. Another path out of the "maze" is for the
physician to frame the problem of obesity as psychological
and to look for the "meaning" behind the "symptom." The
treatment, then, flows from this underlying meaning. This
focus could be on any distressing emotional sequelae of
obesity or of treatment. Emotional concomitants of obesity
and obesity treatment include anxiety, depression, feelings

of shame and poor self-esteem.

93

The overall treatment goal, then, becomes simply for
the patient to feel better. The means which can be employed
to achieve this end are extremely varied and include
referral for psychiatric evaluation or psychotherapy,
provision of supportive counseling and/or provision of a
structured program to provide eating and exercise behavior
changes. This is, at times, provided by the physician him
or herself within the format of primary care, or a referral
is made to community resources for structure and support.

The criterion for progress in this approach is
broadly defined because "feeling better" can be
individualized and is subjective. It can be also most any
positive attitude or behavior change. The physician’s self-
efficacy is somewhat protected from disconfirming
information because part of the job is just to try, to "be
there" with the patient, independent of treatment outcome.
This way of conceptualizing the problem and the treatment
cuts across subgroups based on self-efficacy and outcome
scores; it characterized two consistently high doctors and a
consistently low doctor.

A variation of this psychological "reframing" of the
problem is treatment of a significantly obese patient who
has presently no medical sequelae, but who has multiple risk
factors for these sequelae in the future. This patient
associates being obese with a feeling of well-being and

doesn’t feel good when she or he is not. The physician

2:

wt

94
needs not only to educate the patient about the link between
current habits and future quality of life, but will be
pushed to adjust his or her own standards of care and health
to include a diversity of body weights.

One study participant mentioned that the D.O.
philosophy holds that the power of healing, regeneration and
recovery lie within the body itself. This belief could also
support the physician’s ability to see the "bigger picture,"
and maintain self-efficacy, particularly in the face of
disconfirming feedback. This viewpoint is also supportive
of D. O.’s having optimistic outcome expectations.
Doctor-Patient Matching

Consensus between the doctor’s conceptualization of the
problem and the treatment and the patient’s
conceptualization are described by the doctors as important
for their self-efficacy. The interplay between development
and maintenance of the doctor-patient relationship and
development and maintenance of self-efficacy in obesity
treatment seems very close.

Enactment behaviors, particularly those which engender
positive patient response are described as very powerful for
the physician. There may be an on—going, implicit process
of matching which goes on, so that patients "self-select"
out of a doctor’s practice if that doctor is not, from the

patient’s point of view, treating the "right" problem. This

95
"matching" should have the effect of supporting physician

self-efficacy.

Integpretation (Qualitative)

What High and Low Self-Efficacy Doctors Have in
Common. Even given that the interviewees were self-selected
on the basis of their not being reticent and inexpressive,
the degree of opinionatedness and strong feelings expressed
was striking. Self-efficacious or not about obesity
treatment, all study subjects had a lot to say about it and
said it with strong convictions.

All study subjects considered treatment of moderate to
severe obesity (25% or more over ideal body weight) to be a
common problem, affecting between 1/4 and 1/3 of patients
seen in their practices.

All interviewees considered obesity treatment to
necessitate judgment and behaviors which have a high degree
of difficulty. All considered obesity treatment to be as
hard as or harder than treatment for other chronic problems
which are dependent on patient behavior change.

Doctors referred patients to mental health
professionals for help in treating symptoms of major mental
illness, for eating disorders of anorexia nervosa and
bulimia nervosa and sometimes for depression and anxiety.

They did not, generally, refer their patients for treatment

‘L

n»

1“

Hi.

‘5‘.

96
of obesity. They did not seek psychological consultation
for themselves in dealing with their obese patients.

Doctors view patients as referring themselves to
community programs for treatment of the problem of obesity
or treating themselves with information garnered from the
mass media, rather than presenting themselves to their
doctors for treatment of obesity as a medical problem.

All doctors thought that management of chronic disease
was not well-covered in their medical educations. Some felt
this was a real deficit because this would have been helpful
in teaching them about management of obesity.

No physician felt it was appropriate not to treat
obesity "just" because they don’t like it. It was rare for
obese patients to referred out of a practice for primary
care elsewhere. Doctors who said that they like treating
obesity think that they are unique and imagine that their
colleagues generally like it less than they do.

Attributions That Foster Self-Efficacy. The two main
issues that appeared to block the development and
maintenance of high self-efficacy obesity treatment in
physicians (besides the absence of experience) were
ambiguity about the problem (and, therefore, the treatment
goals) and coping with disconfirming evidence of one’s self-
efficacy beliefs and one’s outcome expectations. The

ability to resolve issues of ambiguity and disconfirmation,

97
therefore, was important in the fostering of one’s self-
efficacy.

There appeared to be two main ways of conceptualizing
competence in this treatment area that foster resolution.
One was to consider obesity to be a problem that was a part
of other health problems a patient had and that treating it
(including considering the appropriateness of treating it)
was part and parcel of treatment of the other medical
problems. Medical "ownership" of the problem was clear, as
were the goals of obesity treatment. If treatment was
unsuccessful, there was a medical "fallback" position, in
treatment of the concomitant medical problems via
medications or other medical means.

The other conceptualization of competence was to see
obesity treatment as a part of on-going preventive primary
medical care. This had more ambiguity in terms of treatment
goals and more to deal with in terms of coping with
disconfirming evidence of one’s competence. Medical
"ownership" was less clear. Identification with an
overarching powerful idea, that transcends particular
problems with particular patients, helped in resolving
ambiguity about treatment goals and gave meaning to
disconfirming evidence.

"Health maintenance" as the overarching idea seemed not
to work very well, although it's time may yet come,

particularly if medical education increases its emphasis on

98
management of chronic conditions and if it is given priority
in health care system reforms. Faith in exercise as a
religion appeared to work better in resolving issues of
ambiguity and disconfirmation.

A broad reframing of the problem, to include mental
health goals and a plethora of behavior change goals was
another preventive conceptualization of treatment
competence. The number of treatment criteria increases, but
many of these are ambiguous and difficult to measure. They
are often "internal" or private changes, rather than public
ones that can be shared and which can more easily reinforce
the doctor's self—efficacy. There may be the possibility of
more "successes," but no one knows who they are.

Attributions That Undermine Self-Efficacy. There are
some patient and problem characteristics that hindered self-
efficacy. Apart from "feeling better" being an ambiguous
goal, it is also not invariably attainable. Some patients
will not ever "feel better." Some problems, including
personality problems are very resistant to treatment. This
treatment goal cannot be met by every patient and relying on
it will undermine one’s feeling of competence. Even if
treatment goals are modest, the forces that maintain
unhealthy behaviors may be more powerful than the most
skillful treatment and care.

