.-u)ri1nm

|.a-u
I. v

In

H '1“

\Ir

n ‘ a. .-
rvyfu (

 

.n-u‘ - u

u
n'ul

 

 

 

 

'.
J
W
.4
L“

.
u
.
P

“I

'7
L

I
~
’I

J»
v

I
i
b
'5
.+
<
.
‘v

 

 

.
“L
n

‘n;

W. v ,

‘ .... ,
w k- .
SAW-2"”:
.:~v3;u..<r - '
_h 1., :. ,2?
«. 1 2‘

.IL.-..‘...$‘
" ‘33:? '

K . ‘d
u. . .
”mung”.-
432*"?1' ‘~

L~RA
n I “‘3

'01:.» ‘.

1w “rm-w WW?“

2: ‘z‘“-1*5'~T("’5—‘ “‘9 ‘
3:557”: 9 if .,-r ,Lé‘. .

.\ ., .;-.‘23' - -. ~ '4: 4.31‘
wt“. ‘ fl

may:

a.
a.
‘

ff;

tic,“ (:3 ‘7;

5-; ‘ -

a! “ -¢$T‘1£}':;:’.LJ¢
’45

Au
3: 30¢
25
.fi
.

3,,
1‘.-

w . ~\.
‘ '\=« "‘
2g . 7.:
.l

J3.“
:, €5.15
1,

     

. '1. .»
.~ «1‘

A,“

‘ ‘3. “$32; .
.393. a Mama . ,, -~ .
fiafi «W a; ‘ _"‘~‘:..'- «A - mg,- ”a
“452% :13. ‘ ‘31 «1’ .gr :4 ,fivgég‘ . ,H.
I»:

W52?

.~ .

. ~w

g'lt‘vx
.2...

3’

*3:
‘rv. . .
«Wm. ,z.
w"‘."‘-w‘ ‘ -
.-, "an.

n.

‘ ‘fi'.-.;_2¢
.{gin z fiffi‘:

n-
. 24'
4.

Q~ .‘~.. 4‘ \
u. . ..~
1 . JAE;
v n—
:-

'f—‘JA ‘ ‘-

1?
mm.“-

 

3-: .A E . , ~ , .- 1 m...
ENE: ”if?" _ 3 Es . _ Em .

. .r‘ ‘1“:
few-$27 , A

if:

“I

._'. ‘L‘

1?
.a-l
‘ ’A‘bn‘m}:

iHEEH.
‘ Areumvenm

\lmllfm

Illllunll l H» [Will

3 1293 01025 39

NM

 

l

 

This is to certify that the

thesis entitled

WW:
WWEAWW

presented by

mmmmie

has been accepted towards fulfillment
of the requirements for

 

 

degree in

em Qt .MMW

Major professor

Date32?~0’4

0-7 639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

4- Aififi - -

 

 

LIBRARY
Michigan State

 

University

 

 

PLACE II RETURN BOX to roman this checkout from your noord.

TO AVOID FINES return on or before data duo.

DATE DUE DATE DUE DATE DUE

 

 

I
JED 0 gm
. ig- "

 

 

2311 m3 .

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PHYSICIAN-PATIENT INTERACTION:
REDEFINING SATISFACTION AS A PROCESS VARIABLE

By

Karen M. Maduschke

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

MASTER OF ARTS

Department of Communication

1994

ABSTRACT

PHYSICIAN -P ATIENT INTERACTION:
REDEFINING SATISFACTION AS A PROCESS VARIABLE

By

Karen M. Maduschke

Previous research has identified communicative behaviors related to patient
satisfaction as an outcome of medical interviews. The identification of satisfaction as
a process which develops within interaction, however, would suggest that satisfaction
may be negotiated during the medical interview and thus have important implications
for physician training. By examining the medical interview and correlations between
patient and physician behaviors associated with satisfaction outcomes, we learn more
about how to encourage the development of satisfied responses during the interview,
as well as how to prevent dissatisfaction.

Interaction analysis was applied to 17 transcripts of physician-patient
interviews. The physician behavior of showing agreement and understanding was
strongly correlated with two patient variables associated with satisfaction: tension
release and volunteering information. Significant relationships were also found
between physicians eliciting questions and opinions from patients and patients
expressing commitment to discussed medical regimens. Implications of the findings,

limitations of the study, and directions for future research are discussed.

ACKNOWLEDGMENTS

In the writing of this thesis I have had the support of many people who deserve
more than my gratitude.

I owe a special debt to my advisor, Dr. Alicia Marshall, who introduced me to
this field of research. She provided me with many valuable comments on the
substance and form of this project, as well as with the encouragement to continue. I
would also like to thank my committee members for their contributions: Dr. William
Donohue, in whose seminar this project began, and Dr. Robert Smith.

I am grateful to Dr. Nancy Ainsworth-Vaughn and Dr. Debra Roter for
providing transcripts for use in this research.

The students who helped me with the mechanics of coding, entering and
proofing data have also earned my gratitude. Angela Birgy, Robert Bordeau, Karenn
Florante, Mark Friar, Maria Jekielek, Keith Lagerstrom, Tina Rajski, and Valerie
Tarquini were great fun to work with.

For providing me with the educational foundation to pursue a graduate degree,
as well as for their love and support throughout my studies, I am indebted to my
parents, Thomas and Carolyn Sager.

Finally, words cannot express my appreciation for the love, tolerance, and

assistance I received from my best friend and husband, Sven.

iii

TABLE OF CONTENTS

LISTOFTABLES vm
CHAPTER ONE ............................................... 1
INTRODUCTION ......................................... 1
LITERATURE REVIEW .................................... 2

The Significance of Satisfaction .......................... 2

Defining Satisfaction .................................. 3
Communicative Behaviors Associated with Satisfaction ..... 6

Communicative Behaviors Associated with Dissatisfaction . . 10

Additional Variables in the Satisfaction Equation ........ 13
HYPOTHESES .......................................... l4
Disease-Related Tension Release ......................... 15
Non-Disease Tension Statements ......................... 16
Question Asking .................................... 17
Repetitiveness ...................................... 18

iv

V

CHAPTER TWO .............................................. 20
METHODS ............................................. 20
Estimating Satisfaction ................................ 20

Interaction Analysis .................................. 21

Existing Coding Systems .............................. 21

A New Coding Scheme ............................... 24

Physician Behaviors ............................ 26

Patient Behaviors .............................. 27

PILOT STUDY .......................................... 3O
Transcript Preparation ................................ 3O

