Perceptions of physician behavior

MISSING IMAGE

Material Information

Title:
Perceptions of physician behavior the effect of physician expressive and technical competence on psychological symptom disclosure
Physical Description:
x, 134 leaves : ill. ; 28 cm.
Language:
English
Creator:
Franco, Miguel A., 1961-
Publication Date:

Subjects

Subjects / Keywords:
Physician and patient   ( lcsh )
Interpersonal communication   ( lcsh )
Communication in medicine   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1991.
Bibliography:
Includes bibliographical references (leaves 121-133).
Statement of Responsibility:
by Miguel A. Franco.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001690061
notis - AJA2103
oclc - 25151701
System ID:
AA00003720:00001

Full Text








PERCEPTIONS OF PHYSICIAN BEHAVIOR:
THE EFFECT OF PHYSICIAN EXPRESSIVE AND TECHNICAL
COMPETENCE ON PSYCHOLOGICAL SYMPTOM DISCLOSURE
















By

MIGUEL A. FRANCO


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE
UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1991











ACKNOWLEDGEMENT


There are many people I wish to thank for their

assistance, guidance, and friendship while I have been in

the counseling psychology program at the University of

Florida. Shae Kosch could not have been a better

supervisor. Not only was Shae supportive during difficult

times, she was the one who introduced me to the "real" world

of psychotherapy. Shae's patience and attention to detail

are what I am most grateful for.

Harry Grater is another person who played a major role

in my professional development. Harry's abilities to teach,

and stimulate me to listen to my inner voice in therapy are

what I am most grateful for.

I want to thank the Minority Fellowship Program for

making it financially possible for me to make it through

graduate school at the University of Florida.

Special thanks go to Linda Hellmich. I am happy that

we are in the same field.

Internship year played a critical role in my overall

development. When I was accepted as an intern at Notre

Dame/Oaklawn I had very high expectations. Rita Donley

almost single handedly surpassed all of these expectations.





I want to thank my mother, Myriam, and my father,

Antonio "Chuchito" Franco. I could not have done it without

their support.

I want to thank my big brother, Tony. Every kid needs

a "big" brother. I am grateful that I have him as mine.

I want to thank the rest of my family, Luly, Tio,

Abuela Kela, Abuelo Tin, Abuela Angelina and the boss,

Antonio "Chucho" Franco Tauler.

I want to thank my children, Candice and Daniel, for

bearing with me, and inspiring me. They are proof that God

works overtime.

Finally, I want to thank the most important person in

my life, my wife B.J. I dedicate this work to her.


iii











TABLE OF CONTENTS

PAGE

ACKNOWLEDGEMENTS . ... .. ii

LIST OF FIGURES . ... vi

LIST OF TABLES . . vii

ABSTRACT . . ... vii

CHAPTERS

1. INTRODUCTION . . 1

2. REVIEW OF THE LITERATURE . 6

Psychological Disorders in Primary Care 6
Cost Benefit Analysis of Treating
Psychological Concerns . 8
Research on Physician-Patient Communication 11
Therapist-Client Relationship and its
Salience to the Physician-Patient
Relationship . ... 16
Psychological Research on Self-disclosure 21
Attributes of the Recipient of
Self-disclosure . ... .22
Characteristics of the Disclosing Person 26
Properties of the Disclosure Situation ... .28
Topic or Content of Self-disclosure 33
Purpose of the Study . ... .34

3. METHODS . . 43

The Videotaped Medical Interviews ... .45
Operational Definitions . .. .46
Subjects . . 47
Procedure . . .. 47
Variables. ............ 48
Dependent Variables . ... 53
Hypothesis Testing . ... .56





4. RESULTS . . .

Pretest of the Expressive & Technical
Competence Manipulation .
Demographics . .. .
Hypothesis 1 .. . .
Hypothesis 2 .. . .
Hypothesis 3 .. . .
Hypothesis 4 .. . .
General, Private, and Psychological Symptom
Disclosure . . .
Correlational Analysis . .

5. DISCUSSION . . .

Analysis of Hypotheses One and Four (ANOVA I)
Analysis of Hypothesis Two (ANOVA II) .
Analysis of Hypothesis Three . .
Concluding Remarks . .


APPENDIC

A.


;ES

SCRIPTS OF EXPRESSIVE SKILL/TECHNICAL SKILL
MANIPULATIONS . .


B. CONSENT FORM . ..

C. SYMPTOMS QUESTIONNAIRE . .

D. SMITH-FALVO PATIENT-DOCTOR INTERACTION SCALE .

E. PRE TEST QUESTIONNAIRE . ..

F. PRE TEST ANOVA SUMMARY TABLES .

REFERENCES. ....... .. . .

BIOGRAPHICAL SKETCH ... .


* 102

. 112

* 113

. 116

* 117

* 118

* 121

* 134











LIST OF FIGURES


Figure Page

4.1 ANOVA I Expressive Dimension Manipulation:
Physician Gender x Subject Gender
Interaction . . .. 65

4.2 ANOVA I Expressive Dimension Manipulation:
Physician Gender x Level of Expressive
Skill Interaction . .. 66

4.3 ANOVA II Technical Dimension Manipulation:
Physician Gender x Subject Gender
Interaction . ... 69

4.4 ANOVA IV High Expressive Dimension Condition:
Degree of Psychological Disclosure to
Male vs. Female Physician . ... .73











LIST OF TABLES


Tables Page

4.1 Population of Study . 62

4.2 Hypothesis One: Expressive Skill
Manipulation Summary Table . ... .63

4.3 Hypothesis One: Expressive Skill Manipulation
Test for Interaction Summary Table ... 63

4.4 Hypothesis Two: Technical Skill
Manipulation Summary Table . ... .68

4.5 Hypothesis Two: Technical Skill Manipulation
Test for Interaction Summary Table ...... 68

4.6 Hypothesis Four: Summary Table of High
Expressive Condition When Disclosing
Psychological Symptoms to a Male Versus
Female Physician . ... 72

4.7 Hypoithesis Four: Summary Table of Test of
Interaction Between Subject and Physician
Gender When Disclosing Psychological
Symptoms to a High Expressive Physician .. 74

4.8 Correlations Between Degree of Psychological
Symptom Disclosure and Degree of Importance
of Physician Behaviors . ... 76


vii











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

PERCEPTIONS OF PHYSICIAN BEHAVIOR:
THE EFFECT OF PHYSICIAN EXPRESSIVE AND TECHNICAL
COMPETENCE ON PSYCHOLOGICAL SYMPTOM DISCLOSURE

By

MIGUEL A. FRANCO

Chairman: Franz Epting, Ph.D.
Major Department: Psychology

Physician underdiagnosis of psychological symptoms is

well documented. This study investigates whether

perceptions of specific physician behaviors exert an effect

on willingness to disclose psychological symptoms.

Forty male and forty female subjects, imagining a

future visit of their own to two physicians viewed on

videotape, rated their willingness to disclose symptoms of a

general, private, and psychological nature to the male or

the female physician. The trigger videotapes were of a male

and female physician, in their mid-thirties, white, and

physically pleasant in appearance. The taped behaviors were

pretested as being high, moderate, or low in expressive

competence and technical competence. One analysis focused

on the effect of high versus low technical skill on

psychological symptom disclosure. In this experimental


viii





condition, the physician's level of expressive skill was

controlled at a moderate level. A second analysis focused

on the effect of high versus low expressive competence on

psychological symptom disclosure. In this experimental

condition, the physician's level of technical skill was

controlled at a moderate level.

Results indicated that physicians depicting high

technical competence elicited greater willingness to

disclose psychological symptoms from subjects than did

physicians depicting low technical competence. Results also

indicated significantly greater willingness on the part of

male subjects to disclose psychological symptoms when they

viewed the male physician.

When level of expressive competence was manipulated,

increased willingness to disclose psychological symptoms was

found when physicians demonstrated high expressive

competence, and in male physician-male subject dyads. The

highest level of willingness to disclose psychological

symptoms occurred when a male physician was depicted as

having high expressive skill.

Evidence was found suggesting that subjects may

have espoused sex role stereotypes of male versus female

physician behavior. Results indicated that, when the male

physician behaved contrary to expectations for sex role

stereotypes, he elicited greater willingness to disclose

psychological symptoms from subjects than did a female

physician who behaved comparably.





Implications are drawn concerning the relative

influence of technical versus expressive competence on

psychological symptom disclosure, the physician-patient

relationship, and medical service delivery.











CHAPTER 1
INTRODUCTION


A cyclic change in the pattern of medical practice has

occurred gradually during the twentieth century in the

United States. In the earliest pattern, physicians followed

a model of practice that was characterized by continuous,

longitudinal, comprehensive care offered by a physician to

each member of a family. The physician attended to all

types of medical problems that patients presented and were

very familiar with many aspects of patient's lives, seeing

them often in the community or even during home visits.

However, in the early part of the century, medical

school training began to emphasize greater specialization

and the proportion of physicians calling themselves "general

practitioners" declined (Chisolm, 1978). A pattern

developed in which patients would be referred from a general

practitioner to a specialist, depending upon the particular

ailment with which the patient presented, or the tentative

diagnosis made by the primary care physician. Thus the

specialized physicians often did not know their patients

well, and the emphasis of the medical encounter was on

treating one particular disease by a physician who was an

expert in that disease state. Specialization became the

"ideal" in medical training although many providers







continued to practice as generalists, especially in rural

areas. However, after World War II, more and more

physicians were trained and practiced as specialists and

subspecialists.

During the 1960s, criticisms of this pattern were

voiced, citing problems of overspecialization, exorbitant

costs, and lack of sensitivity toward the patient. A

movement emerged to restore the status of the general

practitioner and to encourage physicians to adopt this style

of practice (Howe, Tapp, & Jackson, 1982). This led to the

advent of the specialty of "family practice" which mandated

that the physician return to a model of practice that

focused on the patient's social and psychological experience

as well as on the biological aspects of medical care.

The most articulate proponent of a biopsychosocial

model of health and disease has been George Engel, a

physician dually trained in the specialties of Internal

Medicine and Psychiatry (Engel, 1977). Engel has contended

that the dominant conceptual model of Western medicine has

its roots in molecular biology and asserts that every

disease can ultimately be explained in terms of deviations

from the norms of measurable biological processes.

According to Engel, this orientation fails to consider the

potentially stabilizing and destabilizing influence of

psychosocial factors, which are instrumental in the

development, expression, prognosis and outcome of a

patient's disease state. The tendency to ignore







psychosocial domains in medical practice was empirically

validated by Silverman, Gartrell, Aaronson, Steer and Ebdvil

(1983). These researchers explored the effects of a four

month long course entitled, "Introduction to Clinical

Medicine" at the Harvard Medical School. In their study, 66

percent of all medical students failed to ask a single

psychological or social question in response to videotaped

simulated patient interviews; conversely, 90 percent of the

students requested additional information regarding

biological functioning. The results of Silverman and

colleagues supported Engel's contention that "the biomedical

model borders on being a culturally specific perspective

about disease our folk model" (p. 130).

In recent years, residency training programs in primary

care specialties, particularly programs in family practice,

have emphasized the integration of psychological, social,

and biological variables in diagnosis and treatment. These

programs are required by the national accrediting

committees, the Residency Review Committee and the American

Board of Family Practice, to maintain faculty members who

specialize in "behavioral science" and to devote a certain

percentage of the curriculum to "behavioral science"

training in particular topics. Residency programs employ

psychologists, social workers, psychiatrists or other social

science specialists to train residents in the psychological

and interpersonal aspects of counseling, interviewing, and

family dynamics (Society of Teachers of Family Medicine Task







Force on Behavioral Science, 1985). Psychologists are also

involved in the development of behavioral science curricula

in residency programs (Johnson, Fisher, Guy, Keith, Keller &

Sherer, 1977).

One goal of the residency programs which emphasize the

biopsychosocial aspects of care is to enhance physicians'

ability to diagnose and treat the psychological concerns of

their patients. Kosch and Dallman (1983) contend that

despite the efforts of these residency programs to develop

methods of teaching behavioral science and evaluating their

impact, many primary care physicians still show major

deficits in handling psychosocial aspects of patient care.

Their claim is supported by a study which showed that

physicians often did not detect psychological illnesses when

they were present (Jones, Badger, Ficken, Leeper & Anderson,

1987). In this study, the diagnoses of mental disorders

among 51 patients that were made by a group of 20 family

practice physicians were compared with diagnoses generated

by the Diagnostic Interview Schedule. The results showed

that 75 of 94 DIS diagnoses (79%) were not detected by the

physicians during regular medical visits. In a similar

study of 20 third year family practice residents, Ficken,

Milo, Badger, Leeper, Anderson and Jones (1986) reported

that when DIS criteria was applied to patients in a family

practice clinic, results indicated that diagnoses of mental

disorder were warranted for 44 patients yet residents

identified 51 diagnoses among 35 patients. Also, residents







were in disagreement with the DIS results or made no

diagnoses involving 41 patients.

Schulberg and Burns (1987) report that physicians

routinely underdiagnose mental disorders despite the clear

evidence that these illnesses are present in approximately

25% of primary care patients. These investigators suggested

that future research focus on explanations for these

findings instead of completing additional studies which

demonstrate that psychiatric morbidity occurs more

frequently than it is detected.