For some patients, an empathic approach that works

towards a doctor-patient alliance is viewed negatively. It

ubb‘ 4

99
may make a patient angry. The whole feedback loop of
relationship—development and efficacy-building gets
interrupted.

Food and eating behaviors have a multiplicity of
meanings for doctors as well as for patients. The
subjectivity of this area is illustrated by this quote from
one questionnaire respondent: "An obese patient is fatter
than his/her doctor. I’m quite fat, I have proportionally
fewer obese patients."

If food is a valued aesthetic experience for a doctor
and treatment outcome expectations are modest at best,
ambivalence about treatment may undermine self-efficacy.
One study doctor wrote about cholesterol-lowering studies:
"the effectively treated groups have substantially greater
risks of other important health problems. A program which
changes patients’ lives so much more than cholesterol
reduction might bring on other health risks - medical or

psychiatric."

AV

oi.

b-

Chapter V
Discussion

The goal of the study was to determine if the construct
of self-efficacy is a useful one for studying primary
medical care treatment of obesity. Towards this end, an
instrument was formulated and tested that measured obesity
treatment self-efficacy. Evidence of its reliability and
validity was been presented. Demographic and practice
variables were tested for their association with self-
efficacy scores.

In addition, physicians’ own attributions about the
development and maintenance of their self-efficacy in this
practice area were explored to see what, of theoretical or
practical interest, could be learned from conceptualizing
medical practice in this way.

Research Question One

The first research question--can self-efficacy in
obesity treatment be measured with a degree of reliability
and validity?--can be answered in the affirmative. Scores
on the Physician’s Self-Efficacy Measure (the PSEM) showed
variability; they were roughly normally distributed.
Individual item scores had uniform variability, the scale
was highly reliable and it had a parsimonious factor

structure. Factor subscales were readily interpretable in

100

101
terms of self—efficacy theory. The scale predicted self-
efficacious physician behaviors, in a manner consistent with
theory.

The self-efficacy scale has three factors. The most
important factor, as measured here--competence in
interpersonal skills--accounts for 40% of the variance in
scores. Competence in one’s professional knowledge base is
also a factor in self-efficacy, accounting for 16% of the
variance in scores. Both of these factor subscales have
good reliability and are nearly as powerful in predicting
physician behaviors, as is the total self-efficacy score.

The affective subscale of the PSEM accounts for 12% of
total scale variance; it is reliable, but differs from the
other two subscales in that it does not, on its own, predict
self—efficacious physician behaviors.

The scores on the second self-efficacy measure, the
Physician’s Outcome Expectations Measure (POEM), also showed-
a variable range. Their distribution is positively skewed:
doctors generally have lower, rather than higher outcome
expectations for obesity treatment. POEM items have uniform
variability and the scale has moderate reliability. The
POEM also predicts self—efficacious physician behaviors.

Factor subscales of the POEM are less parsimonious than
those of the PSEM and are more difficult to articulate
clearly in terms of self-efficacy theory. The first factor

of the outcome expectations measure-~obesity is treatable

4‘

a§~

‘. 4
.3

m:

102
and is a shared responsibility between doctor and patient--
and the second factor--obesity treatment is effective and it
is important--were reliable and were as good at predicting
self-efficacious physician behaviors as is the total POEM
scale. These subscales, together, account for one third of
the total variation in scores. The other three subscales of
the outcome expectations measure have poor reliability and
poor predictive power for physician behaviors.

The correlation between the two measures (self-efficacy
and outcome expectations) is positive and is moderately
strong: (I = .46) and the first and second factor of each
measure correlate moderately strongly with each other and
with each total measure score. For example, the first
factor of the PSEM (interpersonal skills) correlates not
only with the second factor of the PSEM (a competent
professional knowledge base) and with total PSEM score, but
it also correlates with the first factor of the POEM
(treatability of obesity and shared responsibility of
obesity treatment) and with the total POEM score.

Beliefs about one's interpersonal skills and
professional knowledge base about obesity treatment (factors
one and two of the PSEM) are closely related to beliefs that
obesity is treatable and is a shared responsibility between
doctor and patient, and that treatment is effective and
important (factors one and two of the POEM). The degree of

relatedness between self-efficacy factors and outcome

103
expectations factors, as well as total self-efficacy and
outcome expectations scores, is moderately strong. This is
consistent with self-efficacy theory: beliefs and behavior
are linked, but there is not a perfect correlation; they are
not independent, but they are also not identical.
Study Goal Two

High percepts of efficacy and high outcome expectations
are, generally, not characteristic of any group with regard
to demographic or practice variables. This information
meets the second goal of the study. It supports that study
findings can be generalized to primary care physicians,
across these demographic and practice characteristic
variables. It is evidence for the lack of a significant
identified confound in the relationship between self-
efficacy beliefs and self-reported, self-efficacious
behaviors.

The more patients a doctor sees per week, the higher
she or he scores on the self-efficacy measures. In terms of
the theoretical base of the study, these physicians are
engaging in more "enactment behaviors" and their higher
scores are, therefore, expected. The experience of doing a
behavior and the knowledge of one’s capability at this
behavior should mutually influence each other. Beliefs
about one’s capabilities are one of the determinants of
one’s behavior. At the same time, one’s behavior provides

opportunities to gain a sense of competence.

104

A major study strategy was to determine to what degree
beliefs predict behaviors. In Bandura’s theory, beliefs and
behavior mutually influence each other and they develop
together, in tandem, with each reinforcing the other. One
could have, as logically, tested whether behaviors (i.e.
physician-reported practice behaviors) predict beliefs (i.e.
self-efficacy measure scores.)

Study Hypotheses

In addition to these overall study goals and research
questions, specific hypotheses were tested regarding the
relationship between beliefs as measured by the PSEM and
POEM and physician behaviors. Behavioral measures designed
to test the validity of the study belief measures supported
the validity of these measures.

Hypothesis One. High self-efficacy doctors see a
higher proportion of obese patients in their practices each
week, i.e. a higher proportion of the patients seen by high
self-efficacy doctors are obese. Self-efficacy beliefs
influence behavioral choice of which patients are seen and
which problems are identified for them.

Hypothesis Two. High self-efficacy doctors discuss
obesity treatment with a higher proportion of their obese
patients. Self-efficacy beliefs influence behavioral choice
and effort.