Coding Units ...................................... 31

DATA ANALYSIS ....................................... 33
CHAPTER THREE ............................................ 35
RESULTS ............................................. 35
Descriptive Statistics ................................. 35

Tension Release .................................... 36

Non-Disease Tension ................................. 37

Question Asking .................................... 38
Repetitiveness ................................. - ..... 39

Other Relationships .................................. 4O

vi

CHAPTER FOUR ............................................. 42
DISCUSSION ........................................... 42
Implications of the Findings ............................ 44
Disease-Related Tension Release ................... 44

Question Asking ............................... 45

Solidarity .................................... 47

Orientation Giving ............................. 48

Limitations and Directions for Future Research ............... 49
Methodological Issues ........................... 50

Potential Solutions ............................. 51

Conceptual Issues .............................. 52

Future Research ............................... 53

Pragmatic Relevance ................................. 54

Summary ......................................... 54
APPENDICES ................................................ 56
Appendix A: Characteristics of an ”Ideal” Physician ................ 56
Appendix B: Bales’ Interaction Process Analysis Categories .......... 57
Appendix C: Sources of Tension Statements ..................... 58

Appendix D: Roter’s Modified Interaction Process Analysis Categories . . . 59

vii

ENDNOTES ................................................. 60

LIST OF REFERENCES ........................................ 61

LIST OF TABLES

Table 1
Coding Categories for Assessment of Satisfaction ....................... 25

Table 2
Descriptive Statistics for Patient and Physician Behaviors .................. 36

Table 3
Correlation Matrix of Patient Tension Release
and Physician Expressions of A greement/Understanding and Solidarity ........ 37

Table 4
Correlation Matrix of Patient Tension, Question Asking and Summarizing
with Related Physician Behaviors .................................. 38

Table 5

Correlation Matrix of Patient Solidarity and Orientation
with Related Physician Behaviors .................................. 41

viii

CHAPTER ONE
INTRODUCTION

In 1898 Sir William Osler reportedly advised medical students ”Listen to the
patient, he is telling you the diagnosis” (Roter & Hall, 1987, p. 325). Despite this
age-old adage, many patients still complain that their doctors do not listen to them and
many continue to be dissatisfied with the care they receive from physicians. Though
most physicians entering the medical profession have the altruistic goal of healing
people, amazingly little attention is paid either during medical school or post- graduate
residency training to learning how to communicate effectively with those people one
intends to heal. As Sir William pointed out, communication is the link to understand-
ing a patient’s ailments, and likewise the key to alleviating a patient’s pain.

It is argued here that the role of communication in the development of
satisfaction during physician-patient interactions needs to be more closely examined in
order to improve the health care process. Specifically, there are a number of
communicative behaviors in the context of physician-patient interaction which are
linked to increased and decreased patient satisfaction. These behaviors and their
relationships to satisfaction as an outcome of the interview have been identified by
previous researchers. If, however, satisfaction can be identified as a process which
develops and can be recognized within interaction, it would suggest that satisfaction
may also be negotiated during the medical interview. This would have important
implications for physician training.

The verbal behaviors of patients which are known to indicate satisfaction level

and which could thus act as cues to physicians about the patient’s degree of

2

satisfaction have not been synthesized. Furthermore, there is little research which then
correlates those behaviors of the patient with behaviors of the physician. By
examining the medical interview, identifying patient behaviors associated with
satisfaction and dissatisfaction, and then examining the correlations of patient and
physician behaviors, we can learn more about how to encourage the development of
satisfaction during the interview (as well as how to prevent the development of
dissatisfaction).

This research attempts to make these connections in order to identify the role
of communication in the development of satisfaction. It first assesses the body of
literature regarding the nature of physician-patient communication related to
satisfaction. Next, four hypotheses arising out of the literature are presented. A
coding scheme for analyzing particular physician behaviors related to patient satisfac-
tion and relevant to the hypotheses is then introduced. Finally, the results of an
interaction analysis which applies the coding scheme to 17 physician-patient transcripts
is presented and discussed, along with limitations of the study and suggestions for
future research. Additionally, implications for the training of physicians are addressed.

LITERATURE REVIEW
The Significance of Satisfaction

Physician-patient interactions are of vital significance in the health care
process. Cleary and McNeil have pointed out that ”accurate and complete
communication between a physician and patient is often a necessary condition for the
provision of technical care” (1988, p. 25). While forming the basis of the physician-

patient relationship, the interactions also provide the most basic means for medical

3

diagnosis and treatment of disease, the management of illness, and the prevention of
many health problems (W asserman & Inui, 1983). Because of this link, researchers
have begun to investigate the relationships between aspects of the physician-patient
interaction and factors such as patient knowledge, compliance, and satisfaction.

Although research in doctor-patient interaction has been increasing over the last
30 years, there is still a demand for focused research which integrates tested and valid
methodology with issues of pressing concern. Arguably, patient satisfaction may be
first and foremost on the list of such concerns. It has been demonstrated that patient
satisfaction with their physicians is linked to a chain of variables such as compliance,
number of complications, and length of treatment, as well as known efficacy of
prescribed medication (Lane, 1983). These findings have, in part, led to the initiation
of psychosocial teaching programs in several medical residency programs in the
country (see Merkel, Margolis, & Smith, 1990; Smith, Osborn, Hoppe, Lyles, Van
Egeren, Henry, Sego, Alguire, & Stoffelrnayr, 1991). There is still a great deal of
research that needs to be done, however, to improve these programs and encourage the
continued development of similar ones.

Defining Satisfaction

Traditionally, the concept of satisfaction has been interpreted in a variety of
ways. Therefore, it is meaningful to first define its scope. Satisfaction is sometimes
viewed as a patient’s global assessment of the visit. Frequently it is related to some
specific physician behavior or characteristic such as knowledge and skill,
trustworthiness, or humaneness (Lane, 1983). In other instances, satisfaction has

included reactions to factors of the health care system such as waiting time and visit

4

costs, or appointment availability. For the purposes of this research, satisfaction is
conceptualized as a positive, ongoing aim response toward the physician (Lane,
1983; Roter, Hall, & Katz, 1987). This is a very broad definition, but necessarily so.
Because it is assessed during the medical encounter, satisfaction in this research is
largely judged as a function of how thoroughly, considerately, and humanely the
patient perceives s/he is treated by the primary care physician (Anderson, Fleming,
Aday, & Aguirre, 1979; Lane, 1983).