Although there has been considerable research on the

underdiagnosis of psychological symptoms and/or disorders,

most of it has focused primarily on one aspect of the

physician-patient dyad, specifically, the physician's

cognitive and behavioral problem solving style. And,

although the medical literature is replete with articles

regarding the effect of patient satisfaction on compliance

with physician's treatment plans, very little information

has been collected on the patient's role in facilitating

diagnosis and treatment of psychological disorders. One

important aspect, the patient's willingness to disclose

psychological fears, concerns or symptoms to physicians has

been virtually ignored.

The present study is intended to ascertain if subjects'

perceptions of specific physician behaviors exert an effect

on their willingness to disclose psychologically-laden

symptoms and/or fears.











CHAPTER 2
REVIEW OF THE LITERATURE


Psychological Disorders in Primary Care

For patients to receive effective and appropriate

health care services, physicians must adequately diagnose

the basis for their presenting complaints and tailor their

treatment plans accordingly. Although many patients present

to their physician with what appears to be a biologically-

based illness, there are frequently concomitant

psychological aspects to their complaints which are not

disclosed to the physician. Often physicians proceed to

treat physical ailments without inquiring about psychosocial

factors which may be the catalyst or cause of the presenting

problem. There is evidence that a substantial percentage of

patients treated by primary care physicians experience

significant psychosocial problems, with estimates ranging

from 10% to 60% (Brantley, Veitia, Callon, Buss, Sias,

1986). There is also evidence that as many as 50% to 75% of

primary care patients seek medical assistance as a

consequence of psychological distress (Stoeckle, Zola &

Davidson, 1964). In addition, research has suggested that

primary care patients experiencing psychosocial distress

utilize medical facilities with greater frequency than

patients without such distress (Tessler, Mechanic & Dimond,







1976). Stumbo, Good and Good (1982) analyzed data from a

random sample of all active patients at their clinic. They

found that 36% of all the adult females and 26% of all the

adult males warranted a psychosocial diagnosis and that

these patients made more medical visits over a twelve month

period than did any other diagnostic group. Further

evidence of overutilization in a small percentage of primary

care patients is provided by Browne, Pallister, and Crook

(1982). Browne and his colleagues found that 4.5% of

patients in a family practice clinic could be described as

frequent attenders (nine or more visits per year), and this

small group of patients generated 21% of the total visits in

a twelve month period. These patients were found to be

similar to each other in that they exhibited elevated levels

of physical, social, and emotional distress. The

researchers concluded that medical utilization rates for

people with emotional disorders are one and one-half to two

times greater than for patients without such problems.

These researchers also claimed that overutilization would be

substantially reduced by diagnosing and treating

psychological disorders at the onset of utilization of

medical services.

The results of another project underlined the

importance of physician attention to the psychosocial

distress of their patients. Ruberman, Weinblatt, Goldberg

and Chaudhary (1984) discovered that psychosocial distress

factors were associated with mortality in convalescing heart







attack patients. These researchers reported that of the

2320 male survivors of heart attacks participating in their

study, patients who described themselves as being socially

isolated and having a high degree of life stress had more

than four times the risk of death than the patients who

reported low levels of both stress and isolation. Ruberman

and his colleagues concluded that physicians should take

special steps to encourage patients who have suffered heart

attacks to discuss their psychosocial concerns with them.


Cost-Benefit Analysis of Treating Psychological Concerns

Recognizing the high rate of psychosocial problems

experienced by primary care patients and the potential

overutilization of medical services, the medical community

has endeavored to meet the needs of these patients. A

number of researchers have attempted to ascertain if there

are any effects of brief psychological intervention by

mental health professionals on the frequency of medical

visits by patients experiencing psychosocial distress.

These researchers have estimated that brief psychological

intervention leads to decreases in the use of medical

services of 31% to 64% in total number of visits per year

(Goldberg, Krantz & Locke, 1970; Cummings & Follette, 1976;

Longsbardi, 1981).

More recently, Brantley, Veitia, Callon, Buss and Sias

(1986) conducted a study to determine if patients receiving

psychological intervention in a family practice clinic were







more likely than control patients to decrease medical

utilization. These researchers reported that their results

were consistent with earlier work documenting an association

between brief psychosocial intervention and decreases in

frequency of clinic visits. Moreover, these researchers

found that pre-internship level clinical psychology trainees

were as effective in providing services as mental health

professionals who had completed their training. Therefore,

Brantley and his colleagues advocated for the utilization of

psychology trainees in family practice clinics as a cost-

effective method of providing psychological services which

could in turn lead to decreases in medical services and

expenses.

A number of meta-analytic studies have been conducted

to ascertain the cost-benefit effects of outpatient mental

health treatment. Mumford, Schlesinger & Glass (1982)

conducted a meta analysis of controlled studies of the

effect of psychological intervention on patients following

heart attacks or facing surgery. Their results indicated

that patients provided with psychological intervention did

better than control patients on most outcome indicators.

Thirteen of these studies included days of hospitalization

as an outcome indicator, and combined results disclosed that

psychologically treated patients were released approximately

48 hours sooner than were control patients.

Devine and Cook (1983) conducted a meta-analysis of

49 controlled experiments on the effect of psychoeducational







interventions with surgical patients and found that patients

receiving mental health services were discharged

approximately 34 hours earlier than patients provided only

the usual medical treatment. One of the most comprehensive

studies conducted in this area involved the claims filed for

the Federal Employees Blue Cross and Blue Shield plan for

the years 1977-1978 (Mumford, Schlesinger, Glass, Patrick &

Cuerdon, 1984). These researchers concluded that the

reductions in use of medical services as a result of

psychological intervention are most highly associated with

inpatient rather than outpatient utilization of medical

services. Mumford and her colleagues emphasized the

importance of these findings since inpatient charges account

for 73% of total medical expenses, hence, a reduction in

inpatient services would account for a substantial savings

in medical personnel time and patient money. These

researchers also determined that older patients showed

larger cost-benefit offset effects than younger patients,

and they suggested that planned psychological intervention

among the elderly is warranted given the special needs of

this group.

The benefits that stem from addressing psychosocial

concerns are heavily dependent on the communication that

ensues between the physician and patient during medical

interviews. Missed diagnoses and needless expenditures due

to inaccurate treatment regimens can occur as the result of

the physician's inability to elicit the "hidden agenda" from







the patient. For these reasons, family practice educators

have endeavored to teach effective communication skills to

medical students and residents, and a high priority in

family medicine research has involved studying and analyzing

communication patterns between physicians and patients

during medical interviews.


Research on Physician-Patient Communication

Major theoretical premises regarding the role of

communication between physicians and patients were described

by Szasz and Hollender (1956). These authors conceptualized

three types of physician-patient relationships which are

characterized by their modes of communication during medical

encounters. Their first model was the active-passive model:

communication is unilateral, the physician is active and the

patient is passive. The authors proposed that this

authoritarian model is most appropriate when patients are

incapacitated, as when they are anaesthetized, traumatized,

or comatose. The second model was the guidance-cooperative

model. Here the physician guides and the patient follows

instructions. The majority of physician-patient encounters

follow this model today as it is most applicable to cases

where patients present with acute problems (Mazzucca &

Weinberg, 1986). Szasz and Hollender claimed that the

communication which ensues in the guidance-cooperative model

is similar to that of the active-passive model in that the

information provided by the physician carries substantially







more weight than the information provided by the patient.

However, in the final model, the mutual participation model,

communication is markedly different. Here the physician and

the patient are interdependent problem-solving partners who

communicate freely with one another and consult with each

other regarding the patient's physical and/or emotional

condition. Szasz and Hollender stated that, due primarily

to the communication style differences, both physicians and

patients report a greater level of satisfaction with the

mutual participation model than with the other two models.

The communication that transpires during the medical

encounter has been a focus of study for researchers who have

analyzed the physician-patient relationship. Researchers

have endeavored to lend empirical support to the principles

originated by Szasz and Hollender, particularly with regard

to the efficacy of the mutual participation model. Although

results have not been uniformly in favor of the mutual

participation model (Davis, 1968; Davis, 1971; Mazzucca &

Weinberg, 1986), a substantially larger number of studies

have documented its efficacy in terms of high ratings in

reported patient satisfaction and subsequent treatment

compliance (Korsch, Gozzi & Francis, 1968; Francis, Korsch

& Morris, 1969; Svarstad, 1976; Bertakis, 1977; Stiles,

Putnam, Wolf, & James, 1979; Stiles & White, 1981, Stewart,

1984).

A number of other research findings have emphasized the

importance of training physicians in communication skills.







Several investigators have concluded that the way in which a

physician communicates to, and facilitates communication

from their patient is the most influential factor

controlling the level of reciprocal communication which

occurs between them during any routine visit (Svarstad,

1974; Hirsh, 1986).

Effective communication about illness and treatment

is a major component in a physician's technical skill

repertoire. This skill has been documented as being

instrumental in eliciting patient satisfaction and

compliance to treatment regimen (Hays & Dimatteo, 1984).

Also, refined physician communication skills are essential

in making accurate diagnoses. Numerous articles have

illustrated cases in which both diagnosis and treatment plan

have been incorrectly assigned due to a lack of clear

communication between physicians and patients. Cousins

(1984) reports an incident in which a cardiologist's lack of

knowledge of a patient's afternoon sauna led to an

inaccurate diagnosis and treatment regarding a cardiac

arrythmia. Hilfiker (1984) reported how a lack of

communication between a patient and himself contributed to

an inaccurate diagnosis and subsequent inappropriate

treatment with serious consequences.

Adequate communication skills are also essential as the

physician endeavors to obtain information about the

presenting complaint from the patient, as well as other

important historical information. Information gathering is







profoundly influenced by the role adopted by the physician.

To assist the patient in disclosing pertinent information

and history, the physician must be able to relieve patient

discomfort and anxiety about self-disclosure, give the

patient enough time to discuss pertinent issues, and appear

unhurried and empathic (Hirsh, 1986).

There is evidence documenting that patient adherence to

treatment regimens will increase as the degree of patient

activity increases, both in problem definition (via

disclosure of pertinent information), and in the designing

of the treatment plan (Eisenthal & Lazare, 1977; Eisenthal,

1979; Johnson & Leventhal, 1974; Johnson, Kirchoff &

Endress, 1975). The literature further suggests that the

ability to facilitate active patient involvement relies

heavily upon effective verbal communication skills in the

physician.

Research has also been conducted on physician nonverbal

communication skills. Researchers in this area consider

these nonverbal skills to be as essential to patient care as

verbal skills (Dimatteo & DiNicola, 1982; Friedman, 1982).

Two major dimensions of physician nonverbal skills have been

studied. The first dimension has been coined "decoding" or

"sensitivity"; the other, "encoding" or "expressiveness."

Nonverbal decoding has been operationally defined as the

capacity to understand the emotions conveyed through others'

nonverbal cues such as their facial expressions, body

movements, and voice tone. Nonverbal encoding has been







defined as the capacity to express emotion through nonverbal

cues. Dimatteo, Taranta, Friedman and Prince (1980)

initially reported that physician nonverbal skill was

related to patient satisfaction with their medical care.

Dimatteo, Hays, and Prince (1986) later reported that

physician ability to decode patient affect via their voice

tone was significantly related to appointment compliance,

and nonverbal encoding skill was significantly related to

physician workload (i.e., appointment scheduling). These

authors suggest that the nonverbal communication which

occurs during the medical interview can exert a strong

effect on patient behavior, including their verbal

disclosure of pertinent information.

In summary, research on physician-patient communication

has yielded two important conclusions. First, there is a

positive correlation between the amount of medical

information provided by the physician and the level of

satisfaction experienced by the patient with their medical

care. Therefore, a physician's level of skill in

communicating medical information may have an influence on

the satisfaction patients report regarding physician-patient

interactions. Second, there is a nonverbal emotional

dimension to the physician-patient interaction which may

influence patient satisfaction and behavior. This emotional

dimension has been further investigated in psychotherapy

literature.







Therapist-Client Relationship and its Salience to the
Physician-Patient Relationship

Researchers in the field of psychotherapy, more than

any other health field, have conducted research on the

provider-consumer relationship. Studies on the therapist-

client relationship became popular during the 1960s as a

result of Carl Rogers' (1957) proclamations that empathy and

its related concepts of warmth and genuineness were critical

factors in the therapeutic relationship. Accordingly, these

studies were aimed at determining whether these factors were

necessary and sufficient conditions for therapeutic

personality change.