Hypothesis Three. High self-efficacy doctors initiate

a discussion of obesity treatment with a higher proportion

105
of their obese patients. Self-efficacy beliefs influence
behavioral choice and effort.

Hypothesis Four. High self-efficacy doctors have a
higher proportion of their obese patients on some kind of
weight control program. Self-efficacy beliefs influence
behavioral choice, effort and persistence.

Hypotheses Five and Six. These behavioral variables
were not used in the analysis. Hypothesis five (proportion
of patients on which consultation is sought) because of
inconsistency in subject responses and hypothesis six
(proportion of patients referred elsewhere for primary care)
because of the low base rate of this behavior.
ResearchBQuestion Two

Exploring how doctors view their own self-efficacy--how
they understand the problem of obesity, how they see their
own treatment behaviors and what outcomes they expect from
these behaviors-—was done via interviewing a small number of
doctors. They were asked, also, what they felt fostered
and what they felt undermined their self—efficacy in obesity
treatment.

It was important to do this interviewing because the
formulation of the study measures showed that there was,
indeed, a difference amongst physicians in their self-

efficacy, but it did not explain much about this difference.

106

Sources of Efficacy Information. "Behavioral
enactment" or "experience" is the source of information most
relied on for the development and maintenance of self-
efficacy. This finding is consistent across both the
qualitative and quantitative phases of the study and ties
the two parts of the study together. This finding is
consistent with self—efficacy theory. Development and
maintenance of the doctor—patient relationship is
behaviorally closely linked to self-efficacy.

Alternative methods for developing self-efficacy
according to theory-~"vicarious learning" and "exhortation"-
-are underutilized in medical training and practice. Role
modeling occurs fortuitously, not by design. Persuasion is
similarly underutilized in training. It is experienced
more, post-graduately in practice, via continuing education.

"Autonomically-mediated learning" (represented by
reported emotional reactions) was not related to the
behaviors that doctors reported. It was not, therefore,
related to self-efficacy in physicians. Feelings were
strong across doctors, independent of their self-efficacy.
Intensity of feelings--positive ones of satisfaction and
negative ones of frustration--appear to be characteristic of
many primary care doctors engaged in obesity treatment,
whether they are self-efficacious or not.

"Generalization" of efficacious behaviors, whereby

according to theory, skills can be generalized to similar

107
behavioral areas through the transfer of skills, is not
being capitalized on. The management of chronic disease--
the larger area of which obesity treatment is a part—-is
neglected in most medical training.

One study finding in the quantitative analysis was
strongly counter-intuitive and was explored during the
interviews. The "affective" scale on the self-efficacy
measure was ineffective in differentiating between strong
and weak self-efficacy doctors. This finding was consistent
across the quantitative and qualitative phases of the study.
Exploring this issue via interviewing, in the qualitative
phase of the study did not clarify this finding: all
interviewed doctors had strong feelings about their patients
in this area.

One interpretation of this finding is that feelings
really do not matter as much as mental health professionals
think they do. For providers of a service in a complex
practice area, interpersonal skills and a competent
professional knowledge base are more important for self-
efficacy, than is how the providers feel about their clients
and their clients’ problems. Affect is independent of self-
efficacy.

Physicians strive to conform to a model of professional
practice and emulating this model becomes a source of self-
efficacy for them. They do not, it appears, rely on their

own personal reaction to a patient or a patient problem. In

108
a "professional practice model" of self-efficacy, items on
the self—efficacy measures which pertain to the substance of
professional practice (e.g. knowledge) would get endorsed
and those that do not, (e.g. items about feelings,) would
not.

Another interpretation of this finding is that many
doctors--both and high and low self-efficacy doctors—~have
strong, personal feelings, but they are the same feelings
rather then different feelings. Even the high self—efficacy
doctors, in the interview, used words like "futile" to
describe their feelings about obesity treatment. The
affective continuum, for this study, was conceptualized as
"positive" to "negative" feelings about obese patients.

Because feelings items did not differentiate between
high and low self-efficacy doctors, does not mean that they
are not important. Though the affective scale on the
measure had reasonable reliability, it may have had poor
validity. It could be that another "feeling continuum"
should have been used, e.g. "mild" to "intense" feelings or
feelings on a continuum from "professional uncertainty" to
"professional mastery."

It may simply be that a doctor needs to have "positive
enough" or "intense enough" feelings to allow self-
efficacious behaviors to be developed and enacted. If
liking obese patients a certain amount but no more is

necessary, then a relatively small amount (e.g. an amount

109
easily mustered by a professional in practice) may be
(sufficient for physician self-efficacy.

Patterns of Obesity Treatment Experience. Different
ways of viewing the problem of obesity lead to different
ideas about treatment means and goals, different ways of
measuring progress toward goals (or conceptualizing
"success,") and different ways of dealing with information
that disconfirms one’s self-efficacy (or conceptualizing
"failure").

The patterns did not differentiate between high and low
self-efficacy doctors but, rather, provided a conceptual
context of practice in which the dynamics of self-efficacy
development and enactment were played out. Principles of
self-efficacy development and enactment can also be
discussed in these patterns of experience.

"Illness Eradication" Pattern. Viewing obesity as a
precursor to, or sequela of, another medical problem
provides a doctor with clear, unambiguous goals for
treatment and a strong incentive to treat. It is also
provides clarity of steps to attain the goals, e.g.
restricted calorie diet, increased activity level. The
success or failure of treatment is shared, explicitly,
between doctor and patient. It is often shared, socially,
in a broader sense, because patient behavior and even

appearance changes (or fails to change.)

we

he

5

b

anc

110

In the illness eradication pattern, there is a tight
feedback loop between efficacy belief and actual treatment
outcome. There are also medical alternatives to treatment
failure, so that treatment self-efficacy can be maintained
in the face of disconfirming evidence.

"Health Maintenance" Pattern. Viewing obesity as a
problem in and of itself, which is not a part of a
concomitant health problem, is a different undertaking for a
doctor. This conceptualization does not have the full
weight of the medical profession behind it and, therefore,
professional incentives for obesity treatment are less.

Health maintenance, as an end point, does not provide
the same degree of clarity of goal, steps to attain the goal
and clarity of treatment modality--particularly at the level
of the individual patient--as does treatment of a specific
pathology.‘ A goal which it does provide, i.e. maintenance
of a normal body weight over the life span, is extremely
rarely attained by obese persons. At the level of the
individual patient, failure is very clear; progress is less
clear.