As a behavior variable, patient satisfaction can be viewed as either (1) an
outcome variable or (2) a process variable. Past research has almost exclusively
categorized it as a short-term outcome of the medical encounter; one that can be
assessed immediately after the appointment (Beckrnan, Kaplan, & Frankel, 1989;
Schofield & Amtson, 1989). Other examples of short-term outcomes are health or
disease related knowledge acquisition, tension release, and intention to comply with
medical advice (Beckrnan et al., 1989). Short-term outcomes are commonly used to
associate aspects of the medical interview with effects of the encounter. However, a
major limitation of short-term outcomes is that they are not very good predictors of
long-term effects (Beckrnan et al., 1989). In other words, we are not able to
determine in what way the short-term outcomes affect long-term attitudes or behaviors.
There is some evidence that although patients may express moderate satisfaction
immediately after the interview, they may not comply with medical advice in the long-
term (Beckrnan et al., 1989).

In this research, it is argued that satisfaction should be viewed as a process

outcome; one which occurs and can be assessed within the medical encounter itself.

5

Typically, factors such as degree of patient assertiveness, effectiveness of patient
information seeking, amount of patient opinion sharing, and physician empathy are
considered process outcomes. Beckman et a1. (1989) suggest that the improvement of
specific details of the interviewing process serves as a form of early intervention
linked to improvement in long-term outcomes. Furthermore, research on process
outcomes has two major objectives which are similar to those which should be taken
in the study of patient satisfaction during the medical encounter. According to
Beckman et a1. (1989) these objectives are: first, to recognize successful interaction
behavior and improve the quality of the interaction; and second, to demonstrate the
cause and effect relationship of longer term interventions.

The importance of patient satisfaction in the medical setting has been
demonstrated in numerous studies. Although linked to myriad variables including
perceptions of physician competence, overall effectiveness, and even abatement of
malpractice suits (Hall, Roter, & Rand, 1981; also see Donabedian, 1982; Ware,
Davies-Avery, & Stewert, 1977), satisfaction is perhaps most important to study
because of its relationship to compliance with medical advice. Satisfaction has been
linked to greater patient compliance, which reduces complications and the expense of
medical treatment (Lane, 1983; Stewert, 1984; Stone, 1979).

Identifying satisfaction at its root, as a patient first, almost subconsciously,
formulates a satisfied versus dissatisfied response to the encounter, is critical to
improving the quality of interaction and in establishing a causal link between '
satisfaction and longer term interventions. Determining satisfaction or dissatisfaction

upon completion of the medical visit is simply too late to improve the quality of the

6
interaction or to arbitrate long term behaviors. The following section outlines
behaviors which have previously been linked to satisfaction and which aid in the
examination of satisfaction as a process variable. ’
Qmmunigtive Bflvigg Associated with Satisfaction

Physician Communicative Behaviors. Several physician communication
behaviors are associated with patient satisfaction. In particular, research identifies that
behaviors by the physician which make the patient feel more informed about their
health as well as those which make the patient feel more comfortable or reassured
about their prognosis, are correlated with patient satisfaction. Moreover, any behaviors
which increase patient perceptions of physician warmth and openness may likewise
increase patient satisfaction.

In order to understand satisfaction as an ongoing, affective response formulated
during the process of the interview, it is important to begin by examining how patients
describe an effective doctor, one with whom they are generally satisfied. Feletti,
Firman, and Sanson-Fisher (1986) developed a questionnaire which assessed patient
perceptions of an ”ideal" doctor. Patients were asked, prior to a medical visit, to
select the most important characteristics of an ”ideal” doctor. After their visit, patients
were asked to rate their own physician on those same characteristics. The results of
the ”idea ” doctor ratings were then compared to scores on a satisfaction scale
completed by the patient. The data revealed that the ideal physician is one who
possesses a number of various characteristics, most of which relate to communication.
For example, the ideal physician is one who gives adequate time and attention to

patient problems, who provides valuable information, and who treats the patient as a

7

unique individual rather than ”yet another” medical complaint (see Appendix A for a
description of the characteristics). Feletti et a1. (1986) found that doctors possessing
these ideal features had more satisfied patients.

The findings of Feletti et a1. (1986) are consistent with the general conclusions
arrived at by Hall et a1. (1981) in a study of affective communication between patients
and physicians: patients judge the overall medical competence of physicians on the
basis of ”bedside manner.” According to Hall et a1. (1981), ”bedside manner” is
largely a measure of the physician's ability to communicate in a warm, sympathetic,
and personal manner. In particular, reassuring and positive words from the physician
were found to be associated with patient satisfaction.

Freemon, Negrete, Davis, and Korsch (1971) similarly found that outcomes of
medical consultations, such as patient satisfaction, were favorably influenced by a
number of physician behaviors. These included: (1) expressions of solidarity with the
patient (e.g., expressions of emotional support and trust in the patient); (2) the amount
of time spent discussing nonmedical or social subjects; and (3) an impression of the
physician offering information without the patient having to request it or feel
excessively questioned. Their data also revealed that simply ”being nice,” as measured
by the number of positive affect statements, was highly correlated with satisfaction
outcomes (Freemon et al., 1971).

A further finding by Freemon et a1. (1971) is that physician visits containing
some conversation of a general or nonmedical nature were associated with higher
levels of patient satisfaction. One draw back in this study, however, was that positive

affect and nonmedical conversation were often too difficult to distinguish from one

8

another. That is, it was difficult to determine if the expression of warmth and
friendliness by the doctor, regardless of subject matter (medical and non-medical), was
the cause of the greater satisfaction, or if the social conversation was the cause.

The study by Ley et a1. (1976) helps clarify the issue of time spent in medical
interviews relevant to satisfaction. In their research, a group of patients who received
extra interviews designed to increase their understanding of what they had been told
about their ilhiess showed significantly greater satisfaction than patients who received
placebo interviews about their hospital stay. Thus, additional, or increased
communication with the physician, in and of itself, is not sufficient to increase
satisfaction; rather, it is important that the nature of the communication either: a) help
the patient understand her/his illness better; or b) show friendliness and solidarity on
the part of the physician, while helping relieve the patient’s tension by engaging
her/him in social conversation.