In a review of the research conducted on this subject

during the 1960s, Truax and Mitchell concluded that:

Therapists and counselors who are accurately
empathic, non-possessively warm in attitude and
genuine are indeed effective. Also, these
findings seem to hold with a wide variety of
therapists and counselors regardless of their
training or theoretic orientation, and with a wide
variety of clients or patients, including college
underachievers, juvenile delinquents, hospitalized
schizophrenics, college counselees, mild to severe
outpatient neurotics, and a mixed variety of
hospitalized patients. Further, the evidence
suggests that these findings hold in a wide
variety of therapeutic contexts and in both
individual and group psychotherapy and counseling.
(p. 310, 1971)

The latter findings, however, have not gone

uncontested. Parloff, Waskow and Wolfe (1978) stated that

the unqualified claim that high levels of accurate empathy,

warmth, and genuineness represent the necessary and

sufficient conditions for effective therapy have not been







sufficiently supported. These researchers pointed to a

number of important issues which need to be considered when

therapist variables are being analyzed. One of the most

salient issues concerns whose perceptions of the therapist

behaviors are regarded as accurate. Parloff and his

colleagues claimed that, in a large number of studies

involving observer ratings of therapist variables, results

indicated no relationship between the interpersonal skills

of empathy, warmth and genuineness and positive outcome.

Orlinsky and Howard (1978) reviewed only studies that

measured client perceptions of therapist behaviors. They

reported variations in the findings of studies of client

perception of empathy; however, the majority of the studies

confirmed that the sense of being understood by one's

therapist is a consistent feature of beneficial therapy.

Orlinsky and Howard also reported that 13 studies clearly

indicated that the client's perception that the therapist

affirmed the client's value was significantly associated

with good therapeutic outcome.

Bergin and Lambert (1978) reached the following

conclusions: (a) the therapeutic relationship as defined by

the presence of core conditions as perceived by the client

is correlated with positive outcomes, (b) the absence of

these core conditions lead to detriments in the client, and

(c) the strongest correlations between therapeutic outcome

and the quality of the therapeutic relationship correspond

more so with client perceptions than objective observer







ratings. Lambert, Shapiro and Bergin (1986) added the

factors of client's perception of the therapist's

trustworthiness and acceptance to the list of important

factors leading to positive change.

Research conducted on the physician-patient

relationship has been generated from the specialties in the

medical community that emphasize psychosocial factors in

primary care, i.e., Family Practice and Psychiatry. There

is evidence that a number of factors critical to the

therapist-client relationship are also salient to the

physician-patient relationship, namely, the patient's

perception of the physician's level of empathy and warmth

may influence the therapeutic relationship. Studies of

patient satisfaction have detected that the most frequently

encountered criticism by patients of their health care

concerned the quality of their relationship with the

physician and their perception that their physician

expressed a lack of interest in them (Koos, 1955; Friedson,

1961). Numerous studies have found that patient behavior

has been negatively affected as a result of patient

dissatisfaction due to their belief that their physician did

not care about them (Falvo, Woehlke & Deichman, 1980).

The rapport existing between physicians and patients

has been operationally defined by Kaufman (1970) and Headlee

(1973) in terms of interpersonal etiquette, with respect and

compassion expressed by the physician as the primary

factors. Strauss (1968) reported that perceptions of







impersonal care, lack of continuity of care, or being

treated like an inanimate object are consistently associated

with low levels of patient satisfaction with health care.

Caplan and Sussman (1966) stated that ease of communication

between the physician and patient was associated with

patient's high ratings of their relationship. Eisenberg

(1977) contended that for patients to be satisfied in the

relationship with their physician they expect sympathetic

attention, clear communication about their condition and

treatment, as well as expressions of concern for them as

persons.

Despite the parallels existing between the therapist-

client and physician-patient relationships, patients seek

more than interpersonal care from their physicians.

According to Doyle and Ware (1977), physician conduct is the

most significant determinant of general satisfaction with

medical care. They define physician conduct as the

behavioral enactment of technical and interpersonal skills.

In addition to past research indicating that patients were

more concerned with the interpersonal aspects of health

care, recent evidence documents patients' concerns regarding

the technical skills of their physicians. Studies regarding

the therapist-client relationship do not indicate that

clients experience concern regarding the therapist's

technical skill; however, an argument can be made that this

is so because it is difficult for clients to distinguish

between a therapist's technical skills and their







interpersonal skills. Hence, the disparity which exists

between psychotherapy clients and medical patients on this

issue can be best explained in terms of the different

expectations that these two types of health care consumers

have of these different types of health care providers.

Falvo and Smith (1983) reported that patients do not

tend to discriminate between the technical and interpersonal

aspects of medical care provided by physicians. According

to these researchers, patients perceive three skills as

integral parts of health care: (1) technical skill, (2) the

ability to exchange information, and (3) interpersonal

skill. From their sample of 115 patients they also found

that patients considered it important that their physician

behave professionally and respectfully toward them and they

looked favorably upon physician behaviors that made them

feel as if they could talk about any problem.

It is apparent that interpersonal skill, with its

concomitant factors of empathy, warmth, and genuineness are

important ingredients of a medical interview that lead to

patients' satisfaction with their physician. Such a

conclusion is suggested by the large number of studies in

Family Medicine which concluded that the model of patient-

centered interaction is successful in eliciting satisfaction

from patients. Stewart (1984) proposed that patient-

centered interactions are those in which the patient's point

of view is actively sought by the physician. From this it

can be inferred that patient-centered physicians behave in a







manner that facilitates patient self-disclosure, and that

these physician behaviors include the ability to communicate

empathy, warmth and genuineness.

In order to facilitate disclosure from patients, a

large number of variables must be taken into account.

Again, the field of psychology has uncovered a large number

of important variables which exert an effect on self-

disclosure. This body of knowledge is applicable to the

practice of medicine in identifying factors which would

facilitate disclosure from patients.


Psychological Research on Self-disclosure

Within the past 30 years, research in the field of

psychology on self-disclosure has been extensive. Recently,

physicians have begun to examine this construct because the

information disclosed by patients is considered to be

instrumental in their endeavor to accurately diagnose and

adequately formulate treatment plans.

The most salient variables which researchers have

consistently found to affect patient self-disclosure

include: (1) the attributes of the recipient or target of

the disclosure, (2) the characteristics of the disclosing

person, (3) the properties of the disclosure situation, and

(4) the topic or content of the disclosure. These variables

received extensive coverage in literature reviews on self-

disclosure by Goodstein and Reineker (1974) and Berg and

Derlega (1987).







Attributes of the Recipient of Self-disclosure

Goodstein and Reineker (1974) reported that the degree

to which people self-disclose depends on the relationship

between the recipient and the discloser. Evidence suggests

that people are likely to self-disclose both to those with

whom they are intimate (Jourard and Landsman, 1960) and to

total strangers whom they never expect to see again

(Rickers-Ovsiankina and Kusmin, 1958). Apparently, people

self-disclose to others who have already demonstrated that

they will not punish the discloser or who have no capacity

for punishing the person for self-disclosing.

The degree to which a person likes the recipient also

affects the likelihood of self-disclosure (Certner, 1973).

Pederson and Higbee (1969) reported that disclosers

attribute positive qualities such as warmth, closeness, and

friendliness to their targets of self-disclosure. In their

review, Goodstein and Reineker concluded that "liking" is

the most important characteristic leading to the selection

of another as the recipient of a self-disclosure.

A large number of psychotherapy studies have supported

Goodstein and Reineker's conclusion. These studies imply

that client perception of the psychotherapist's likability

has an effect on the outcome of therapy. For example, Lorr

(1965) conducted a factor analysis of therapist variables in

which over 500 clients participated. The only two factors

which correlated with positive outcome of therapy were

patient perception that the psychotherapist accepted and







understood them. In an uncontrolled study, Lazarus (1971)

asked over 100 of his former clients to report those factors

which they believed were most instrumental to their positive

change. These clients reported that therapist gentleness,

honesty, and sensitivity were most instrumental to their

positive change. The dependent variable in both of these

studies was positive outcome and not the degree of client

self-disclosure. However, research findings do indicate

that successful psychotherapy cases differ from unsuccessful

cases depending on the extent to which clients/patients

engage in self-exploration (i.e., provide cognitive or

affective verbal responses about their experiences in

therapy) (Truax & Carkhuff, 1967). There is also evidence

that the level of the facilitative conditions such as

warmth, empathy and respect provided by the therapist is

significantly related to the depth of patient self-

exploration (Truax & Carkhuff, 1964).

One experimental study provided direct evidence of the

causal relationship between therapist level of empathy and

warmth and the degree of patient self-exploration (Truax &

Carkhuff, 1965). These experimenters initially established

a baseline of patient self-exploration during the first

twenty minutes of an interview in which relatively high

conditions of empathy and warmth were present.

Subsequently, the experimental variable of lowered

conditions was introduced and maintained for the next twenty

minutes. The experiment was conducted on three







schizophrenic patients who had recently been admitted to an

acute treatment ward. Results indicated that the

experimental lowering of conditions during the middle

portion of the interviews did indeed lower the amount of

self-exploration established at baseline. Truax and

Carkhuff (1967) concluded, "Where there is an experimentally

induced drop of accurate empathy and warmth, a clear effect

is produced in the form of a consequent drop in the level of

patient self-exploration" (p. 106).

Given the results of these studies, particularly Lorr

(1965) and Lazarus (1971), one can infer that therapists who

express warmth, respect, and genuineness are more likely to

be experienced by clients as being more likeable than

therapists who do not behave in such manner. In turn,

therapists demonstrating these behaviors facilitate greater

depth of client self-exploration (which implies a greater

amount of client self-disclosure). Halpern (1977) studied

whether therapist facilitative behaviors were related to

client self-disclosure. Halpern found that clients who

perceived their therapists as being warm and empathic

reported greater willingness to self-disclose than did

clients who did not perceive their therapist as warm and

empathic.

Another variable which has been shown to enhance the

level of therapist attractiveness/likability is therapist

self-disclosure (Halpern, 1977). Therapist self-disclosure

has been differentiated in terms of self involving versus







self-disclosing communication (McCarthy & Betz, 1978).

These researchers define self-disclosing statements as those

referring to factual information about the counselor. Self-

involving statements are those that reveal the counselor's

cognitions and emotions as they relate to the client.

Evidence exists indicating that self-involving statements

elicit counseling relevant information from clients

(McCarthy, 1979, 1982), and that it enhances client

attraction to the counselor (Reynolds & Fischer, 1983;

VandeCreek & Angstadt, 1985; Peka-Baker & Friedlander,

1987). There is a general consensus that client perception

of therapist attractiveness and likability have a

significant impact on the outcome of therapy. It is also

postulated that during the process of therapy, therapist

attractiveness and likability has an effect on client self-

disclosure. It is important to note, however, that research

on the effect of therapist self-disclosure on the degree of

the client liking of the therapist has not yielded uniform

results (Derlega, Lovell, & Chaikin, 1976; Curtis, 1981).

The gender of the recipient is another variable

regarding self-disclosure which appears to have an effect on

willingness to disclose. There is evidence that same-sexed

friends are more frequent recipients of self-disclosures

than opposite-sexed friends (Jourard, 1964, 1971). For many

years it was believed that same-sexed dyads were more

facilitative in initiating self-disclosure than opposite-

sexed dyads. However, Hill and Stull (1987) claimed that







female-female disclosure is highest in initiating self-

disclosures, opposite sex disclosure is next highest, and

male-male disclosure is lowest. Psychotherapy literature

lends support to Hill and Stull's claims. Evidence exists

that female psychotherapists are superior to male

psychotherapists in their ability to facilitate self-

disclosure from both male and female clients (Fuller, 1963;

Brooks, 1974).

In summary, it appears that liking is the

characteristic most responsible for the selection of another

person as the recipient of a self-disclosure. Liking

another person is a function of previous acquaintanceship,

perceived warmth, and perceived friendliness (Berg &

Derlega, 1987). This factor interacts with the gender of

the target person to influence the degree to which another

individual self-discloses.


Characteristics of the Disclosing Person

Research on the characteristics of the disclosing

person is extensive. The demographic variables affecting

patterns of self-disclosure include gender, race,

nationality, age, marital status, social class, and birth

order. Goodstein and Reineker concluded that patterns of

self-disclosures are complexly determined as many

demographic variables interact with one another and with

other factors as well. A discussion of these interactions

is beyond the scope of this paper and the interested reader







is encouraged to read Goodstein and Reineker (1974) for an

extended review of this topic.

The variable which has received the most extensive

attention is the gender of the discloser. The relationship

between gender and self-disclosure has been a topic for

which clear predictions have been made yet puzzling results

have emerged. The earliest research on self-disclosure

found that men revealed less about themselves than did women

(Jourard and Lasakow, 1958). However, subsequent studies

did not always find sex differences (Rickers-Ovsiankina &

Kusmin, 1958; Melikian, 1962; Plog, 1965), and other studies

reported greater disclosure by men under particular

circumstances, such as during the beginning stages in the

development of opposite sex relationships (Derlega,

Winstead, Wong & Hunter, 1985).

Research conducted within the context of the health

care provider and consumer relationship has not provided

definitive answers to the question of whether or not gender

differences exist. Brooks (1974) reported that within the

context of the therapist-client relationship there is no

difference between males and females with regard to self-

disclosure of emotional information. Conversely, in an

analogue study, Young (1979) found that male and female

subjects reported a greater willingness to disclose symptoms

to a physician of their respective sex. However, Young also

found that males reported more symptoms than did females.

In another analogue study, Young (1980) stated that male







subjects reported a greater willingness to disclose all

types of symptoms than females and that this difference was

due to a pattern of resistance of females to disclose to

female physicians.