In a health maintenance approach, the feedback loop
between behavior and outcomes is less tight. Outcomes, in a
positive sense, are less shared and, consequently, less
reinforcing of physician self-efficacy. There is no
convenient way of dealing with evidence which undermines

one’s self-efficacy.

‘5
is

2 : F:

111

A health maintenance approach would work better, if
practice is conceptualized in a "public health" model of
health care delivery, rather than in a "private practice"
model of health care delivery. The goal, in this
conceptualization, would be population behavior change,
rather than individual patient behavior change. Individual
patient "failure" would not disconfirm physician self-
efficacy because it would be the "success" of the whole
population, not the individual patient, which would be the
primary concern.

Focusing on concomitant mental health sequelae of
obesity, as a part of health care in the broadest sense,
provides a plethora of treatment goals. However, the goals
may be ambiguous and progress toward them difficult to
measure. The link between physician behavior and treatment
outcome, is obscured, particularly information which serves
to support self-efficacy. The link between physician
treatment and failure of treatment may well be clearer. In
fact, physicians may only routinely get feedback on their
treatment when it fails and not, consistently, when it
succeeds.

Features That are Independent of Pattern and of Self-
Ef_fi_cm

There are some aspects of obesity treatment behavior

which appear to be consistent across pattern of obesity

treatment experience and across high and low self-efficacy

dc

HE

St

.0.

112
doctors. In terms of self-efficacy theory, the "level" of
necessary behaviors is "difficult." All interviewed doctor
stated this.

On physician report, obesity treatment behaviors
generalize to other areas of management of patients with any
kind of chronic disease or condition. Physicians also
report that consensus or "matching" between doctor and
patient on the definition of the problem is important
because it supports self-efficacy and, more specifically,
its absence undermines self-efficacy.

Treatment goals which are specific and which are shared
facilitates the feedback loop between physician behavior and
its different levels of expected outcome: patient behavior
change and, in turn, patient health improvement.

Conditions exist in the obesity treatment which should,
according to self-efficacy theory, promote misjudgments
about one’s capabilities and undermine the development of
self-efficacy. Goals for treatment are often ambiguous and
the quality of one’s performance is often unclear.
Therefore, making goals as concrete as possible and
formulating them so that there exists a clearly articulated
feedback loop between behaviors (physicians’) and outcomes
(patients’) would promote self-efficacy development and
maintenance.

A Model of Physician Self-Efficacy in the Treatment of

Obesity

113
A summary of the features of self-efficacy theory
which, when applied to obesity treatment should promote the
development and maintenance of self-efficacy according to
study findings follows.
Physician Self-Efficacy = Experience or "enactment
behaviors," that are optimized by [ professional incentives '3

(persuasion, role-modeling) + goal-setting (specific,

 

measurable, consensual) + clear feedback loop (including a 5
means of dealing with disconfirming evidence) ] + individual '~
differences + error. I

Practical Applications

Given this model, guidelines from theory can be applied
to this practice area and could go across "illness" and
"health" conceptualizations of obesity treatment.

These include: 1.) the formulation of treatment goals
which are specific, measurable and about which there is a
high degree of consensus between doctor and patient;

2.) the integration, within the treatment approach, of
specific provisions for the completion of a feedback loop
between physician behavior and outcomes of physician
behavior, (often patient behavior);

3.) the formulation of a way of dealing with patient
behavioral "failure," so that it is not construed as
disconfirming evidence of physician self—efficacy; and

4.) the support for links that tie physician beliefs

and physician behaviors together.

114

Clarity of goals and clarity of feedback on whether
they are attained, support for links which tie beliefs and
behaviors together, and a means of dealing with evidence
which disconfirms one’s self-efficacy are all prominent
features of self—efficacy theory.

Specific guidelines from theory, which could serve to
tie beliefs and behaviors together include: a.) the
teaching of necessary "subskills" of obesity treatment
behavior (e.g. interpersonal skills),

b.) providing clear professional "incentives" for
treatment, (including obesity treatment being remunerated by
health plans and its having a more legitimized place in
medical training); and

c.) providing more explicit "social pressure," so that
doctors are encouraged to have a high personal investment
in, a feeling effectiveness in and a means of being rewarded
in this treatment area.

Other implications for medical education, which are
specifically gleaned from the interview data, include:

1.) tying teaching about doctor-patient relationships
to self—efficacy training in practice areas which rely on
interpersonal skills, since self-efficacy appears to be
intimately connected to these skills;

2.) using role-modeling and encouragement for obesity

treatment in medical school and residency;

115

3.) encouraging physicians to resolve subjective issues
that they may have regarding experiences of food and eating
behavior, so that their own feelings of helplessness about
patient failure or ambivalence about treating obesity
because of not wishing to be punitive with patients, do not
interfere with their treatment self—efficacy.

Counseling psychologists could have a role in the
"relationship skills" training of primary care doctors, in
general, and specifically, in the teaching of behavior
management skills to doctors.

Theoretical Implications

The original idea of studying physician self-efficacy
because of its possible influence on patient self-efficacy
(and, in turn, on positive patient health outcomes) is
oreinforced by this study. Almost without exception,
interviewed physicians, as well as physicians who wrote
comments on their returned questionnaires, spontaneously
described their self-efficacy in obesity treatment as being
dependent on their overall relationship with their patient.
They report that there was something like an interplay
between their patients’ self-efficacy and their own.

This emphasis on the interpersonal relationship did not
appear to be a based on the physicians voicing a fashionable
catch phrase, but on stating a fact of life. Obesity

treatment was virtually never considered independent of the

116
doctor-patient relationship, because it doesn’t "work"
(isn’t effective) any other way.

This is one aspect of primary care medical practice
that is similar to mental health practice. In mental health
treatment, it is known that the relationship between
therapist and client is important. In research on the k?
differential effectiveness of various mental health

treatment modalities, the importance of the "therapeutic" or

 

"helping alliance," (the patient’s experience that the y-
treatment or the relationship with the therapist is helpful
or potentially helpful) is documented in the research of
Luborsky, Crits-Christoph, Alexander, Margolis and Cohen
(1983), among others. It has been identified as one of the
"common features" (Stiles, Shapiro, and Torrington, 1986) of
treatment effectiveness that is consistent across treatment
approach. It appears to be a kind of "necessary condition"
for therapy to take place, independent of other features of
the therapist, the client or the type of therapy.

The interviewed doctors, in the present study,
described forming a "therapeutic alliance" with their obese
patients and that it was within this alliance that obesity
treatment took place.