Question asking is another form of physician cormnunication which is
correlated with patient satisfaction. High levels of patient satisfaction have been
related to physicians asking open-ended questions, clarifying questions, empathetic
questions, and encouraging questions, and by physicians eliciting the patient’s concerns
and expectations (Comstock, Hooper, Goodwin, & Goodwin, 1982; Woolley, Kane,
Hughes, & Wright, 1978). Moreover, Rowland-Morin and Carroll (1990) suggest that
a high rate of physician questioning may facilitate information transfer. The transfer
of infomation may serve two functions relevant to patient satisfaction. One, it may
lead to increased patient perceptions of physician involvement and concern. Second,

by increasing the time spent with the patient, it may lead to perceptions of the

9
physician giving appropriate time and attention to the patient’s problem. Both of these
are physician behaviors consistent with factors identified by Feletti et a1. (1986) as
characteristics of an ideal physician (see Appendix A). Thus, on the basis of this
relationship, physician question asking should be likely to increase patient satisfaction.

Patient Communicative Behaviors. While physicians may communicate in
ways directly related to patient satisfaction, patients themselves exhibit communicative
behaviors which may indicate satisfaction. One communication behavior of patients
shown to be positively correlated with satisfaction is identified as giving ”orientation”
(see Bales’ Interaction Process Analysis categories, Appendix B) (Carter et al., 1982).
Giving orientation is characterized by verbal behaviors which provide extended detail,
such as lengthy story telling, whereby the patient extensively relates her/his
experiences in her/his own words. Research has shown that the opportunity to engage
in this type of expressive behavior seems to encourage patient perceptions of
satisfaction with their medical encounters (Stiles, Putnam, Wolf, & James, 1979). It
may be that this opportunity helps the patient release tension, or perhaps it
demonstrates to the patient that the physician is willing to spend time and listen to
her/his concerns. In either event, it increases the patient’s level of satisfaction and is
worthy of further investigation.

The opportunity for patients to relieve tension is another communication
variable associated with satisfaction. In their research on outcome-based doctor-
patient interactions, Carter, Inui, Kukull, and Haigh (1982) coded six different sources
of patient tension statements (see Appendix C) and examined relationships between the

various kinds of tension and patient satisfaction. To do this, they counted how many

10

patient tension statements in each of the six categories were made by the most
satisfied patients, and how many in each category were made by the least satisfied
patients. One category of tension, labeled ”disease-related tension" was found to be
associated with more satisfied patients. They described disease-related tension as
utterances regarding the patient’s pain or restriction of activities. It seems that the
opportunity to share this type of information is a source of tension-release for patients
and aids in the development of satisfaction. Patient communication regarding other
sources of tension are specifically related to dissatisfaction and are discussed in a later
section.

anmrum'cative Behaviors Associated with Dissatisfaction

Physician Communicative Behaviors. In many cases, physician behaviors
which were associated with satisfaction are the same as those associated with
dissatisfaction. That is, while certain verbal behaviors are related to satisfaction, the
absence of those behaviors, or, at time, the opposite behaviors, are linked to
dissatisfaction. However, research does indicate that at least two variables are
distinctly related to dissatisfaction. Ironically, these are both variable also associated
with satisfaction: time and question asking.

Just as the amount of time spent with the patient was an important variable in
satisfaction, it appears to be significant in the development of patient dissatisfaction.
It was reported earlier that the time the doctor spent discussing the cause of an illness
was positively correlated with satisfaction. But, Freemon et a1. (1971) also found that
the time the doctor devoted specifically to history taking was negatively associated

with satisfaction. According to their research, patients seemed to become frustrated

1 1
and dissatisfied by lengthy discussions about their medical history. Freemon et a1.

(1971) hypothesize that this negative relationship between history taking time and
satisfaction is a function of patients’ perceiving longer history taking as ineffective
communication.

Although Freemon et a1. (1971) do not specifically identify what verbal
behaviors might be perceived as ineffective, one might expect that the excessive
question asking which commonly occurs in this period of the medical encounter
contributes to such perceptions. This would be true to the extent that patients interpret
the question asking as a ”not dealing with” the immediate issues or concerns that the
patients has. That is, they may not see any relevance in the questions and would
prefer to be told immediately what causes their pain or discomfort and how to treat it.

Patient Communicative Behaviors. Based on the literature, at least three
patient communication behaviors seem to indicate dissatisfaction. Specifically,
research shows positive associations between dissatisfaction and the variables of
patient tension statements, repetitive behaviors, and question asking. Each of these are
discussed below.

Findings regarding patients’ verbal indicators of dissatisfaction are found in the
study by Carter et a1. (1982). In their research on doctor-patient interaction, these
researchers observe that ”Tense verbal behaviors of patients were negatively associated
to satisfaction with the encounter" (p. 565). As described earlier, Carter et a1. (1982)
coded six categories of patient tension statements. Examples of tense behaviors.
include complaints about treatment, or other aspects of the medical encounter, such as

appointment availability or length of time waiting to be seen by the doctor (see

12

Appendix C). In each of the categories except disease-related tension, there were
significantly more tension statements made by the least satisfied patients than by the
satisfied patients.

Freemon et al. (1971) have similar conclusions regarding dissatisfaction. With
an index measuring the proportion of negative affect (as expressed in statements of
disagreement, tension, and antagonism) to the total amount of expressed affect, these
researchers illustrate an inverse relationship between negative affect and both
satisfaction and compliance. That is, the more a patient expressed negative emotion,
the less satisfied the patient was with the interaction and the less compliant the patient
was with prescribed medical treatments. These findings appear to be consistent with
those in the research by Korsch, Gozzi, and Francis (1968), who found that the
physician’s lack of warmth was related to the patient’s dissatisfaction. It may be that
the physician’s lack of warmth is a source of tension for the patient, creating the
negative affect which is associated with dissatisfaction.

Patient communication which is repetitive in nature has also been linked to
dissatisfaction. Carter et a1. (1982) found that repetitive patient disclosure in the
introduction and history taldng segment of the visit was negatively associated with
subsequent satisfaction scores. The conclusions they draw regarding this relationship
are that patients who perceive their doctors do not understand the severity of their
symptoms feel the need to reiterate the extent and degree of their pain. Therefore,
patients who repeatedly make statements regarding their condition are likely to “feel
that the doctor does not understand them, and consequently, these patients will be less

satisfied with the encounter.