In attempts to understand these inconsistent findings,

researchers have employed strategies that explore mediating

factors. Hill and Stull (1987) claim that interactions

between situational factors, sex role attitudes, sex role

identity, sex-role norms, and self-disclosure measures have

each exerted an effect and have confounded results. Hill

and Stull have encouraged future researchers to control as

many external factors as possible in order to provide more

definitive findings.


Properties of the Disclosure Situation

The situation in which an individual chooses to self-

disclose plays an important role in subsequent disclosure.

Psychological and physical factors, such as the privacy of

the setting and the physical and psychological distance

between the discloser and the recipient, have been found to

have an effect on self-disclosure. However, it appears that

the most salient situational factor involves the mutuality

or reciprocity of self-disclosure. Chittick and Himelstein

(1967) found that subjects who were placed with two highly

self-disclosing confederates were themselves more self-

disclosing in their self-introductions than subjects placed

with low self-revealing confederates. Jourard and Resnick







(1970) reported that pairs of subjects with self-reported

high levels of disclosure did self-disclose more than pairs

of subjects with low self-reported levels of self-

disclosure. When a high self-discloser was paired with a

low disclosing subject, however, the low disclosing subject

increased his or her level of disclosure. Hence, there is

evidence to conclude that low disclosers increase their

level of disclosure when paired with high disclosers in

mutual interview situations. Also, a consistent and

frequent finding about the interpersonal effects of self-

disclosure is that it leads to disclosure reciprocity (Bloch

& Goodstein, 1991).

Three different explanations have been proposed to

explain the latter findings (Berg & Derlega, 1987). The

first explanation is based on a trust-liking approach and

holds that receiving intimate disclosure increases the trust

in and liking for the discloser; the recipient is then

expected to reciprocate in order to demonstrate these

feelings. The second explanation holds that much of

disclosure reciprocity is the result of modeling: the

recipient simply does what they have seen the discloser

already do. The third explanation, and the one which has

gained the most support (Derlega, Harris & Chaikin, 1973;

Davis & Skinner, 1974; Derlega, Chaikin & Herndon, 1975),

emphasizes the influence of social norms: those norms

similar to ones of equity theory govern exchanges of

disclosure and force the recipient of the disclosure to







reveal, in turn, information that is of a comparable level

of intimacy.

Although the effect of health care provider self-

disclosure on client/patient self-disclosure has received

attention in the field of psychotherapy, this body of

research has not led to definitive conclusions. Beutler,

Crago and Arizmendi (1986) reviewed eight studies conducted

before 1986 on the effect of therapist self-disclosure on

client self-disclosure and found that only one naturalistic

study had been conducted and its results were non-

significant. However, seven analogue/process studies had

been conducted and all suggested that therapist self-

disclosure enhances the therapeutic relationship,

facilitates patient engagement in the treatment process, and

enhances self-inspection. Beutler et al. also reported that

a number of investigations suggest a moderately strong

relationship between therapist self-disclosure and patient

subsequent self-disclosure (Derlega et al. 1973; Davis &

Skinner, 1974; Davis & Sloan, 1974; Mann & Murphy, 1975).

In one often cited study, Bundza and Simonson (1973)

presented subjects with transcripts of simulated counseling

sessions. Two of the conditions under study were "warm

support" and "warm support plus self-disclosure." Results

revealed that counselors who were both warm and disclosing

were rated as more nurturant and elicited a greater amount

of patient willingness to self-disclose than did counselors

who were depicted as warm but did not self-disclose. Hence,







these results imply that counselors who rely solely on

facilitative behaviors (such as expressing warmth, respect,

and genuineness), to facilitate disclosure from clients

should consider the effect of their own self-disclosures on

maximizing client disclosure.

Bergin and Lambert (1978) also reviewed this body of

literature and warned that results of analogue studies

should be interpreted with some skepticism. These

researchers stated that results are much less consistent

when applied to clinical populations. Beutler et al.

commented that "Nowhere in the literature does one observe

such a clear disparity between findings of analogue and

clinical investigation as is true with regard to therapist

self-disclosure" (p. 289). Despite the negative tone of

these reviews, a large number of naturalistic studies have

strongly supported Jourard's (1964) contention that self-

disclosure begets self-disclosure. These results, one

should note however, have been collected outside the field

of psychotherapy, and their generalizability to the

therapist-client relationship remains questionable. Despite

the lack of clarity which exists on this issue, the value of

therapist self-disclosure has been emphasized in many core

counseling skill text books. For example, Ivey and Simek-

Downing (1980) write, "[therapist] self-disclosure is

important in the interview as a builder of counselor and

client rapport and adds to the mutuality of the

relationship. It gives a feeling of sharing and can serve







as a model of interpersonal openness. It can serve as a

facilitator for client self-disclosure, sharing and self-

exploration" (p. 84).

Another situational variable which has received

attention is the role or status of the disclosure target.

Slobin, Miller, and Porter (1968) studied employees of a

large company and found that persons were most likely to

self-disclose to peers, next most likely to self-disclose to

superiors and least likely to self-disclose to subordinates.

Furthermore, Cansler and Stiles (1981) gave college subjects

instructions to discuss a topic or a problem with others;

they found that freshmen disclosed more to seniors than vice

versa. Consistent with the latter findings, Stiles and

White (1981) reported that children self-disclose more to

parents than vice versa.

Researchers in the field of medicine have focused their

attention on situational variables found in the context of

patients visiting their physicians. Stimson and Webb (1975)

point out that physician office visits can be very anxiety

provoking to patients. They have reported that symptom

disclosure is reduced when the patient expects to be

embarrassed or scolded by the physician due to the content

of their disclosure. Consistent with this finding, Truax

and Carkhuff (1967) reported that disclosure of important

but potentially harmful information, like that which may be

socially inappropriate or condemnable is increased in

situations where the person in the superior role, such as a







physician or psychotherapist, provides a socially accepting

and supportive problem-solving relationship with the person

in a subordinate role, such as a client or patient.

In summary, the two most salient situational factors

facilitating self-disclosure are the mutual reciprocity of

self-disclosure and the role of the target of the

disclosure. It also appears that the degree of self-

disclosure may be influenced by the interaction of these two

factors.


Topic or Content of Self-disclosure

Goodstein and Reineker claim that all interpersonal

behavior involves some self-disclosure and that since some

information about one's self is rather public, it is readily

disclosed. However, there is other information about one's

self that is private or intimate and is disclosed only under

special circumstances. A number of research findings have

supported these claims. Early studies asked subjects to

rate either their past experiences in sharing information

about themselves or their present willingness to share such

information (Jourard, 1958). Jourard developed a

questionnaire whose items produced two clusters of self-

disclosure: a high disclosure cluster with information about

one's attitudes and opinions, tastes, interests and work;

and a low disclosure cluster with information about one's

finances, personality, and body. Research findings have

determined that these patterns regarding self-disclosure are







highly similar across race, gender, and several national

groups (Jourard and Lasakow, 1958; Jourard, 1959, 1961;

Melikian, 1962; Plog, 1965).

The finding that disclosure clusters exist received

further attention by Berger (1978) who found that when

specific disclosure content is of an emotional or

idiosyncratic nature, self-disclosure is less likely to

occur than when the content is unemotional or task related.

Consistent with these findings, Young (1980) reported that

disclosure of psychological symptoms to physicians was more

difficult than disclosure of private symptoms (genital/

elimination content) or nonprivate symptoms (complaining of

chest pain).

At present, there is a substantial body of literature

which suggests that disclosures regarding personality, body,

and mental state are more difficult to make than other types

of disclosures. This situation poses constraints on the

physician's ability to obtain accurate information for

either effective diagnosis and treatment of psychological

problems or for referral of patients to other health

providers elsewhere for psychological/psychiatric treatment.


Purpose of the Study

For years, the academic medical community has encour-

aged and trained physicians to address the psychological

concerns of their patients. Researchers have indicated that

physicians either ignore or misdiagnose a large number of







psychological disorders (Cassata & Kirkman-Liff, 1981;

Cohen-Cole, Baker, Bird & Freeman, 1982; Block, Herbert,

Schulberg, Coulehan, McClelland & Gooding, 1988; Kamerow,

Pincus & McDonald, 1986; Jones et al., 1987). Physicians'

inability to accurately diagnose and treat psychological

disorders has encouraged researchers to study variables

which hinder or facilitate physician ability to engage

effectively in these endeavors. Research has focused on

physician beliefs that foster avoidance of psychosocial

aspects of health care (Williamson, Beitman & Katon, 1981)

as well as physician attitudes toward psychosocial issues

(Maisiak, Meredith, Bokerm, Rider-Gordon & Scott, 1980).

The majority of research projects have focused on the

physician's role within the context of the physician-patient

relationship. There exists a paucity of research analyzing

the patient's role and how patient behaviors can hinder

physician ability to diagnose and treat psychological

problems. Steinert and Rosenberg (1981) state that

psychosocial care is not provided to patients because many

patients do not feel comfortable enough with their

physicians to describe what is bothering them. These

researchers suggest that future research should attempt to

uncover how to ameliorate patient reluctance to disclose

psychosocial concerns to physicians.

Research has indicated that patient perceptions of

physician behaviors has an effect on their level of

satisfaction with the physician-patient relationship. The







majority of these studies have not focused on whether

physician behaviors which affect patient satisfaction and

compliance also affect patient psychological symptom

disclosure. Young's (1980) study is among the few that have

given attention to patient willingness to disclose

psychological symptoms to physicians. However,

methodological constraints of this study (i.e., using

scripted vignettes as independent variables instead of

actual physician behaviors in the context of medical

interviews) restrict the generalizability of the results.

Further study is warranted in order to ascertain if patient

perceptions of particular physician behaviors have an effect

on patient symptom disclosure. Hence, the purpose of this

study is to examine if subjects' perceptions of physician

competency in conducting a medical interview affects their

psychological symptom disclosure.

Patients are reluctant to disclose symptoms for a

number of reasons, such as, their socially taboo nature,

fear of being sanctioned by a physician, and the anxiety

involved in seeking medical treatment (Stimson & Webb, 1975;

Johnson, 1979; Falvo & Smith, 1983). These factors can

independently or collectively interact to prevent patients

from informing their physician about their psychological or

intimate concerns.

Bloom (1963) divulged two dimensions which collectively

comprise the construct of physician competence during a

medical interview. The first dimension which Bloom coined







"expressive" pertains to the physician's ability to elicit

from the patient all pertinent information and symptoms.

Bloom contended that the more socially skilled the

physician, the greater the likelihood that an accurate

diagnosis could be made. Because of this the physician

would be able to elicit more information from the patient

and consequently have a larger pool of information from

which to base a diagnosis. The second dimension, which

Bloom coined "instrumental," entails the physician's ability

to utilize the information disclosed by the patient to

arrive at an accurate diagnosis and treatment plan. Doyle

and Ware (1977) and Young (1980) have defined Bloom's

expressive dimension as "physician interpersonal or social

skill" and the instrumental dimension as "physician

technical skill." Other researchers have documented

contradictory findings as to whether patients can

differentiate between these two types of skills (Ware &

Snyder, 1975; Falvo & Smith, 1983). However, there is

considerable evidence supporting the argument that patients

expect their physicians to be skilled in both the expressive

and instrumental domains. Consequently, questionnaires have

been designed to tap patient satisfaction with the medical

interview such as the Smith-Falvo Patient-Doctor Interaction

Scale (Falvo & Smith, 1982), and the Medical Interview

Satisfaction Scale (Wolf, Putnam, James & Stiles, 1978) that

are composed of items that load on these two factors.

Again, it can be argued that during a medical interview,







parallel assessment processes occur. While a physician

assesses the patient's problem, the patient assesses the

physician's expressive and instrumental abilities.

Research conducted by Truax and Carkhuff (1967) has

pointed out that the disclosure of important, yet

potentially harmful or embarrassing information is increased

in situations in which the person in the superior role

provides a socially accepting and supportive relationship.

Given that within the context of the physician-patient

relationship, the physician is in the superior role, it is

expected that:

Hypothesis 1: Physicians who are highly skilled in the

expressive dimension will elicit a greater degree of

psychological symptom/fear disclosure from subjects than

physicians having low expressive dimension skills.

Because the expressive dimension has been defined in

terms of the interpersonal/social skills demonstrated by

physicians, this dimension has been readily identifiable to

patients and has received extensive attention in patient

satisfaction research. However, since the instrumental

dimension has been defined in terms of the technical skills

of the physician, this dimension has received less attention

in patient satisfaction literature. This is likely to be

the case because patients are usually lay persons who are

untrained in medical matters, and therefore are not expected

by researchers to be able to provide reliable or valid

assessments of a physician's actual level of technical







skill. Nevertheless, patient perception of a physician's

technical ability is a valid measurement in terms of the

effect that such a perception has on patient behavior,

particularly, their willingness to disclose symptoms. In

this context, the accuracy of subjects' assessment of a

physician's actual technical abilities is inconsequential.

The variable to be studied is the effect of this assessment

on subjects' likelihood to disclose psychological symptoms.