Luborsky, Mintz, Auerbach, Cristoph, Bachrach, Johnson,
Cohen and O’Brien (1980) have also looked at the similarity
between therapist and client and found the "match" between

client and therapist was positively related to treatment

117
effectiveness. In obesity treatment, the primary care
doctors described that a consensus between themselves and
their patients on what the problem and goals of treatment
were was important for the maintenance of obesity treatment
self-efficacy. They described the importance of the "match"
between themselves and their patients.

The importance of the relationship between therapist
and client, in counseling and psychotherapy, is well
established in research, but the importance of the
relationship in medical practice between doctor and patient
is not. The doctor-patient relationship is typically
considered adjunctive rather than central to medical
treatment. Its influence on patient health outcomes has not
been studied. It is rare for it to be stressed in medical
training; students whose relationship skills are poor do not
fail to graduate from medical school. It is not a basis for
medical malpractice suits.

The importance of relationship may be central to
‘medical treatment, especially in areas where a treatment is
not unitary and straightforward, where treatment goals are
‘unclear or rarely met, and where disease management is
highly dependent on patient behavior. Using a framework
from self-efficacy theory to focus on interpersonal factors
‘within health care has the advantage of an established
research base to draw from, including research about patient

self-efficacy.

118
It is only a small step, from physician self-efficacy,
to think about the influence of physician self-efficacy on
patient self-efficacy (the doctor’s ability to foster or to
undermine patient self—efficacy) and, in turn, the
relationship between physician self-efficacy, patient self-

efficacy and patient health outcomes.

Appendices

Appendix A

PHYSICIAN SELF-EFFICACY MEASURE

completely 5
quite a bit 4
moderately 3
a little bit 2
not at all 1

How confident are you that you can:

Diagnose obesity?

Select and explain a weight loss treatment
plan to obese patients?

Deal with any negative feelings you may
have regarding obese persons, so that

they do not interfere with your effectiveness
in treating obese patients.

Help obese patients to deal with their
weight problem?

Diagnose obesity with health risks?
Provide a compelling rationale to obese
patients for making changes in their diet
and exercise habits?

Adjust your advice to obese patients
depending on their circumstances?

Help obese patients to lose weight?

Explain the health risks related to
obesity to patients?

10 . Engage obese patients in a treatment

plan to lose weight?

11 . Maintain a good relationship with obese

patients who do not lose weight.

119

Appendix B

PHYSICIAN OUTCOME EXPECTATIONS MEASURE

completely 5
quite a bit 4
moderately 3
little bit 2
not at all 1

Indicate your agreement
with the following statements :

1 . Obese patients who fafl to lose weight are at
significant health risk.

2. Losing weight can make a significantly
positive difference in a patient's life.

3. Obese patients who restrict their calorie
intake can usually lose weight.

4. Obese patients do what they want, rather
than following their doctor's advice
regarding diet and exercise.

5 . Most patients who conscienciously try to
change their diet and exercise habits
are able to lose weight.

6. Physicians have influence on whether
their obese patients change their
diet and exercise habits.

7. Doctors can be effective counselors
(educating patients about their obesity,
engaging them in weight loss programs, -
monitoring their progress) for their patients
who are trying to lose weight.

8. Physician lack of effectiveness in obesity
treatment is a factor in a patient's
failure to lose weight.

9 . Most patients will never even begin a weight
loss attempt .

10. Most patients will begin an attempt to
change their diet and exercise habits,
but will not lose any weight.

11. Most patients are able to lose weight,
but will gain most of it back.

12. Most patients are able to lose weight and
to maintain the weight loss .

13 . Whatever doctors or patients do,
weight loss is only temporary .

14. Effective weight loss treatment is largely
via pharmacalogic agents or surgical
measures, rather than physician guidance.

 

120

Appendix C

PHYSICIAN PRACTICE MEASURE

1.
2.

5.

8.
9.

Specialty: Fam. Pract._ Int. Med._ Gen. Pract._.

 

 

 

Age 3. Sex: M F 4. Year of med. sch.
grad.

Which type of practice best describes your own:
private practice , university clinic-based ,
HMO , VAMC , PMS/IHS , USAF .
Which size of practice best describes your own:

solo _____, 2-3 doctors , 4—6 doctors , more than 6
doctors__.

How many patients do you see in a typical week?
none , less than 20 , 20—75 , 75-125 ,
125-175 , 175-225 , more than 225 .

What percentage of your patients are female? .

Do you have obese patients in your practice? yes ,
no .

10. If you do, thinking back, over the past month,

a.) Approximately how many patients did you see who were
obese?

b .) With approximately how many of these patients did you
discuss weight loss?

c.) With approximately how many of these patients did you
initiate a discussion of weight loss?

d.) Approximately how many of these obese patients are
currently on a weight loss program of some type?

 

11. Thinking back over the past year,

a.) For how many of your obese patients did you
seek consultation from a medical colleague, a
psychologist or consult the medical literature about?

 

b .) How many of your obese patients did you
refer out of your practice and to another physician for

on—going primary care?

 

12.) Do you currently have a patient in your practice that

you would like to refer for medical care to a physician
who specializes in treating obesity? yes __ no _.

Would you be willing to be interviewed on this

subject? yes—no

Please make any additional comments you have on the back of
this sheet.

121

Appendix D

Physician Self Efficacy Interview Guide

The self—efficacy interview is used in different ways.
In addition to the structured questions which are asked
everyone , I can use the written answers that the person had
already provided as a spring board and ask the person to
elaborate on points of interest.

The main body of questions is right out of the theory.
I ask about choice of, effort in and persistence in
behaviors which pertain to obesity treatment. I ask about
the perceived level of difficulty of obesity treatment
behaviors. I ask about the strength of efficacy beliefs,
particularly in the face of disconfirming experience. I ask
about generality of beliefs, in comparing them to physician
beliefs in their efficacy in treating other chronic health
problems, particularly those requiring dilligent patient
effort and behavior change .

Short Introduction:

Thank you for agreeing to talk with me about your
practice with obese patients. I would like to begin by
asking you some general questions, and then, there is some
specific ones that I'd like to ask also. Please feel free
to elaborate, on these questions or on your answers, as we
go along, or to ask me questions. If there is any question
you would not like to discuss, please say so. What I am
interested in is how you feel about your practice, your own
perceptions about your work with obese patients . Do you
have any questions to begin with? ok, let's start.