13

An inconclusive finding of the research by Freemon et a1. (1971) may hold an
interesting clue regarding patient dissatisfaction. An expected, positive relationship
between the opportunity to ask many questions of the physician and higher levels of
satisfaction was not confirmed. Freemon et a1. (1971) suggest that patients prefer a
doctor who freely offers information, without the patient needing to extract the desired
information through numerous questions. Research by Cartwright (1964) confirms
this, concluding that patients who said they had to ask for more information expressed
greater dissatisfaction than those who said they were freely given all the information
they required. As mentioned earlier, Comstock et a1. ( 1982) and Freemon et al. (1971)
similarly found that physicians freely offering information to the patient was positively
correlated with high levels of patient satisfaction. Thus, if this is true, the number of
questions a patient asks during the medical visit would indicate degree of satisfaction.
Additional Variables in the Satisfaction Euation

Thus far, the research has identified a number of communicative behaviors of
physicians and patients which are linked in some way to patient satisfaction. Research
has shown, however, that these associations with satisfaction or dissatisfaction may
only be valid during particular segments of the interview. In other words, the timing
of physician and patient behaviors plays a significant role in their relationship to
patient satisfaction.

Some researchers suggest that the medical encounter can be divided into
different sections with varying objectives and processes (Carter et al., 1982; Stiles et
al., 1979). In formulating their research, Carter et a1. (1982) separate the medical

encounter into three, independent segments: (1) an introduction-history; (2) a physical

l4

examination; and (3) a conclusion. They argue that the timing of behaviors within and
across these segments is critical to subsequent outcomes. In particular, Carter et al.
(1982) cite an example dealing with patient question asking. For instance, if a patient
requests medication early in the encounter, this behavior has been positively related to
subsequent satisfaction. However, requests for medication which are made in the
concluding segment have been negatively associated with reports of satisfaction.

Further evidence for the role of appropriate timing is found in the research by
Freemon et al. (1971) who indicated that patient satisfaction was correlated with the
amount of time the physician spent with the patient, while the length of time
specifically spent in the history taking segment was correlated with dissatisfaction. It
was suggested that the difference is attributable to perceived effectiveness of
communication. Questions by the physician, much like some questions from the
patient, need to be delivered at the appropriate time. Because of this relationship
between the timing of behaviors and outcomes such as satisfaction, it is important that
research continue to evaluate when specific behaviors occur, thus figuring the timing
of behaviors into the satisfaction equation.

HYPOTHESES

The literature confirms that a number of patient communicative behaviors are
associated with satisfaction. There is also evidence supporting the relationship
between physician communication and satisfaction. However, in order to understand
how physicians may affect the process of satisfaction development, it is essential that
physician behaviors which are correlated with patient satisfaction and dissatisfaction

behaviors can be identified. Thus, four hypotheses regarding the relationship between

15

physician and patient communication behaviors have been developed. The hypotheses
focus on four specific patient behaviors: (1) disease-related tension release; (2) non-
disease related tension; (3) question asking; and (4) repetitiveness. Each of the
hypothesized relationships is addressed separately below.

Disease-Related Tension Release

Expressions of tension during the medical encounter may have different
meanings. Carter et al. (1982) found that patient expressions of tension generally bear
strong negative relationships to patient satisfaction, while patient expressions of
tension rel+ease are positively related to satisfaction. Tension release is operationalized
as statements of disease-related tension, whereby the patient ”unveils” their concerns to
the physician about their medical or socioemotional condition. One would expect that
that, in order to lead to higher levels of satisfaction for the patients, these behaviors
must be met or encouraged by statements from the physician which demonstrate
concern, empathy or support. Such communicative behaviors of the physician might
be characterized as those which show agreement or understanding or which express
solidarity with the patient. An absence of such physician behaviors would be expected
to result in either the patient not engaging in (or quickly discontinuing) this type of
disclosure, or in the development of dissatisfaction.

Returning to the arguments made by Carter et al. (1982), it is anticipated that
this would only be true during the history taking segment, which provides the
appropriate outlet for tension-release. Later in the interview, such behaviors may be
interpreted more as repetitive (which will be discussed further in a later section).

Therefore, it is hypothesized that:

16
H1, Utterances of tension release made by the patient during history
taking are positively correlated with the physician showing

agreement/understanding during the same segment.

H", Utterances of tension release made by the patient during history
taking are positively correlated with the physician expressing solidarity

during the same segment.

Non-Disease Tension Statements

Non-disease related tension is characterized as verbal behaviors which express
discontent with aspects of the medical visit not directly associated with the patient’s
medical or socioemotional condition (Carter et al., 1982). This may include
complaints regarding the length of time the patient waited before being seen by the
physician, or difficulty in finding transportation or obtaining the appointment.
Expressions of these types of tension, regardless of when they occurred, were found to
be made more frequently by the least satisfied patients. It is hypothesized that these
statements continue when the physician does not express appropriate levels of
agreement and understanding or solidarity with the patient. Thus, it is specifically

hypothesized that:

H2, Non—disease related tension utterances made by the patient
throughout the interview are negatively correlated with physician

statements of agreement/understanding

17
H25 Non-disease related tension utterances made by the patient
throughout the interview are negatively correlated with physician

expressions of solidarity.

Question Asking

Freemon et al. (1971) found that patients more favorably rate a physician who
freely offers information than one who requires the patient to extract the needed
information through extensive question asking. Research by Cartwright (1964)
confirms that patients who said they had to ask for more information expressed higher
levels of dissatisfaction than those who said they were offered all the information they
required. Thus, the number of questions a patient asks during the medical visit would
be an indicator of level of satisfaction. In particular, the more questions that a patient
asks, the less satisfied the patient is.

One would expect patient behaviors which are associated with dissatisfaction to
be negatively correlated with physician behaviors which are associated with
satisfaction. Furthermore, patient questioning behaviors should be negatively
correlated with physician behaviors intended to increase understanding and reduce the
need for questions. Several researcher identified physician question asking as one
important behavior which is correlated with patient satisfaction (Comstock et al., 1982;
Woolley, et al., 197 8). Additionally, one might expect that when a physician
summarizes her/himself, that it would serve to clarify and synthesize information,
thereby decreasing the need for patient questions. These conclusions lead to the third

hypothesis:

18

H3. The frequency of the patient asking questions is negatively

correlated with physician questioning behaviors

Ha, The frequency of the patient asking questions is negatively

correlated with the physician summarizing her\himself.

Repetitiveness

The research on repetitive behaviors indicates that the more patients repeat
themselves, the less satisfied they are with the physician (Carter et al., 1982).
Repetitiveness is considered to be statements made by the patient which repeat earlier
expressions or which summarize in new words what the patient has already said.
Carter et al. concluded that behavior is considered necessary by the patient when the
patient is under the impression that the physician is not listening or responding to what
s/he is saying. Therefore, such behaviors should be negatively correlated with the
physician demonstrating understanding either by summarizing the patient or by directly
showing agreement or understanding. These behaviors would serve to demonstrate
that the physician is listening, is attuned to the patient’s concerns, and empathizes.