Research on trusting behaviors has found that when

subjects believe that the consequence of taking a risk leads

to a positive outcome, they are more likely to take that

risk (Deutsch, 1973; Young, 1980). Hence, if a patient

perceives their physician as being well skilled in the

instrumental dimension (i.e., high technical skill) during

the course of a medical interview, the patient may deduce

that the physician has the skill to cure their ailment, and

therefore may be more likely to take the risk of disclosing

potentially threatening information to that physician.

Conversely, a patient will be less likely to take the risk

of disclosing potentially threatening information to a

physician whom they perceive as having poor instrumental

skills (i.e., low technical skill) during the course of a

medical interview. Therefore, it is expected that:

Hypothesis 2: Physicians who are highly skilled in the

technical dimension will elicit a greater degree of

psychological symptom/fear disclosure from subjects than

physicians with low technical skills.







Commenting on physician-patient interactions, Stewart

(1984) has indicated that patients are very satisfied with

relationships that include warm, genuine, and respectful

physicians. The effect of instrumental dimension behaviors

(i.e., degree of technical skill), on patient satisfaction

has not been documented, although patient satisfaction has

been found to be highly associated with patient perception

of the amount of communication provided to them by their

physician during a medical interview (Hirsh, 1986). Some

researchers have inferred from the latter finding that

patients are highly satisfied with physicians who

communicate more with them because patients perceive this

behavior as warm and caring. However, other researchers

believe that patients are more satisfied because of the

amount of valuable medical information they receive

concerning their illness and treatment. Research has yet to

be conducted which has empirically analyzed what features

create satisfaction in patients during the medical

interview, is it: (a) the amount of medical information

provided to them about their presenting complaint by the

physician (an instrumental dimension behavior), or (b) the

perception that a physician who truly cares about them would

spend many minutes discussing their treatment plan with them

(an expressive dimension perception). Another question

which merits attention is which dimension, the instrumental

(technical skills), or the expressive (interpersonal

skills), is most influential in eliciting psychological







symptom disclosure from patients. Psychological research

focusing on the therapeutic relationship has repeatedly

documented that client disclosure of social, emotional/

psychological content is highly dependent on client/patient

perception of the therapist's interpersonal skills (Lorr,

1965; Lazarus, 1971; Garfield & Bergin, 1986). Given this

finding, it is expected that:

Hypothesis 3: The degree to which a subject discloses

psychological symptoms/fears is affected to a greater degree

by physician level of expressive skill than physician level

of technical skill.

A myriad of research has focused on gender differences

in self-disclosure studies. To date, the results have been

inconclusive. Some researchers have claimed that females

disclose more than males (Kopfstein & Kopfstein, 1973;

Littlefield, 1974), whereas others have reported that no

differences exist, or that males disclose more readily than

females (Kohen, 1975; Young, 1980). Research out of the

field of psychotherapy has suggested that self-disclosure is

greatest when the therapist-client dyad includes a female,

regardless of whether the female is the therapist or the

client (Fuller, 1963; Brooks, 1974). Also, psychological

research of self-disclosure has shown that perceived warmth

and friendliness of a target person increases the likelihood

that a disclosure will be made to that target person

(Goodstein & Reineker, 1974). Hence, it appears that

researchers have uncovered two circumstances which







facilitate disclosure. One is that socioemotional

disclosure frequently takes place in a context where a

target person is perceived as being warm and friendly, the

other is that the presence of a female in a

psychotherapeutic context increases the amount of self-

disclosure. Given that the skills falling under the heading

of the "expressive" dimension entail interpersonal skills

that can be perceived as being warm and friendly, and that

the presence of a female when psychological issues are

discussed has increased self-disclosure, it is expected

that:

Hypothesis 4: A female physician having high expressive

dimension skills will elicit a greater degree of

psychological symptom/fear disclosure from subjects than a

male physician having high expressive dimension skills.











CHAPTER 3
METHODS


This study was designed to investigate the effect of

four independent variables on the willingness of potential

patients to disclose psychological symptoms and fears:

(a) physician gender, (b) subject gender, (c) subject

perception of physician expressive competence, and

(d) subject perception of physician technical competence. A

series of different factorial designs (i.e., four one-way,

two three-way, and one two-way ANOVA) were used with

physician gender (male or female), subject gender (male or

female), expressive skill (high or low), and technical skill

(high or low) constituting the variables under study.

The level of physician competence was manipulated along

two dimensions, an expressive skill dimension and a

technical skill dimension. High or low expressive dimension

competence was manipulated by providing a videotaped medical

interview of a simulated patient and a physician. One

videotape depicted the physician possessing high expressive

dimension skills. In this interview the physician

pleasantly greets and bids farewell to the patient, behaves

in an accepting manner towards the patient (i.e., does not

criticize the patient), uses language the patient

understands (i.e., uses words that are not too technical),







asks questions that are socially appropriate (not highly

personal), and self-discloses personal information to the

patient. A second videotape depicted the physician as

possessing low expressive dimension skills. In this

interview the physician does not pleasantly greet or bid

farewell to the patient, criticizes the patient for not

taking better care of himself, uses words that are very

technical and unlikely to be understood by lay persons, asks

questions that are very personal, and does not self-disclose

any personal information to the patient. In both of these

interviews, the level of physician medical technical skill

was held constant at a moderate level. A third videotape

depicted the physician as possessing high medical technical

skills. In this interview, the physician spends three

minutes presenting information and discussing prevention

without appearing rushed, provides extensive information to

the patient, asks the patient five questions about the

presenting complaint, tells the patient what the physical

examination will entail, and alerts the patient as to when

to expect discomfort during the examination. A fourth

videotape depicted a physician possessing low medical

technical skills. In this interview the physician spends 20

seconds presenting information and discussing prevention

while appearing rushed, provides only minimal information to

the patient about their diagnosis and treatment, asks only

one question about the presenting complaint, initiates a

physical examination without telling the patient what the







physical examination will entail, and does not tell the

patient when to expect discomfort during the examination.

In both of these interviews, the level of physician

expressive dimension skill was held constant at a moderate

level. A male and a female physician were employed in each

of the four conditions, therefore creating eight separate

videotaped interviews (four interviews conducted by a male

physician, and four interviews conducted by a female

physician). After viewing one of the eight interviews, a

subject's willingness to disclose symptoms to the particular

physician they viewed was assessed via a questionnaire.


The Videotaped Medical Interviews

Two experienced family practice physicians (one male,

one female), and one male simulated patient engaged in eight

simulated medical interviews. The physicians and the

simulated patient were briefed as to the purpose of this

study and to their designated roles. These three volunteers

were provided with scripts to guide their behavior during

the interviews (see Appendix A for the scripts). Each

physician and the simulated patient engaged in four medical

interviews which were videotaped. One medical interview

depicted the physician behaviorally enacting high expressive

dimension skills. A second interview depicted the physician

behaviorally enacting low expressive dimension skills. A

third interview depicted the physician behaviorally enacting

high medical technical skills. The final interview depicted







the physician behaviorally enacting low medical technical

skills. A total of eight interviews were videotaped (four

depicting a male physician with the simulated patient, and

four depicting a female physician with the simulated

patient).


Operational Definitions

In this study, physician competence was defined along

two dimensions, an expressive skill dimension, and a

technical skill dimension. The expressive dimension refers

to physician behaviors which may facilitate disclosure (from

a patient) of information pertinent to the patient's health.

More specifically, high or low levels of expressive skill

were operationally defined in terms of the presence (high)

or absence (low) of the following behaviors: (a) pleasantly

greeting and bidding farewell to a patient, (b) behaving

uncritically toward a patient, (c) using language that lay

persons would understand, (d) asking socially appropriate

questions, and (e) self-disclosing personal information to a

patient.

Physician competence on the technical dimension refers

to behaviors which may facilitate a physician's ability to

make an accurate diagnosis, formulate an appropriate treat-

ment plan, and articulate this plan to a patient. A high or

low level of technical dimension competence is operationally

defined in terms of the presence (high) or absence (low) of

the following behaviors: (a) spending three minutes







presenting information and discussing prevention without

appearing rushed, (b) providing extensive information to the

patient when asked to do so, (c) asking the patient five

questions about the presenting complaint, (d) telling the

patient what the physical examination entails, and

(e) warning the patient as to when to expect discomfort

during the examination.


Subjects

The subject pool of this study consisted of one-hundred

twenty-eight college students from a midwestern university

who received undergraduate psychology course credit for

their participation. Forty-eight subjects (24 males and

24 females) participated in the pretest study. Eighty

subjects (40 males and 40 females) participated in the final

study.


Procedure

Students currently enrolled in an undergraduate

psychology course who had acknowledged in writing that they

had an interest in obtaining extra credit were contacted

by phone and asked to participate in the experiment which

took place in an on-campus building. As groups of five male

and five female students entered the testing room,

they were asked to read, then sign, a consent form prior

to participating in the study (see Appendix B). Subjects

were assigned experimental numbers and seated in one of

20 desks. The desks were placed in front of a 24-inch







television screen so that subjects would have an

unobstructed view of the screen. After all subjects had

been seated, the proctor informed them that they would be

viewing a videotaped medical interview between a physician

and a patient. The proctor then presented the videotape to

the subjects.

After viewing the videotape, each subject was handed

a questionnaire (see Appendix C for the Symptoms

Questionnaire). Each questionnaire required that

demographic information be provided, that the subjects

imagine that they were about to engage in a medical

interview with the physician they viewed, and then report

their willingness to disclose 18 separate symptoms or fears

to that physician. Subjects who responded having been

willing to disclose at least one psychological symptom were

required to respond to a Likert-type scale. The scale

addressed the relative importance of ten particular

physician behaviors (i.e., the manipulated expressive and

technical dimension behaviors) on the subject's willingness

to disclose the psychological symptomss. Upon completing

the questionnaire, each subject was asked to hand the

questionnaire to the proctor, then proceed to the back of

the room to receive course credit.


Variables

Four independent variable/factors, i.e., physician

gender, subject gender, physician expressive competence







(high or low), and physician technical competence (high or

low) were manipulated in a series of four ANOVA designs.

One male and one female physician conducted four

simulated medical interviews, all of which were videotaped.

Forty male and forty female subjects viewed one of eight

randomly assigned videotaped medical interviews.

The expressive competence of the physician was

manipulated by providing depictions on the videotape

indicating high or low levels of expressive skill.

In the high expressive competence manipulation, the

physician: (a) pleasantly greeted and bade farewell to the

patient, (b) behaved in an accepting manner toward the

patient, i.e., did not criticize the patient for any reason,

(c) used language that the patient would understand, i.e.,

not highly technical, (d) asked questions that were socially

acceptable, i.e., not very personal, and (e) self-disclosed

personal and emotional information to the patient.

In the low expressive competence manipulation, the

physician: (1) went directly to the medical problem without

first greeting the patient and did not bid the patient a

pleasant farewell at the end of the interview,

(2) criticized the patient for not taking care of himself,

(3) used language that was too technical for the patient to

understand, (4) asked questions that were too personal, and

(5) did not self-disclose any personal or emotional

information to the patient. In both high and low expressive

dimension interviews, the physician's level of technical







skill was held constant by depicting the physician:

(1) asking the patient to tell him/her the present

complaint, (2) asking the patient if he had eaten anything

which has made the patient ill, (3) examining the patient's

chest, throat and ears, (4) discussing the diagnosis and

treatment for one and one-half minutes, and (5) asking the

patient at the end of the interview if he had additional

questions.

The level of medical technical skill of the physician

was manipulated by providing depictions on the videotape of

high or low levels. In the high technical skill

manipulation, the physician: (1) spent three minutes

presenting information and discussing prevention, without

appearing rushed, (2) provided enough information about the

diagnosis and treatment, and asked the patient if more

information was needed, (3) asked the patient a series of

questions regarding his presenting complaint, (4) told the

patient what the physical examination would entail, and

(5) told the patient when to expect discomfort, and the

simulated patient acted as if he was not handled roughly by

the physician.

In the low technical skill manipulation, the physician:

(1) spent 20 seconds presenting information and discussing

prevention while appearing rushed, (2) provided very little

information to the patient, i.e., provided two statements,

(3) asked only one question to the patient regarding the

patient's presenting complaint, (4) did not tell the patient







what the physical examination would entail, and (5) did not

tell the patient when to expect discomfort during the

examination, and the simulated patient reacted as if he was

handled roughly by the physician. Held constant in both

levels of technical competence was the physician's level of

expressive competence. In both high and low technical

competence interviews the physicians' level of expressive

skill was held constant by depicting the physician as:

(1) introducing themselves to the patient, (2) complimenting

the patient for appearing physically fit, (3) acknowledging

all of the patient's questions, (4) shaking hands with the

patient at the end of the session.