General Questions:
1 . )How do you feel about treating obese patients? What's it
like?
prompts - I'd like a general idea, are there things you
like about it, things you don't like about it? Funny
experiences, frutratlng ones, ones where you feel a real
sense of satisfaction? a real feeling of failure?
2.) What has been your hardest case (or cases) to treat?
3 .) What's been the easiest case to treat?
4.) What's a typical case (or cases) like, in your practice?
5.) how important do you think it is, to address the problem
of obesity ?
6 .) How important do you think it is, to other doctors in
your practice , to address the problrm of obesity?

Thinldng of yourself and what is typical behavior for
you, with patients in your practice, I'd like to ask you
about some of the specifics about treatment:

122

._P""

123

7.b .) Do they ever initiate it with you?
if 7.a.) is yes, then,
8.) How do you decide when (and whether) to initiate a
discussion of weight loss with an obese patient?
9.) Are there times when you decide not to address it?
Under what circumstances?
10.a.) What kind of treatment goals do you set for/with your
obese patients?

10.b.) How do you decide on these goals?
11.) How do you measure success in treating obesity, i.e.
how do you know when the treatment is working?
12 .) are there other measures besides weight loss?
13.) How do you know when you're doing a good job, in
treating your obese patients?
14.) How do you decide whether to continue addressing the
problem of obesity with patients or to stop addressing it?
15.a.) If an obese patient continues coming to you for
treatment, but does not lose weight, even though it is
medically recommended, how do you handle it?

prompts: b .) do you seek consultation from colleagues?

c .) from the medical literature?
d.) refer to a community weight loss program,
e. g. Weight Watchers?

e .) do you seek other sources of assistance ,

which you've found to be helpful?

f.) is it frustrating, when this happens? If so,
what helps you to cope? How does it affect
your relationship with the patient, when this
happens?

9.) If it's not, what do you think helps you not

to get frustrated?

15.) On a scale of one to 10. how hard is it, in your
estimation , to do the following activities which are
involved in treating obesity (say what they are briefly):
a.) identifying when obesity is a significant health risk
factorforapatient? 12345678910
b.) educating a patient about obesity as a risk factor?
1 10
c.) engaging a patient in a treatment program to lose
weight? 1—--10
16.) Compared to other chronic health problems, how hard do
you think it is to treat obesity? more difficult, less diffi

cult, equally difficult? Why?
17.) Compared to other problems which can only be addressed
through diligent patient effort (smoking cessation , exercis-
ing regularly, diet control of high blood pressure or
diabetes, limiting or eliminating alcohol consumption), is
treating obesity more difficult, less diffcult, equally
difficult? Why?

18.) Compared to other problem areas you have to address
commonly in primary practice , how would you rate your own

124

confidence in treating obesity? more confident, than
many other areas; about as confident as other areas;
__ less confident , than in many other areas .

19.) How often, if ever, do you suggest patients try
community based support groups for weight loss?

What is you opinion of these groups? prompt: what
kinds of experiences have patients of yours had in these
groups?

15.) In term of your experience in your training, what
experiences did you have which taught you about treating
obesitfi
1.) Was there didactic material on obesity as a risk
factor?
2 .) didactic material on educating patients about
obesity?
3 .) didactic material about engaging patients in treating
obesity?
a.)during your pre-clinical years? 1.) Y N, 2.) Y N,
3 .) Y N
b.) during your clinical years? 1. Y N, 2.) Y N, 3.) Y
c.) as an intern? 1.) Y N, 2.) Y N, 3.) Y N
d.) a resident? 1.) Y N 2.) Y N 3.) Y N
e.) other experiences, formal or informal, during your
training? 1.) Y N, 2.) Y N, 3.) Y N
What were they?
16.) What was the nature of this experience?
a.) did you have the opportunity to see your supervisors and
mentors treat (diagnose , educate , engage in treatment) obese
patients? a.) during pre-clinical years? b .) clinical
years? c.) as an intern? d.) as a resident?
16b.) were you allowed and encouraged by supervisors to have
experience treating (diagnose, educate, engage in treatment)
obese patients?
c.) do you think that your training experiences were
adequate to help you be confident in this area of practice?
prompts: can you tell me a little more about that? What
was that like? , what had the greatest influence, do you
think, in your learning how to treatment obesity? if answers
are sparce .
1‘7 .) Has your confidence in your ability to treat obese
patients changed over time? How? To what do you attribute
the change?

 

References

Bandura, A. (1989). Human agency in social cognitive
theory. American Psychologist, 34, 1175-1184.

Bandura, A., O’Leary A., Taylor, C.B., Gauthier J., Gossard,
D. (1987). Perceived self-efficacy pain control: opoid
and nonopoid mechanisms. Journal of Personality and

Social Psychology, 5;, 563-571.

Bandura, A. (1986). Social foundations of thought and
action. Englewood Cliffs, NJ: Prentice-Hall.

Bandura, A. (in Press). Self-efficacy mechanism in
physiological activation and health-promoting
behavior. In J. Madden IV, S. Matthysse, & J.
Barchas (Eds.), Adaptation, learning and affect.

New York: Raven Press.

Bandura, A. (1982). Self-efficacy mechanism in human
agency. American Psychologist, 31, 122-147.
Bandura, A. (1977). Self—efficacy: Toward a unifying theory

of behavior change. Psychological Review, g3, 191-215.

Bandura, A., & Adams, N.E. (1977). Analysis of self-efficacy
theory and behavior change. Cognitive Therapy and
Research,;, 287-308.

Bartlett, E. E. (1989). How can patient education help
attain objectives for the nation? American Journal of
Preventive Medicine, 5, 230-239.

Becker M.H., & Janz, N.K. (1990). Practicing health
promotion: the doctor's dilemma. Annals of Internal
Medicine, 113, 419-422.

Betz, N.E. & Hackett,G. (1981). The relationships of
career-related self-efficacy expectations to percieved
career options in college women and men. Journal of

Counseling Psychology,;§, 399-410.

Bloom, S.W. (1989). The medical school as a social
organization: The sources of resistance to change.
Medical Education, g1, 228-241.

125

 

 

126

Bray, G.A. (1984). The role of weight control in health
promotion and disease prevention. In J.D. Matarazzo,
S.M. Weiss, J.A. Herd, N.E. Miller, & S.M. Weiss
(Eds.), Behavioral health: A handbook of health
enhancement and disease prevention (pp.632-656). New
York: John Wiley & Sons.

Brownell, K.D. (1982). Obesity: Understanding and treating
a serious, prevalent, and refractory disorder. Journal
of Counseling and Clinical Psychology, 5Q, 820-840.

Condotti, M.M. & Lichtenstein, E. (1981). Self-efficacy and
relapse in smoking cessation. Journal of Consulting
and Clinical Psychology, 49, 648-658.