This the following is hypothesized:

H4. Repetitive utterances made by the patient are negatively correlated

with the physician summarizing the other.

19
H“, Repetitive utterances made by the patient are negatively correlated

with the physician expressing signs of a greement/understanding.

CHAPTER TWO
METHODS
Estimating Satisfaction

Finding estimates of patient satisfaction with various aspects of health-care
delivery has been the impetus of many studies. However, the vast majority of these
studies measure satisfaction through the use of self-report questionnaires such as the
Medical Interview Satisfaction Scale developed by Wolf, Putnam, James, and Stiles
(1978; in Rowland-Morin & Carroll, 1990) or a modification of the Interpersonal 1'
Communication Satisfaction Scale (Lane, 1983). Although Roter et a1. (1987) used a
third party observation measure of satisfaction, and others have used personal
interviews (Stewart, 1984), there have been almost no efforts made to measure
satisfaction in any way other than through personal reports. Such reports are
frequently subject to a variety of reliability and validity threats including self-selection
bias (i.e., patients who respond to questionnaires tend to be more satisfied; Ley,
Bradshaw, Kincey, & Atherton, 1976) and subject apprehension to admit
dissatisfaction (Oksenberg & Cannell, 1977). Recognizing that there are serious
limitations to this method, Freemon et al. (1971) have specifically appealed for more
recorded interaction and direct analysis.

In this study interaction analysis is used in attempts to answer that appeal; to
develop more insight into patient satisfaction with the physician in the context of
interaction. It focuses on patients and their consultations with new physicians. While
the literature review identified physician and patient cornrrrunication behaviors related

to either satisfaction or dissatisfaction as an outcome, the proposed research seeks to

20

21

identify correlations between specific physician and patient verbal behaviors which can
help determine how satisfaction develops in the process of the medical interview.
Interaction Analysis

Interaction analysis (IA) is a process of coding and quantifying qualitative data
in order to allow for sophisticated and systematic statistical analysis (Wasserman &
Inui, 1983). Several distinct IA coding systems have been developed within the last
thirty-five years to assess aspects of dyadic or small group interaction. Among these,
the Bales’ Interaction Process Analysis (IPA) and Roter’s modification of the Bales’
system (Modified IPA) have been the most widely adopted in medical interview
research. Each of these has been very useful for the research to which they have been
previously applied, however, neither is completely adequate for the proposed study.
Therefore, a new scheme has been developed which draws upon the significant
contributions of each of these, plus one additional system which has not been
previously applied to research in the medical setting. The following section will first
review the relative merits and weakness of each of these systems and then present the
new coding system designed specifically for examining the four hypotheses of this
research.

Existing Coding Systems

Wasserman and Inui (1983) present a systematic review and critique of IA
approaches, highlighting the relative strengths and weaknesses of each. In their review
these authors call for interaction research which has, as its ultimate goal, modifying

the interactions of doctor-patient dyads, and thereby affecting the health care process.

 

22
Their evaluation of each of the analysis systems is geared to assessing the utility of
each system to that end.

The Bales system was developed in 1950 specifically ”for the study of the
social and emotional behavior of individuals in small groups” (see Appendix B)
(Wasserman et al., 1983). This system, which has been shown to have strong validity
and reliability, is particularly well suited for the study of relationships. The instrument
is sensitive to the feelings of the interactants towards one-another, since it concentrates
on statements of affect. Furthermore, the system has a strong potential for analyzing
the interaction sequence, and is applicable to many diverse situations. The primary
weakness of Bales’ IPA is that each utterance can only be coded into one category.
Unfortunately, there are many instances of statements which transfer both content
information, and relational information. These would represent distinct categories with
the Bales method. Wasserman et al. (1983) cite the example of ”Doctor, am I going
to die?” It is difficult to determine whether this should be coded as ”asks for
opinion,” or as ”shows tension;” it appears to do both simultaneously.

A second shortcoming in the Bales system is a limited selection of neutral
classifications appropriate to physician-patient interactions. The choices for a
physician's eXpression of neutrality include only: ”su g gestion,” ”opinion,” and
”orientation.” Modifications of the system by Freemon et al. (1971) and Roter (197 7)
have provided some solutions to this problem (W asserman et al., 1983). For instance,
Freemon et al. (1971) changed ”suggestion" to ”instructions” and added subcategories

for simple acknowledgements such as ”Yes, I see.”

.1"

23

Roter’s adaptation of IPA provides two unique sets of categories; one for the
physician’s communication and one for the patient’s (W asserman & Inui, 1983; see
Appendix D). Although the two sets of categories are similar, the classifications are
altered to more appropriately fit the roles of each interactant. These modifications are
helpful in applying interaction analysis specifically to the health care setting, however,
the modified system is subject to many of the same criticisms as the original Bales
model. Both methods suffer, additionally, from an inability to specify the
informational content of the interactions, thus prohibiting any contextual references in
the analysis. Wasserman et a1. (1983) note that this is potentially dangerous, since the
entire course of the interaction and the outcomes are influenced tremendously by the
subject of the conversation.

Furthermore, Roter’s modified IPA does not have a category for patient asking
questions, other than requests specifically for medication or for a clarification. Given
the relationship between the frequency of questioning by patients and their satisfaction,
it is imperative that all questioning behavior be included in any coding scheme meant
for the analysis of satisfaction.

The Carter et a1. (1982) study to identify effective provider and patient
behavior also compares and contrasts the advantages of these two systems. They
measured the ability of the Bales and Roter systems (among others) to predict various
outcomes (knowledge, compliance and satisfaction) in each of the three interaction
time periods (history-physical-conclusion). The results of their research indicated that
the Bales and Roter systems both have moderate explanatory power for the dependent

variable of satisfaction.

 

24

Based on the studies by Wasserman et al. (1983) and Carter et al. (1982), the
Bales’ Interaction Process Analysis and Roter’s modified IPA both appear to be useful
for investigating the role of affective communication. However, these systems are still
incomplete. By looking outside of the medical literature, one can expand the range of
possible coding schemes significantly. For instance, Gottman (1979) has developed a
procedure for investigating marital interactions called the Couples Interaction Scoring
System (CIS S). This system has additional content categories not previously
considered. Some of these are categories such as: (1) summarizing oneself; (2)
summarizing the other person; and (3) diverting the topic (”communication talk”).
Each of these are as relevant to a medical interview as to any other dyadic encounter
and are not directly addressed by the other two systems.