The preceding criteria for what constitutes patient

perceived physician competence in high and low levels of

expressive and technical dimensions are derived from items

on the Smith-Falvo Doctor-Patient Interaction Scale (Falvo &

Smith, 1983; see Appendix D). This scale was constructed

from a multidimensional scaling of physician behaviors that

patients considered to be important during the course of

medical interviews. The technical and expressive dimensions

were represented by items on the scale, and this researcher

utilized items which clustered on the technical skill

dimension as criteria reflecting technical competence, and

items which clustered on the expressive dimension as

criteria reflecting expressive competence (Smith, Falvo,

McKillip & Pitz, 1984). Falvo and Smith (1983) reported a

convergent validity coefficient between the SFPDIS and the







Medical Interview Satisfaction Scale of .74. The Medical

Interview Satisfaction scale is a measure of physician

behavior which reports favorable reliability and validity

coefficients. Concurrent validity of the SFPDIS was also

reported; there was a high positive correlation between

scale scores and patient's reported intentions to return to

their physician for future health care (r = .73). The

authors also reported test-retest reliability coefficients

of .76 and an internal reliability coefficient of .85 as

measured by Cronbach's Alpha (Falvo & Smith, 1983).

One item, whether the physician self-disclosed social

or emotional information to the patient, was not derived

from the items on the SFPDIS. This item was included based

on literature indicating that the most consistent finding

regarding interpersonal effects of self-disclosure is that

of disclosure reciprocity (Berg & Derlega, 1987). Garfield

and Bergin (1986) also reported that self-disclosure of

psychotherapists increased the amount of self-disclosure by

clients. The inclusion of a physician self-disclosing to

the simulated patient attempted to ascertain if the same

dynamic ensues within the context of the physician-patient

relationship. Hence, physician self-disclosure is included

within the domain of the expressive dimension because this

dimension entails, according to Bloom (1963), any physician

behavior which may increase information elicited from the

patient.







A pretest was conducted to confirm that the videotaped

interviews created statistically distinct perceptions of

the high versus low levels of expressive and technical

competence.



Dependent Variables

Willingness to disclose symptoms and fears was measured

by requiring subjects to answer the following question to

each of the 18 symptoms listed below: "How easy or difficult

would it be to talk about each of the following issues with

the physician you just saw on the videotape?" The scale

ranged from +3 to -3 (in which + 3 represents "extremely

easy to discuss fully, no resistance, would hold nothing

back"; +2 represents "easy to discuss, little resistance to

discuss it"; +1 represents "fairly easy to discuss, very

little hesitation to discuss it"; 0 represents "moderately

easy, some resistance, would hold back some of my fears and

symptoms"; -1 represents "hard to discuss, hesitant to

discuss it"; -2 represents "very hard to discuss, very

hesitant to discuss it"; and -3 represents "extremely hard

to discuss, great resistance, would hold back most

information"). The dependent variables were computed by

summing the subject's willingness to disclose private,

nonprivate, and psychological symptoms or fears.

Although the dependent variable of interest in this

study was disclosure of psychological symptoms and fears,

the content of this disclosure could also be labeled as







"private" because of the social embarrassment this culture

ascribes to it. However, subjects can experience problems

of a private nature that are not primarily psychological in

origin. In order to differentiate between the different

content areas, the following symptoms and fears were

designated: (a) private medical symptoms and fears regarding

venereal disease, acquired immunodeficiency syndrome (AIDS),

urinary bleeding, hemorrhoids, genital sores, and loss of

bowel control; (b) nonprivate medical symptoms and fears

regarding appendicitis, loss of vision, loss of functioning

in a limb, heart attack, ulcer, and migraine headaches;

(c) psychological symptoms and fears including depression,

anxiety, obesity, marital discord, substance abuse, and

sexual problems.

The 18 items selected were based on the results of a

factor analysis conducted by Young (1979). The factor

analysis identified three areas which were difficult for

subjects to disclose to physicians, i.e., the private,

nonprivate, and psychological domains under study here. The

items used in this study within the domains of private and

nonprivate medical concerns include some of Young's (1980)

items. However, this researcher added the item "AIDS" to

the private domain given that this disease has gained much

attention and is considered to be a taboo or private

disease. Also, this researcher added "heart attack" to the

nonprivate domain items given the recent upsurge of media

attention to this ailment as a consequence of an elevated







cholesterol level. The items used in this study of a

psychological nature are those which Cassata and Kirkman-

Liff (1981) report are the most common mental health

problems encountered by family physicians during office

visits. The items used in this study as dependent variables

were ones which loaded on the three factors described by

Young and, in addition, items which have recently become a

source of concern to the general public. An internal

reliability coefficient of .94 as measured by Cronbach's

Alpha was found for the symptoms questionnaire.


Hypothesis Testing

Hypothesis 1, stating that physicians who are highly

skilled in the expressive dimension will elicit a greater

degree of psychological symptom/fear disclosure from

subjects than physicians having low expressive dimension

skills, was tested by using a one-way ANOVA with level of

expressive skill (high or low) as independent variables and

degree of psychological symptom disclosure as the dependent

variable. A 2x2x2 between subjects' factorial design was

conducted to test for possible interaction effects between

physician gender, subject gender and level of expressive

skill. The first variable was physician gender (male or

female), the second variable was subject gender (male or

female), and the third variable was the level of the

physician expressive skill (high or low with the physician's

level of technical skill held constant). The symptom







questionnaire assessed subject willingness to disclose

psychological symptoms (the dependent variable under study),

to the physician viewed on videotape.

Hypothesis 2, stating that physicians who are highly

skilled in the technical dimension will elicit a greater

degree of psychological symptom/fear disclosure from

subjects than physicians with low technical skills, was

tested by using a one-way Anova with level of technical

skill (high or low) as independent variables and degree of

psychological symptom disclosure as the dependent variable.

A 2x2x2 between subjects' factorial design was conducted to

test for possible interaction effects between physician

gender, subject gender and level of expressive skill. The

first variable was physician gender (male or female), the

second variable was subject gender (male or female), and the

third variable was the level of the physician's technical

skill (high or low with the physician's level of expressive

skill held constant). The symptom questionnaire assessed

subject willingness to disclose psychological symptoms to

the physician viewed on videotape.

In order to test Hypothesis 3, stating that the degree

to which a subject discloses psychological symptoms/fears

will be affected to a greater degree by physician level of

expressive skill than physician level of technical skill, a

one-way, four level between subject's factorial design was

used. The variable under study was labeled as physician

competence. The four levels of physician competence were







(1) high expressive competence with technical competence

held constant, (2) low expressive competence with level of

technical competence held constant, (3) high technical

competence with level of expressive competence held

constant, and (4) low technical competence with level of

expressive competence held constant. The symptoms

questionnaire assessed subject willingness to disclose

psychological symptoms to the physician viewed on videotape.

Hypothesis 3 was indirectly tested by employing a

multiple comparison of means between the four groups under

study. The characteristics of this statistical test

precluded that: (1) an overall ANOVA treatment effect would

not have to be significant, (2) the means of the groups in

question would be used, and (3) that the comparison would

assess two or more means.

The statistical procedure included: (1) deciding on the

types of comparisons to be made; (2) arranging co-efficients

such that, when multiplied by the means being compared, they

sum to zero within a comparison, i.e., when the null

hypothesis is true, and that they are orthogonal across all

pairs of comparisons; (3) the calculation of the weighted

sum, standard error, and t-value for each comparison; and

(4) comparing the obtained t-value from these calculations

with the critical t-value.

The third hypothesis was indirectly tested by

implementing a multiple comparison among the respective

means of the high versus low expressive dimension







manipulations and the high versus low technical dimension

manipulations.

This comparison of means would detect if the expressive

dimension manipulation from high to low level led to more

pronounced differences in subject psychological symptom

disclosure than the high to low technical dimension

manipulation. To the extent that a statistically

significant difference existed between the two respective

differences, i.e., the expressive dimension manipulation

difference is greater than the technical dimension

manipulation difference, then the third hypothesis would be

supported.

The fourth hypothesis, stating that a female physician

having high expressive dimension skills will elicit a

greater degree of psychological symptom/fear disclosure from

subjects than a male physician having high expressive

dimension skills, was tested by conducting a one-way ANOVA

with physician gender as the independent variable and degree

of psychological symptom disclosure to the high expressive

physicians as the dependent variable. A 2x2 between

subjects' factorial design was conducted to test for

possible physician gender-subject gender interaction. The

first variable refers to physician's gender, and the second

variable refers to subject gender.











CHAPTER 4
RESULTS


Pretest of the Expressive &
Technical Competence Manipulations

A pretest confirmed that the videotaped interviews

created statistically distinct perceptions of the high

versus low levels of expressive competence and technical

competence. The pretest also confirmed that the competence

variables, which were held constant in the experimental

design, in each condition, were perceived by subjects as

constant.

Subjects (n = 24 male and 24 female undergraduates at a

midwestern university) were randomly assigned to one of

eight videotaped interviews and then completed a pretest

questionnaire (see Appendix E). Five items measured level

of expressive competence on a 5-point Likert type scale.

These five items assessed the subject's opinion of: (1) the

degree to which the subject perceived the physician to care

about the patient, (2) the degree of friendliness of the

physician, (3) the degree of kindness/warmth of the

physician, (4) how relaxing the physician's behavior was,

and (5) how comforting the physician's social style was.

An additional five items measured the physician's level

of technical skill on a 5-point Likert type scale. These







five items assessed the subject's opinion of: (1) the level

of the physician's technical competence, (2) the physician's

capacity to heal, (3) how fit the physician is to be a

physician, (4) how qualified the physician is to be a

physician, and (5) how well trained the physician is. An

internal reliability coefficient of .88, as measured by

Cronbach's alpha was found for the pretest questionnaire. A

concurrent validity coefficient of .71 was also found; there

was a high positive relationship between pretest scale

scores and subject willingness to disclose symptoms of a

general, private and psychological nature to the physician

viewed on videotape.

A series of four one-way ANOVA between subject designs

were conducted (see Appendix F). The first ANOVA analyzed

the high versus low expressive competence manipulation.

Results indicate a significant between groups effect of

expressive condition [high expressive mean = 4.28 vs. low

expressive mean = 1.58, F(1,22) = 220.25, p < .01]. The

second ANOVA analyzed the high versus low technical

competence manipulation. Results indicate a significant

between groups effect of technical condition [high technical

mean = 4.13 vs. low technical mean = 1.66, F(1,22) = 25.38,

p < .01). The third ANOVA analyzed whether technical skill

ratings remained constant as expressive skill was

manipulated from high to low levels. Results indicate that

there were no significant between groups technical skill

effects [F < 4.49]. The final ANOVA analyzed whether







expressive skill ratings remained constant as technical

skill was manipulated from high to low levels. Results

indicated that there were no significant between groups

expressive skill effects [F < 4.49]. In order to test for

the possibility of interaction effects between physician

gender, subject gender and level of skill, a series of four

2x2x2 ANOVA between subjects' designs were conducted (i.e.,

with the first variable referring to physician gender, the

second variable to subject gender and the third variable to

conditoin level). Result indicated that there were no

significant interaction effects found in any of the four

ANOVAS.


Demographics

The sex and racial/ethnic group distributions of the

sample of 80 undergraduates enrolled in a psychology course

are presented in Table 4.1. Males and females were equally

represented in the sample. The mean age of the sample was

18.99 years (males = 19.45, females = 18.52). Whites

comprised 90% of the sample, blacks comprised 6.25% and the

remaining subjects, 3.75%, were Asian or of Asian descent.

Members of the freshman class comprised 37.5% of the sample,

sophomores comprised 27.5%, juniors comprised 27.5% and

seniors comprised 7.5%.


Hypothesis 1

The findings supported the hypothesis that physicians

who demonstrate a high level of skill in the expressive







Table 4.1. Population of Study


White Black Other*

Year 1 2 3 4 1 2 3 4 1 2 3 4


Male 7 6 18 4 1 1 2 0 0 1 0 0

Female 21 12 2 2 0 1 0 0 1 1 0 0

Total 28 18 20 6 1 2 2 0 1 2 0 0


*Asian or Asian-American.





dimension elicit greater willingness to disclose psycho-

logical symptoms from subjects than physicians having low

expressive dimension skills [high expressive mean = 3.50;

low expressive mean = -11.0 F(1,38) = 145.13, p < .01] (see

Table 4.2). Results of a test of interactions indicated

that there were two main effects and two first-order inter-

action effects (see Table 4.3). When disclosing to a physi-

cian showing high expressive skills compared to a physician

showing low expressive skills, subjects reported greater

willingness to disclose psychological symptoms [high expres-

sive mean = 3.50; low expressive mean = -11.0, F(1,32) =

96.72, p < .01]. The second main effect indi-cated that

willingness to disclose psychological symptoms was greater

to male physicians than to female physicians (disclosure to

the male physician mean = -2.2 vs. disclosure to the female

physician mean = -5.3, F(1,32) = 4.59, p < .05].