Dembo, M.H. & Gibson, S. (1984). Teachers’ sense of
efficacy: An important factor in school

improvement. The Elementapy School Journal, 86,
173-184.

Craig, K.D. (1984). Emotional aspects of pain. In P.C. Wall
& R. Melzack (Eds.) Textbook of pain. London:
Churchill/Livingston.

DiClementi, C.C. (1981). Self-efficacy and smoking cessation

maintenance: A preliminary report. Cognitive Therapy
and Research, 5, 175-187.

Dillman, D.A. (1978). Mail and telehone surveys: The total
design method. New York: John Wiley & Sons.

Ewart, C.K., Taylor, C.B., Reese, L.B., Debusk, R. F.
(1984). Effects of early postmyocardial infarction
exercise testing on self-perception and subsequent

physical activity. American Journal of Cardiology,
1076-1080.

1,

Ell, K.O. (1986). Coping with serious illness: On
integrating constructs to enhance clinical research,
assessment and intervention. International Journal of

Psychiatry in Medicine, 15, 335-357.

Fordyce, W., Fowler, R., Lehmann, J., DeLateur, B., Sand,
P., & Treischmann, R. (1973). Operant conditonning in
the treatment of chronic pain. Archives of Physical
Medical Rehabilitation, pg, 399-408.

Freeman, S.H. (1987). Health promotion talk in family
practice encounters. Social Science and Medicine, 2;,
961-966.

127

Fuller, B., Wood, K., Rappoport, T., & Dornbusch, S.M.
(1982). The organizational context of individual
efficacy. Review of Educational Research, 55, 7-30.

Glaser B.G. & Strauss A.L. (1967). The discovery of
grounded theory: Strategies for qualitative research.
Chicago: Aldine Publishing.

Glass G. V. & Hopkins, K.D. (1984). Statistical methods in

education and psychology (2nd Ed.) New Jersey:
Prentiss-Hall, Inc.

Gorden, R.L. (1980). Interviewing: Strategies, technigpes,
and tactics. Illinois: The Dorsey Press.

Gottlieb, N.H., Mullen, P.D., & McAlister, A.L. (1987). Pa—
tients’ substance abuse-and the primary care physician.
Addictive Behaviors. 12, 23-32.

Gruppen, L.D., Wolf, R.M., & Stross, J.K. (1990). Physician
practice characteristics as a context for primary care
treatment decision making: A preliminary study.
Academic Medicine, 65L (September Suppl.), 89-810.

Heppner, P.P., Kivlighan Jr., D.M., & Wampold, B.E. (1992).

Research design in counseling. Pacific Grove:
Brooks/Cole Publishing.

Hoshmand, L.L.S. (1991). Clinical inquiry as scientific
training. The Counseling Psychologist, 19, 431-453.

Kaplan, R.M. (1990). Behavior as the central outcome in
health care. American Psychologist, 45, 1211-1220.

Kaplan, R.M., Atkins, C.J., & Reinsch, S. (1984). Specific
efficacy expectations mediate exercise compliance in
patients with COPD. Health Psychology, 55 223-242.

Lenker, S.-L., & Lorig, K. (1984). Reasons for the lack of
association between changes in health behavior and
improved health status: An exploratory study. gapippt

Education and Counseling, 5, 69-72.

Lewis, C.E., Clancy, C., Leake, B., & Schwartz, J.S. (1991).
The counseling practices of internists. Appgl§_pf
Internal Medicine, 114, 54-58.

Lorig, K., Chastian, R.L., Ung, B., Shoor, S., & Holman
H.R. (1989). Development and evaluation of a scale to
measure perceived self-efficacy in people with
arthritis. Arthritis and Rheumatology, 5;, 37-44.

-9‘

128

Lorig, K., Seleznick, M. Lubeck, D., Ung B., Chastian, R.L.,
& Holman, H.R. (1989). The beneficial-outcomes of the
arthritis management course are not adequately
explained by behavior change. Arthritis and
Rheumatologyy 52, 91-95.

Lorig, K., Lubeck D., Kraines, R.G., Seleznick, M., &
Holman, H.R. (1985). Outcomes of self-help education
for patients with arthritis. Arthritis and

Rheumatology, 55, 680-685.

Luborsky, L. & Auerbach, A.H. (1985). The therapuetic
relationship in psychodynamic psychotherapy: The
research evidence and its meaning for practice.

American Psychiatric Association Annual Review, 3.

Luborsky, L.,Crit-Christof, P., Alexander, L., Margolis, M.
& Cohen, M. (1983). Two helping alliance methods for
predicting outcomes of psychotherapy: A counting signs
vs. A global rating method. The Journal of Nervous and
Mental Disease, 51;, 480-491.

Luborsky, L., Singer, B. & Luborsky, L. (1975). Comparative
studies of psychotherapies: Is it true that "everyone
had won and all must have prizes?" Archives of General

Psychiatry, 52, 995-1008.

Lorig, K. & Laurin, J. (1985). Some notions about
assumptions underlying health education. Health

Education Quarterly 12, 331-343.

Maddux, J.E., Sherer, M. & Rogers, R.W. (1982).
Self-efficacy expectancy and outcome expectancy: Their
relationship and their effects on behavioral

intentions. Cognitive Therapy and Research, 5,
207-211.

Maddux, J.E., & Stanley, M.A. (1986). Self-efficacy theory
in contemporary psychology. Journal of Social and

Clinical Psychology,4, 249-255.

Maddux, J.E., & Rogers, R.W. (1983). Protection motivation
and self-efficacy: A revised theory of fear appraisals

and attitude change. Journal of Experimental Social

Psychology, 19, 469-479.
Maheux,B., Legault, C., & Lambert, J. (1989). Increasing

response rates in physicians’ mail surveys: An
experimentalstudy. American Journal of Public Health,
15, 638-639.

129

Marshall, C., & Rossman G.B.. (1989). Designing,gualitative
research. Newbury Park, CA:Sage.

Mechanic, D. (1974). Social structure and personal
adaptation: Some neglected dimensions. In Coelho
C.V., D.A. Hamburg & J.E.Adams (Eds.), Coping and
adaptation. New York: Basic Books.

Mehrens, W.A., & Lehmann, I.J. (1984). Measurement and
evaluation in education and psychology, 3rd. Ed. New
York: Holt, Rinehart and Winston.

Mintz, J. (1983). Integrating research evidence: A
commentary on meta-analysis. Journal of Consulting and
Clinical Psychology, 1&1 71-75.