A New Coding Scheme

The following system is a specially adapted translation of these three coding
procedures which has been devised for coding physician and patient verbal
communication behaviors. The scheme helps to identify the correlations between
patient behaviors associated with satisfaction and dissatisfaction and related physician
behaviors (see Table l). The specific purpose of this scheme is to identify
correlations between physician and patient behaviors associated with the development
of satisfaction. As such, each item was selected to represent relevant behaviors
previously described in the literature.

While the coding scheme is not intended to represent an ordinal scale, the
behavior variables are generally arranged on a continuum of patient satisfaction. That

is, lower numbered variables are less frequently correlated with patient satisfaction (or

may be correlated with dissatisfaction), while variables with higher numbers tend to be

more strongly correlated with satisfaction. For instance, question asking by the

patient, when observed to occur excessively, is considered to be negatively correlated

with satisfaction (or more highly correlated with dissatisfaction), whereas frequent

expressions of solidarity are positively correlated with patient satisfaction. The same

is true with the physician behaviors. The more behaviors at the higher end of the

continuum occur, the more frequently they demonstrate a positive correlation with

patient satisfaction. The coding categories will be used to test each of the

hypothesized relationships.

Table 1

Coding Categories for Assessment of Satisfaction

Patient Behaviors

l
2
3
4
5
6
7
8
9

10
11

12

asks questions

shows disagreement or criticism
summarizes/repeats self
expresses non-disease tension
expresses disease tension
requests opinions/suggestions
summarizes other

personal remarks or digressions
responds directly to question
gives voluntary orientation

shows agreement/understanding

offers solidarity

Physician Behaviors

1

\DOOQO'ttlt-fi-UJN

Hp...
HO

p—s
N

shows disagreement or criticism
gives direction/instruction
summarizes/repeats self
personal remarks or digressions
gives information or opinion
indicates listening

summarizes other

requests questions

requests opinions/su g gestions
asks questions

shows agreement/understanding

shows solidarity

 

 

Ell

26
Physician Behaviors

Twelve categories of physician communication behaviors were developed based
upon previous research. The categories are designed to be inclusive of a vast array of
possible behaviors, while specifically addressing variables related to satisfaction. The
first physician behavior listed in the coding scheme is ”shows disagreement.” This is
defined as an instance of the physician directly contradicting the patient, or responding
negatively to an assumption or concern of the patient. The next category of physician
behavior is ” gives direction/instructions” and includes things such as ”you may put
your shirt on now” or ”take one of these tablets every four hours with a glass of
milk." Any request for the patient to alter her/his behavior would fall under ”gives
direction or instruction.” Disagreement and directions/instructions fall relatively low
on the satisfaction continuum because they are associated with perceptions of the
physician as dominant and distant.

The ”summarizes/repeats self” code represents any attempts on the part of the
physician to rephrase statements that s/he previously made. Greetings, introductions,
and other comments not related to the medical encounter are categorized as ”personal
remarks or digressions.” This latter category usually represents neutral comments, but
may also be considered positive, if social conversation or positive affect is
communicated.

Listening behaviors of the physician are generally very short, neutral
comments, such as ”uh=huh,” ”okay,” or ”I see.” These are indicators that the
physician has heard the patient and will often precede other statements by the

physician, or interrupt the patient during longer turns at speaking. Attempts to

27
summarize, paraphrase, repeat or confirm what the patient has previously expressed is
counted under the ”summarizes other” category. This is seen as relatively positive
behavior, since it demonstrates active listening and an attempt to fully understand the
patient.

Requesting questions and requesting opinions or suggestions from the patient
are also very positive behaviors. Such behavior demonstrates that the physician is
concerned about the patient’s ability to understand the communication and that the
physician feels that opinions and suggestions of the patient are valuable. As such, it
would be considered quite positive. A request for questions or opinions is generally
very direct and easily recognized, such as ”Do you have any questions?" or ”How long
do you think you'll need to quit smoking?”

One of the most positive behaviors a physician may exhibit is the offering of
solidarity or commitment. This is a behavior which demonstrates confidence in the
patient’s ability to overcome their challenges, raises the patient’s status, or offers help,
reassurance or rewards to the patient. It shows support for the patient and lets the
patient know that the physician is ”in it together” with the patient. An example of
such as statement would be: ”I know you can stop smoking. Keep in touch with me
when it gets tough and we’ll get through this."

Pag'ent Behaviors

The codes for patient communicative behaviors also address a wide option. of
possible utterances. As with physicians, there are twelve distinct categories. One of
the less positive of the categories is ”ask questions.” This category encompasses a

wide range of questions but excludes direct requests for physician opinions or

28
suggestions. Therefore, ”May I get dressed now?” would be included, but ”What do
you think I should do?” is not. The latter example would be coded as ”request for
opinions/suggestions.” Question asking is considered low on the satisfaction
continuum because of the expectation that physicians should provide the desired
information without prodding from the patient.

The next code for patient behaviors is ”shows disagreement or criticism.” This
category incorporates utterances from patients which indicate a conflict with the
physician. This would include statements such as ”I don’t think the patch will do any
good for my smoking problem.”

Several patient behaviors on the scale are identical to ones on the physician
scale, such as ”surnmarize self” and ”personal remarks/digressions.” The most unique
statements on the patient scale concern the expression of tension. Two types of
tension statements are distinguished: those which are directly related to the patient’s
physical or socioemotional state (termed disease tension); and those which are not
associated with a physical or socioemotional condition (non-disease tension). As
discussed earlier, the expression of disease tension can be a positive behavior when it
occurs at the appropriate time, since it allows patients to share concerns and relieve
themselves of burdensome worries. N on—disease related tension, hoWever, tends to be
negative, since it is less seldom associated with positive changes which may take place
during the medical visit, and more often represents unproductive complaints.

There are also two categories for patients to provide information. The first is a

direct response to a question posed by the physician. The second category pertains to

 

29

information which is offered by the patient beyond the physician’s question. For
instance, consider the following exchange:

Physician: ”Where does it hurt?”