Table 4.2


Hypothesis One: Expressive Skill Manipulation
Summary Table


Source SS df MS F


Between 263.175 1 2638.175 145.13

Within 690.800 38 18.178

Total 3328.975 39











Table 4.3. Hypothesis One: Expressive Skill Manipulation
Test for Interaction Summary Table


Source SS df MS F


Main Effects 2247.275 3 749.092 34.70**
Physician Gender 99.225 1 99.225 4.60*
Subject Gender 60.025 1 60.025 2.78
Expressive Skill 2088.025 1 2088.025 96.72**

2-Way Interactions 390.675 3 130.225 6.03**
Phys. X Subject Gen. 172.225 1 172.2 7.97**
Phys. Gender X Exp Skill 172.225 1 172.200 7.97**
Sub. Gender X Exp Skill 46.225 1 46.200 2.10

3-Way Interactions .225 1 .225 .01
Phys. X Sub. X Exp Skill .225 1 .225 .01


Explained
Residual
Total

*p < .05
**p < .01


2638.175
690.800
3328.975


376.882
21.588
85.358


17.458







A first order interaction effect was found between

physician gender and subject gender [F(1,32) = 7.97,

p < .01]. Specifically, the mean self-disclosure score for

male subjects viewing a male physician was significantly

higher than the mean scores of the other conditions (i.e.,

female subject-male physician condition, female subject-

female physician condition and male subject-female physician

condition) (see Figure 4.1). Female subjects did not report

a greater willingness to disclose psychological symptoms

based on physician gender [F < 4.15], whereas male subjects

reported a greater willingness to disclose to male

physicians than female physicians [male subject disclosure

to male physician mean = 1.1 vs. male subject disclosure to

female physician mean = -6.20, F(1,32) = 12.18, p < .01].

Also, results indicated that males were more reluctant to

report psychological symptoms to a physician of the opposite

sex than were female subjects.

Physician gender also interacted with physician level

of expressive competence [F(1,32) = 7.97, p < .01]. This

result indicated that when disclosing to a male physician

showing high expressive competence, compared to a female

physician showing high expressive competence, subjects

reported a greater willingness to disclose psychological

symptoms to the male physician (see Figure 4.2) [mean

disclosure to male physician = 7.1 vs. mean disclosure to

female physician = -.10]. More specifically the mean self-

disclosure score for all subjects viewing the high










+180


30


(11)

0-

DPSD*

-30

(-44)


60 (-62)

-180
Male Physician Female Physician


X = Male Subjects
= Female Subjects

DPSD = Degree of Psychological Symptom Disclosure


Figure 4.1. ANOVA I Expressive Dimension Manipulation:
(Physician Gender x Subject Gender
Interaction).











+180 :

75




0 -

DPSD*

75




-150

-180 -


High Expressive


Low Expressive


X = Male Physician
*= Female Physician

* DPSD = Degree of Psychological Symptom Disclosure


Figure 4.2.


ANOVA I Expressive Dimension Manipulation:
(Physician Gender by Level of Expressive
Skill Interaction).


(-105)
(-115)







expressive male physician was significantly greater than the

mean disclosure scores of the other three conditions (i.e.,

high expressive female physician condition, low expressive

female physician condition, low expressive male physician

condition). Also, the mean self-disclosure score for all

subjects viewing the high expressive female physician was

significantly greater than the mean disclosure scores to the

low expressive female physician and the low expressive male

physician. Results also indicated that the manipulation

from high to low level of expressive skill led to a greater

decrease in subject willingness to disclose psychological

symptoms to the male physician than to the female physician.


Hypothesis 2

The findings supported the hypothesis that physicians

who demonstrated a level of high skill in the technical

dimension elicited a greater willingness to disclose

psychological symptoms from subjects than physicians

demonstrating low technical skill [high technical mean

2.5; low technical mean = -8.2 F(1,38) = 47.97, p < .01]

(see Table 4.4). Results of a test of interaction found a

first order interaction effect between physician gender and

subject gender [F(1,32) = 4.42, p < .05] (see Table 4.5 and

Figure 4.3). Specifically, the mean psychological dis-

closure score for male subjects viewing a male physician

was significantly higher than the mean disclosure scores of

the other conditions (i.e., female subject-male physician

condition, female subject-female physician condition and







Table 4.4


Hypothesis Two:
Summary Table


Technical Skill Manipulation


Source SS df MS F


Between Group 1462.7 1 1462.70 47.97

Within Group 1158.8 38 30.49

Total 2621.5 39








Table 4.5. Hypothesis Two: Technical Skill Manipulation
Summary Table


Source SS df MS F


Main Effects 1151.00 3 383.6 10.59**
Physician Gender 48.40 1 48.4 1.34
Subject Gender .10 1 .1 .00
Technical Skill 1102.50 1 1102.5 30.40**

2-Way Interactions 297.30 3 99.1 2.73
Phys. X Subject Gender 160.00 1 160.0 4.40*
Phys. X Technical Skill .40 1 .4 .01
Sub. Gender X Tech. Skill 136.90 1 136.9 3.78

3-Way Interactions 14.40 1 14.4 0.40
Phys. X Gender X Subject
Gender X Tech. Skill 14.40 1 14.4 0.40


Explained
Residual
Total

*p < .05
**p < .01


1462.70
1158.80
2621.50


208.960
36.213
67.218


5.77


























(-37)


(-28)




(-49)


Male Physician


Female Physician


X = Male Subject
* = Female Subjects

* DPSD = Degree of Psychological Symptom Disclosure


Figure 4.3.


ANOVA II Technical Dimension Manipulation:
(Physician Gender x Subject Gender
Interaction).


+ 180 :


10 -

0-

-10-

-20


DPSD*


- 30 -

-40-

- 50
-S -

60 -


-180 3







male subject-female physical condition). Further analysis

indicated that the physician gender effect was more

influential to male subjects than female subjects. Whereas

female subjects did not report significantly greater

willingness to disclose psychological symptoms based on

physician gender [F < 4.15], male subjects did [disclosure

to male physician mean = 0.40 vs. disclosure to female

physician mean = -5.3, F(1,32) = 4.50, p < .05]. Also,

results indicated that male subjects were considerably more

reluctant to disclose psychological symptoms to a physician

of the opposite sex.


Hypothesis 3

The findings did not support the hypothesis that

willingness to disclose psychological symptoms was greater

to a physician who demonstrated high expressive skill than

to one who showed high technical skill. Results of a

planned comparison of means indicated that the difference in

willingness to disclose psychological symptoms between high

versus low expressive physicians and the difference between

high versus low technical physicians was not significantly

different (t < 1.98). However, the difference between the

high versus low expressive skill manipulation and the high

versus low technical skill manipulation was in the expected

direction (mean differences = 14.5 for the expressive skill

manipulation difference versus 10.7 for the technical skill

manipulation difference).







A post hoc analysis was conducted to determine if

Hypothesis 3 received support among male compared to female

subjects. Whereas the difference between the high vs.

low technical dimension difference and the high vs. low

expressive dimension difference was not statistically

significant amongst male subjects (t < 2.04), a significant

difference was found among female subjects (t > 2.04,

df = 30, p < .05).


Hypothesis 4

The findings did not support the hypothesis that a

female physician high in expressive skill would elicit

greater willingness to disclose psychological symptoms than

a male physician high in expressive skill. When disclosing

to either a male or a female physician high in expressive

competence, results indicated that subjects reported a

greater willingness to disclose psychological symptoms to a

male physician than to a female physician [disclosure to

male physician mean = 7.1; disclosure to female physician

mean = -.10, F(1,18) = 13.56, p < .05]. These results are

reported in Table 4.6 (see Figure 4.4). A test of physician

gender and subject gender interaction found no significant

interaction effect (see Table 4.7).


General. Private, and Psychological Symptom Disclosure

Results indicated that willingness to disclose general

symptoms was greater than the willingness to disclose both

private symptoms (general symptoms mean = 9.01; private







Table 4.6.


Hypothesis 4: Summary Table of High
Expressive Condition When Disclosing
Psychological Symptoms to a Male vs. Female
Physician


Source SS df MS F


359.35 1 359.35 13.56


Between Group

Within Group


503.60 18

862.95 19


26.50


Total











+ 180


50

40

30

20
DPSD*


0

-10

- 20

- 30


-180 g


(47)


(10)




(-11)


Male Physician


Female Physician


X = Male Subjects
S= Female Subjects

* DPSD = Degree or Psychological Symptom Disclosure


Figure 4.4.


ANOVA II High Expressive Dimension Condition:
Degree of Psychological Disclosure to Male
vs. Female Physician.







Table 4.7.


Hypothesis 4: Summary Table of Test of
Interaction Between Subject and Physician
Gender When Disclosing Psychological Symptoms
to a High Expressive Physician


Source SS df MS F


Main Effects 266.90 2 133.45 4.24
Physician Gender 266.45 1 266.45 8.46*
Subject Gender .45 1 .45 .01

2-Way Interactions 92.45 1 92.45 2.93
Phys. X Subject Gender 92.45 1 92.45 2.93


Residual
Total

*p < .05


503.60 16
862 19


31.47
45.40







symptoms mean = -6.25; F = 227.2, df = 79, p < .01) and

psychological symptoms (general symptoms mean = 9.01;

psychological symptoms mean = -3.26, F = 228.78, df = 79,

p < .01). However, unexpectedly, results indicate less

reluctance to disclose psychological symptoms than private

symptoms (psychological symptoms mean = -3.26 vs. private

symptoms mean = -6.25, F = 17.92, df = 79, p < .01).


Correlational Analysis

The Pearson correlations between the degree of

psychological symptom disclosure and the degree of

importance of ten physician behaviors (five expressive

dimension and five technical dimension behaviors) can be

found in Table 4.8.

Forty-nine subjects (25 males, 24 females) reported

that it would be "fairly easy to discuss," "easy to

discuss," or "extremely easy to discuss" at least one of six

psychological symptoms. Results indicate that no particular

physician behavior was significantly correlated with male

and female subject's willingness to disclose psychological

symptoms.

Correlations were also calculated among the different

physician-subject gender dyads. The only statistically

significant correlation was between a technical dimension

behavior ("The thorough way in which the physician ques-

tioned the patient") and willingness to disclose psychologi-

cal symptoms (r = .69, df = 9, p < .01). This correlation

was found in the female physician-male subject dyad.







Table 4.8 Correlations Between Degree of Psychological
Symptom Disclosure and Degree of Importance of
Physician Behaviors.


N=49 N=15 N=10 N=11 N=13
All M.Phys- M.Phys- F.Phys- F.Phys-
Subjects M.Sub. F.Sub. M.Sub. F.Sub.


P.B.*

1) -.10 -.25 -.02 -.25 .24
2) .15 .23 .04 .32 .01
3) .12 -.05 .18 .44 .25
4) .06 .05 -.15 .03 .25
5) .19 .00 .12 .10 .40
6) .16 .21 .28 -.07 .23
7) -.05 -.20 .04 .31 .17
8) .14 -.10 .03 .69** .38
9) .02 -.17 -.38 .41 .25
10) .00 -.24 .15 .68 -.12
GE) .13 -.02 .05 .33 .30
GT) .06 -.16 .01 .67 .19


*Physician Behavior:
**p < .01

1. The greeting provided by the physician.
2. The acceptance of the patient expressed by the
physician (i.e., the lack of criticism expressed by the
physician to the patient.
3. The language the physician used (i.e., not too
technical).
4. The questions asked by the physician were socially
appropriate (i.e., not too personal).
5. The way the physician disclosed personal information to
the patient.
6. The way the physician took time to speak with the
patient.
7. The way the physician made sure to provide enough
information to the patient.
8. The thorough way in which the physician questioned the
patient.
9. The way the physician informed the patient about the
forthcoming physical examination.
10. The way the physician told the patient when to expect
discomfort.
GE. The summed total of the five expressive dimension
behavior scores (questions 1-5).
TE. The summed total of the five technical dimension
behavior scores (i.e., questions 6-10).




77


Finally, correlations were also calculated collapsing

the five expressive dimension behavior scores into one

"general" expressive dimension score and collapsing the five

technical dimension scores into one "general" technical

dimension score. Again, no statistically significant

correlations were found.











CHAPTER 5
DISCUSSION


Support was found for the hypotheses that predicted

greater willingness to disclose psychological symptoms to

physicians showing: (a) high expressive competence

(Hypothesis 1) and (b) high technical competence (Hypothesis

2) than those demonstrating low levels of competence on

these dimensions. There was not sufficient evidence to

conclude that the expressive dimension has a greater effect

on psychological symptom disclosure than the technical

dimension for both male and female subjects (Hypothesis 3).

However, it is important to note that the design employed to

test the third hypothesis did so indirectly (by testing

whether the difference between high and low expressive skill

was significantly different than the difference between high

and low technical skill). Also, contrary to expectations,

male physicians demonstrating high expressive dimension

skills elicited greater psychological symptom disclosure

from subjects than female physicians showing high expressive

dimension skills (Hypothesis 4). It is warranted that a

finding of the pretest manipulations (testing high vs. low

levels of expressive and technical skill) be taken into

consideration prior to forming any definitive conclusions

of previously described results. A comparison of the







relative F-values of the expressive skill pretest and the

technical skill pretest indicates that the expressive

condition F-value (F = 220.25) is nearly eight times greater

than the technical condition F-value (F = 28.38). Although

both of these F-values are significant at the p < .01 level,

one value is nevertheless indicating a much greater effect.

Hence, the impact of the expressive dimension on subject's

psychological symptom disclosure is much stronger than the

impact of the technical dimension. An explanation for this

discrepancy may be that subjects are more certain of how to

assess physician interpresonal skill than how to assess

physician technical skill. This degree of certainty may

have been at least partly responsible for the disparity in

the two respective F-values. This disparity in F-values

must be taken into consideration when discussing the results

of tests analyzing the relative influence of the expressive

and technical dimension on psychological symptom disclosure.