Neufeld, R.W.J., & Thomas, P. (1977). Effects of percieved
efficacy of a prophylactic controlling meachanism on
self-control under painful situations. Canadian
Journal of Behavioural Science, 5, 224-232.

Newmann, F.M., Rutter, R.A., & Smith M.P. (1989).
Organizational factors that affect school sense of
efficacy, community, and expectations. Sociology of
Education, 55, 221-238. '

O’Leary, A. (1985). Self-efficacy and health. Behavior
Therapy and Research, 55, 437-451.

Nunnally, J.C. (1978). Psychometric Theory. New York:
McGraw- Hill

Orlandi, M.A. (1987). Promoting health and preventing
disease in health care settings: An analySis of
barriers. Preventive Medicine, 55, 119-130.

Raudenbush, S.W., Rowan B., & Cheong, Y.F. (in Press).
Contextual effects on the self-efficacy of high school

teachers. Sociology of Education.

Reid, D.W. (1984). Participatory control and the chronic
illness adjustment process. In B. Dohenwend & R.
Dohenwend (Eds.), Research with the Locus of Control
Construct (Vol.3) (pp.361-389).' Orlando:Academic

Press.

' ' 'tive

Reese, L.B. (1982). Pain reduction through cogni ' .
self-relaxation and placebo means: A self-efficacy d
analysis. Unpublished Doctoral Dissertation, Stan or ,

CA: Stanford University.

130

Rodin, J. (1986). Aging and health: effects of the sense of
control. Science, 233, 1271-1276.

Rodin J., & Janis, I. (1982). The social influence of
physicians and other health care practioners as agents

of change. In (Eds.). Interpersonal issues in health
care. (pp.33-49). Orlando: Academic Press.

Rosenthal, R. (1983). Assessing the statistical and social
importance of the effects of psychotherapy. Journal of
Consulting and Clinical, 5;, (1), 4-13.

Rubin, J.D., Sobal, J., & Moran, M.T. (1990). Health
promotion beliefs and practices of fourth year medical
students. American Journal of Preventive Medicine, 5,
106-111.

Schunk D.H., & Carbonari, J.P. (1984). Self-efficacy
models. In J.D. Matarazzo, S.M. Weiss, J.A. Herd, N.E.
Miller, & S.M. Weiss (Eds.), Behavioral health: A
handbook of health enhancement and disease prevention
(pp. 230-247). New York: John Wiley & Sons.

Shoor, S.M., & Holman (1985). Development of an instrument
to explore psychological mediators of outcome in
chronic arthritis patients. Trans.Association of
American Physicians. In Press.

 

Sobal, J., Valente, C.M., Muncie, Jr., H.L., Levine, D.M. &
DeForge, B.R. (1985). Physicians’ beliefs about the
importance of 25 health promoting behaviors. American
Journal of Public Health, 15, 1427-1428.

Spradley, J.P. (1979). The ethnographic interview. New
York: Holt, Rinehart and Winston.

Stiles, W.B., Shapiro, D.A., & Elliot, R. (1986). Are all
psychotherapies equivalent? American Psychologist,

él.(2).

Stone, G.L. (Ed.) (1991). Counseling psychology and health
applications. The Counseling Psychologist, 55 (3).

Strauss, A. & Corbin, J, (1990). Basics of qualitative
research: Grounded theory procedures and techniques.

Newbury Park, Ca: Sage.

Strupp, H.H. (1986). Psychotherapy research, practice, and
public policy (how to avoid dead ends). American

Psychologist, 51 (2).

 

 

131

Stunkard, A.J. (1975). From explanation to action in
psychosomatic medicine: The case of obesity.
Psychosomatic Medicine, 51, 195-236.

Sullivan, L. (1990). An evolving agenda for the AAMC’s
second century. Academic Medicine, 55, 226-230.

Taylor, C.B., Bandura, A., Ewart, C.K., Miller, N.H., &
DeBusk, R.F. (1985). Exercise testing to enhance
wives’ confidence in their husbands' cardiac capability
soon after clinically uncomplicated acute myocardial

infarction. American Journal of Cardiology, 55,
635-638.

Tresolini C.P. & Stritter, P.T. (1992). Medical students'
development of self-efficacy in conducting patient
education for health promotion: An analysis of learning
experiences. Presentation at the annual meeting of
AERA, San Francisco, CA.

Turk, D.C., Meichenbaum, D., & Genst, M. (1983). Pain and
behavioral medicine: A Cognitive-behavioral approach.
New York: Guildford Press

U.S. Preventive Services Task Force. (1989). Guide to
clinical preventive services: An assessment of 169
interventions. Baltimore: Williams & Wilkins.

Valente, C.M., Sobal, J., Muncie, H.L., Levine, D.M. &
Antlitz, A.M. (1986). Health promotion: Physicians'
beliefs, attitudes, and practices. American Journal of
Preventive Medicine, 5, 82-88.

 

Wechsler, H., Levine, S., Idelson, R.K., Rohman, M., &
Taylor, J.O. (1983). The physician’s role on health
promotion-~A survey of primary care practitioners. 155
New England Journal of Medicine, 308, 97-100.

Weinberg, R.S., Gould, D., & Jackson, A. (1979).
Expectations and performance: An empirical test of
Bandura’s self-efficacy theory. Journal of Sport

Psychology, 1, 320-331.

Weinberg, R.S., Hughes, H.H., Critelli, J.W., Englan, R., &
Jackson, A, (1984). Effects of pre-eXisting and
manipulated self-efficacy on weight loss in a
self-control program. Journal of Research on

Personality, 15, 352-358.

 

' ' ht reduction
Weinber , R.S., & Agras, W.S. (1984) The weig‘
efgicacy questionnaire. Stanford UniverSity, Stanford,

CA.

 

132

Weiner, H. (1989). A medicine of human relationships. The
Pharos, 55, 2-6.

Weisenberg, M., Kreindler, M.L., Schachat, R., & Werbof, J.
(1975). Pain: Anxiety and attitudes in black, white and
Puerto Rican patients. Psychosomatic Medicine, 51,
123-135.

Wilson, T.G. (1984). Weight control treatments. In J.D.
Matarazzo, S.M. Weiss, J.A., Herd, N.E. Miller & S.M.
Weiss (Eds.) Behavioral health: A handbook of health

enhancement and disease prevention. (pp. 657-670). New
York: John Wiley & Sons.

Yin, R.K. (1989). Case study research: Design and methods.
Newbury Park, CA: Sage.

"111111111311111111“