Analysis of Hypotheses One and Four (ANOVA 1)

Results suggest that during the initial medical

interview, patients will be assessing physician behavior, as

well as seeking advice about their medical problems. In the

absence of any clear technical skill deficit in the

physician's behavioral repertoire, patients are likely to

focus on and assess other aspects of the physician's

behavior; i.e., bedside manner (Bloom, 1963). The ANOVA

results of Hypothesis 1 suggest that when a physician's







level of demonstrated technical skill was moderate, with no

indication of a deficit in technical skill, subjects focused

on other behaviors to assess the physician's competence.

Hence, physicians who showed a high level of skill in

"bedside manner" (high expressive skill), and were

moderately technically skilled, elicited greater willingness

to disclose psychological symptoms than did physicians with

a comparable level of technical skill but poor "bedside

manner" (low expressive skill).

These results are consistent with other research

findings indicating that the disclosure of important, but

potentially harmful, personal information is increased in

environments in which the person in the superior role (in

this case, the physician) creates a socially accepting and

supportive problem-solving relationship with the person in

the subordinate role (in this case the subject; Balint &

Balint, 1957; Truax and Carkhuff, 1967; O'Reilly, 1978).

These results are also consistent with Young's (1980)

contention that physicians high in expressive skill would

elicit greater symptom disclosure from subjects than would

physicians low in expressive skill.

Given that technical skill, as construed by patients,

involves behaviors that are likely to occur during the

course of any medical interview, it can be assumed that in

most medical interviews physicians are likely to be

perceived by patients as at least moderately technically

skilled. Consistent with this argument is the finding that







the extent of consumer ignorance is greater in the case of

medical care than for the vast majority of consumer goods

and services (Arrow, 1971). State regulations and licensing

of physicians have been undertaken in order to guarantee

minimal levels of training and competency. In the absence

of behavioral indicators of either high or low levels of

technical skill, knowledge of physician licensing

regulations may lead some consumers to ascribe at least a

moderate level of technical skill to a physician. The

results of ANOVA I suggest that in such instances, the

physician's level of expressive skill can play a crucial

role in the interchange of important information from the

patient to the physician.

Another result from ANOVA I suggests that male

physicians elicit greater willingness to disclose

psychological symptoms from subjects than do female

physicians. Although the ANOVA detected a physician gender

main effect, it would be imprudent to come to this

conclusion without considering the interaction between

physician gender and subject gender and between physician

gender and level of expressive skill. Results indicated

that male subjects reported greater willingness to disclose

psychological symptoms to the male physician than to the

female physician. Male subjects also reported greater

reluctance to disclose psychological symptoms to a physician

of the opposite sex than did female subjects. Conversely,

female subjects did not report greater willingness to






disclose psychological symptoms based on physician gender.

However, there was not sufficient evidence to conclude that

same-sex dyads lead to greater willingness to disclose

psychological symptoms among females, there was a trend in

this direction. Hence, the main effect of physician gender

is best explained by the variance accounted for by the

greater reported willingness of male subjects to disclose

psychological symptoms to a male physician, and the greater

reluctance by male subjects to report these symptoms to a

female physician.

To summarize the results, female subjects did not

appear to be affected by physician gender, but male subjects

did. To date, there have been very few studies of

physician-patient interactions that attempted to compare the

patient care process or its outcomes for same sex versus

opposite sex physician-patient dyads. However, the existing

research suggests that the social status of the provider can

influence symptom reporting. Field studies of psychiatric

disorders have found that women consistently report more

symptoms than men (Tueting, Koslow, & Hirschfeld, 1981;

Clancy & Gove, 1974), and these sex differences have

generally been interpreted as a true sex difference in

symptoms rather than response bias (Gove, 1984).

Nevertheless, holding other variables constant, status

congruence between the physician and patient may influence

the reporting of symptoms by patients. There is evidence

that less social distance between an interviewer and a







respondent can increase rapport and reporting behavior

(Dohrenwend, Colombotos, & Dohrenwend, 1973). The results

of this study would support these previous findings. This

phenomenon, however, appears to reveal itself more clearly

in the case of male in contrast to female respondents.

Other research findings suggest that male compared to female

respondents report more symptoms to higher status

interviewers (Reissman, 1979). Hence, the perception of

"lesser social distance" perceived by male subjects who

considered a male physician interviewing a male patient,

coupled with their perception of the physician as being of

high status, may account for the male subjects' greater

willingness to disclose psychological symptoms to a male

physician. Conversely, males may have been more reluctant

to disclose psychological symptoms to a female physician

because of (a) the "greater social distance" between them by

virtue of being of the opposite sex and/or (b) perceiving a

female health care provider as being of lower status than a

male physician with comparable skills (Horman, Campbell, &

DeGregory, 1987).

Female subjects appeared to be unaffected by the impact

of physician gender on subject willingness or reluctance to

disclose psychological symptoms. Although other research

findings have suggested that same-sex dyads led to greater

disclosure among females (Highlen & Gillis, 1978; Young,

1979), it was not the case in this study. A number of

plausible explanations can be proposed for this finding.







There is evidence that higher status interviewers may

"threaten" female respondents and reduce their reporting

(Reissman, 1979). Reissman reported that in a population

based study in New York City, women respondents tended to

report more symptoms of functional disorder than did men and

also tended to report more symptoms when the interviewer was

a psychiatrist who did not identify himself as a physician,

as compared to interviewers who did identify themselves as

physicians. Hence, it is possible that, female subjects

were reluctant to report greater willingness to disclose to

either a male or female physician because of the perceived

"threat" of the higher status interviewer. These results

are consistent with other research findings which have

indicated that females are particularly resistant in

disclosing psychological symptoms to male physicians (Young,

1980).

A second explanation addresses the tendency of greater

reluctance by female subjects to disclose psychological

symptoms to male physicians. It is possible that the

"social distance" phenomena can account for some of the

reluctance. For instance, if women perceive male physicians

as being of particularly high status, this could make them

more "threatening" as recipients of disclosure. There is

also evidence that male physicians may discourage

communication and information exchange with female patients

(Wallen, Waitzkin, & Stoeckle, 1979). These researchers

found evidence suggesting that male physicians are more







likely to attribute psychological causes to the illness of

female patients than to those of men and to report being

more pessimistic in their prognosis for their female

patients as compared to their male patients. It is

therefore possible that some female subjects in this study

had past contact with male physicians, who had behaved as

Wallen and colleagues described, and these interactions

predisposed them to be more reluctant about disclosing

psychological symptoms to a male physician. This, coupled

with the female subjects' reluctance to disclose to higher

status individuals, or a perception of high social distance,

may explain why reluctance to disclose psychological

symptoms was slightly higher to male physicians than female

physicians.

In summary, these results are consistent with others'

findings that same-sex physician dyads may result in greater

symptom disclosure. However, these results suggest that the

same-sex dynamic exists more strongly among male dyads than

female dyads.

A final result of ANOVA I disconfirmed Hypothesis 4.

These findings indicated that, contrary to expectations,

when a male physician demonstrated high expressive skill, he

elicited greater psychological symptom disclosure from

subjects than did the female physician high in expressive

skill. One explanation of this finding may reflect the

influence of sex role stereotyping of physicians by both

male and female subjects. There is evidence suggesting that







male and female patients have different expectations of male

versus female physician behavior during the course of a

medical interview (Shapiro, McGrath, & Anderson, 1983;

Weisman & Teitelbaum, 1985). Shapiro and colleagues found

that male physicians are expected to be instrumentally

competent, while women physicians are expected to be

expressively skilled as well. Weisman and Teitelbaum claim

that:

sex-role socialization could result in patients
bringing traditional role expectations or
stereotypes to the encounter and responding to
these physicians based on those expectations.
Patients might expect female physicians to be
empathic and nurturant, and male physicians to be
less demonstrative and more directive. In
addition, owing to the historical fact that up
until recently the vast majority of physicians
were men, patients may identify the male physician
with the norm of 'affective neutrality'-i.e.
professional objectivity to preserve technical
judgement. [And] patients may expect and desire
that female physicians be more expressive and
humane. Thus both male and female patients might
respond to female physicians with greater self-
expression and symptom disclosure expecting
greater affectivity in response. (p. 1120, 1985)

The final ANOVA I result suggests that when a patient is

interviewed by a physician for the first time, the salience

placed on particular physician behaviors in either the

technical or expressive domain may vary according to the

physician's gender. Taking previous findings (Shapiro et

al., 1983) into account in reviewing the findings of this

study may assist in understanding the present results. It

may be that the male physician high in expressive competence

was found to be superior to the female physician high in







expressive competence in eliciting psychological symptom

disclosures because of the unexpected presence of a high

level of expressive competence in the male physician as

perceived by both male and female subjects. It is possible

that this unexpected perception served to increase the level

of trust that subjects had for the male physician since he

appeared more expressively competent than he was expected to

be. Consequently, this increase in level of trust may have

led to subjects reporting greater willingness to disclose

psychological symptoms. This unexpected and favorable

perception would not have impacted disclosure to the female

physician since some level of expressive skill was likely to

have been expected of her from the onset. Hence, this

"unexpected but favorable" dynamic may account for the

finding that male physicians high in the expressive

dimension elicited a greater degree of psychological symptom

disclosure from subjects than did female physicians high in

the expressive dimension.

Correlational analysis between (a) subject willingness

to disclose psychological symptoms and (b) the relative

importance of specific physician behaviors on making a

disclosure support a sex-role expectation explanation.

Results indicate that male subjects value behavioral cues

indicative of technical skill in female physicians to a

greater extent than do female subjects. When asked how

important particular physician behaviors were to their

decision to disclose psychological symptoms, male subjects







reported that "the thorough way the [female] physician

questioned the patient"-a technical competence behavioral

cue-was the most important behavior. A case could be made

for the argument that male subjects are less suspect of a

male physician's level of technical skill than they are of a

female physician's level of technical skill. Consistent

with this explanation, this study found that when a female

physician behaviorally demonstrates aptitude in a technical

skill behavior, this "unexpected but favorable" finding

impacted male subject willingness to disclose psychological

symptoms to a greater extent than female subjects'

willingness. Moreover, two other variables indicative of

the female physician's level of technical skill correlated

highly, with male subject willingness to disclose at least

one psychological symptom. These physician behaviors were

(a) the way the [female] physician informed the patient when

to expect discomfort, and (b) the "sum total" of technical

behavior ratings among male subjects who viewed a female

physician.

Sex role socialization may account for another finding.

Closer analysis of the low expressive dimension condition

indicates that female subjects are more sensitive to the

absence of physician expressive competence than are male

subjects. It is possible that the differential in sex role

socialization sensitizes females more than males to

expressive dimension characteristics in their interactions

with physicians. Hence, in the presence of physicians with







low expressive skills, female patients may be more

sensitized to the absence of expressive dimension behaviors

and therefore react differently than a male patient would.

Given that the domain of the expressive dimension is

consistent with traditional female sex-role

qualities/behaviors, it would not have been surprising to

find that the low expressive female physician (i.e., an

unexpected and unfavorable perception) elicited greater

reluctance to disclose psychological symptoms than the low

expressive male physician. However, this was not the case

as both male and female subjects reported approximately the

same degree of reluctance to disclose to the female

physician.


Analysis of Hypothesis Two (ANOVA II)

Results of ANOVA II indicate a main effect of level of

technical skill. Hence, physicians, with a high level of

technical skill and a moderate level of expressive skill

elicited greater willingness to disclose psychological

symptoms from subjects than physicians with low technical

skill and a moderate level of expressive skill.

These results suggest that when patients interact with

a physician in the context of a medical interview, they are

assessing the physician's level of technical competence. In

contexts where there is no evidence that the physician is

either very caring and feeling, or uncaring and unfeeling,

the physician's level of technical skill has a significant







effect on the degree of psychological symptom disclosure

elicited from the patient.

Research has indicated that as consumers of health care

services, patients expect quality care (Arrow, 1971). When

patients "go to the doctor" they are expecting to pay a

financial price in return for accurate diagnoses and

treatment. Hence, they are relying on the physician's

knowledge and experiential base to help them "get better."

In spite of patient's desire to be well provided for, they

are nevertheless reluctant to provide certain disclosures.

Research results from other studies suggest that disclosures

are more likely to be made to physicians who appear to have

the ability (technical skill) to help them "get better"

(Young, 1980).

The results of ANOVA II suggest that when patients

experience a choice dilemma (whether or not to disclose

information that could lead to an undesirable result) they

are faced with a risky decision (Antonovsky & Hartman,

1974). Research on trusting behavior has found that the

expectation of a preferred or positive outcome will increase

the likelihood of risk taking behavior (Tedeschi, Schlenker

& Bonoma, 1973). Hence, in a context where a patient must

decide whether or not to disclose information that is

"risky" (but may or may not be clinically relevant), the

patient's assessment of whether or not the physician could

do anything meaningful with this information could have an

effect on the likelihood of making the disclosure. The