Patient: ”The whole knee hurts. The pain started under the knee cap

two weeks ago during a basketball game, but spread and got worse in

the last couple of days.”
The patient’s first sentence ”The whole knee hurts.” would be an example of
responding directly to a question (9). The second sentence elaborates beyond the
question, providing additional information which the patient has volunteered. Thus,
the second sentence would be coded as giving voluntary orientation (10). Voluntary
orientation must be distinguished from direct responses to questions because of the
relationship between orientation giving and satisfaction identified by Carter et al.
(1982). Orientation giving is a highly positive patient behavior that aids the physician
in diagnosis and which is associated with increased levels of patient satisfaction.

The most positive of patient behaviors is the same as the most positive of
physician behaviors: ”shows solidarity.” Such statements indicate that the two
interactants are prepared to work together to ensure the health of the patient.
Examples of the patient expressing solidarity would include "I'm sure I’ll be able to

quit smoking with this new patch.”

3O
PILOT STUDY

The research was executed in two stages: a pilot study for confirmation of the
coding scheme and reliability testing, followed by a full scale research endeavor to
address the four hypotheses. At the time of the pilot study, the coding scheme was
not yet fully developed. Rather, a preliminary scheme had been devised, which
underwent some modifications as a result of the pilot study.

Two undergraduate research assistants were originally trained in the
preliminary coding scheme. An intercoder agreement ratio of .91 was achieved during
training, before the coders began working with the actual transcripts. The assistants
then applied the coding scheme to six transcripts in order to test the scheme for
appropriate focus and scope, and work out any potential difficulties. Although the
assistants were provided with background information about physician-patient
communication issues and interaction analysis, they remained blind to the hypotheses.

Transcript Preparation

The six test transcripts‘ were each full-length interviews, ranging in length
from 98 to 260 turns. In the transcripts, names of the physicians were changed and
any personal names of patients or their friends and families were blacked out to ensure
anonymity. Demographic information for the physicians and patients in the test
transcripts is unavailable.

In order to analyze where important behaviors such as tension release occur,
each transcript was divided into the three sections proposed by Carter et al. (1982):
(l) introduction-history taking; (2) physical examination; and (3) conclusion. The first

and last segments of the encounter are defmed by their exclusion from the physical

31

examination section. The first section, introduction-history, begins with the very first
verbal message, and ends when the physical exam begins. The physical examination
section is then defined as beginning with the first verbal event that suggests an
examination activity (such as instructions for undressing, or assuming a specific
position or a statement such as ”Let’s take a look at...”). The physical examination
section ends with the last verbal message indicating that the examination has
concluded (such as ”You may get dressed now”). The conclusion segment begins
when the examination ends, and continues until the last verbal message is delivered
(Carter et al., 1982).
Coding Units

The primary unit of analysis in the study was a single person’s turn. A turn is
defined as one person’s uninterrupted speaking time (Hopper, Koch, & Mandelbaum,
1985). Occasionally, when it was evident that clearly different behaviors were
expressed in the course of a single tum, then the turn was divided into independent
thought units. Only when the code of a thought unit changed from a previous one
within the same turn was it coded differently. It is possible that multiple thought units
could be used to represent the same coded behavior. To avoid tremendous
redundancy, multiple thought units representing the same behavior were not coded
separately.

Each transcript was coded for characteristics such as the physician
identification, the total number of turns in the transcript, the total number of questions
asked by the physician and the total number of questions asked by the patient. Each

individual turn was then coded for the segment of the transcript in which the turn

 

32

occurred (introduction-history, examination, conclusion), then for the speaker
(physician, patient, third party), and finally for the behavior the turn represents
according to the newly developed coding scheme (such as ”asks a question” or
”summarizes the other”). Multiple thought units representing different behaviors
within a single turn were also coded.

In coding the behaviors of the physicians and patients, coders were instructed

to carefully examine the preceding utterances, in order to distinguish subtle differences

 

between categories such as ”shows agreement or understanding” and ”shows
solidarity.” A physician’s thought unit was only coded as showing solidarity when it
was an independently offered gesture, and not a response to a previous comment by
the patient (unless the patient was expressing disease-related tension). The same is
true for statements made by patients; differences between offering commitment and
showing agreement or understanding could be detected through the careful scrutiny of
previous statements by the physician.

Further clarifications of the coding scheme included the ”priority system” by
which utterances were coded. Any utterance which seemed to fall into more than one
category, was coded into the category with the most narrow scope. For instance, if a
patient were to say: ”How do you think I should go about reducing my weight?” the
utterance would be coded as ’requests opinion/suggestion’ as opposed to simply ’asks
questions.’ Likewise giving ’direction or instruction’ is a more narrow category than

’gives information or opinion.’

33
DATA ANALYSIS

After the original coding process was completed and the data from the pilot
test entered, the coding scheme was carefully reviewed and several adjustments were
made to the coding categories. These adjustments resulted in the revised coding
scheme presented in Table 1. The primary change was the addition of the ”listening”
category for physicians and the ”responds directly to questions” category for patients.
These refinements were necessary because the original categories were sometimes too
broad. For instance, most ”listening behaviors” had previously been subsumed under
the ”shows agreement/understanding” category. However, expressions such as ”yeah”
and ”okay” made by the physician after every utterance by the patient are not
necessarily expressions of agreement or understanding, but rather are mere indicators
of listening or hearing.

For the patient, it was necessary to separate out the type of information giving
which was voluntary, from that which directly answered a physician’s questions.
Thus, another distinction in the coding scheme, would be the difference between
”responds directly to a question” (9) and ” gives voluntary orientation” (10) on the part
of the patient. "Gives voluntary orientation” is any information offered by the patient
which is not directly solicited by the doctor. In many cases, a patient will provide
more information than is necessary to answer a question. When this is the case, the
portion of the answer which is necessary to respond to the question is coded as a ”9”
and the remaining portion is coded as a ”10.”

Once the coding scheme was fully developed, it was applied to seventeen

transcripts of physician-patient interaction? All the transcripts represented either first

34
time or relatively early visits between the particular patient and physician. Only first
time and very early visit transcripts were used because once a relationship develops
between a physician and patient, it is likely that the dynamic of verbal interaction
changes substantially, and more frequently non-verbal cues are relied upon.
Of the seventeen transcripts, 10 represented same sex dyads, while seven were
mixed sex dyads. Of the same sex dyads, four were male physicians and patients,

while six were female physicians and patients. Two transcripts had male physicians

 

with female patients, and five had female physicians with male patients. Transcripts
ranged in length from 86 to 437 independent thought units, with a mean of 231
thought units (e.g., representing interviews ranging from approximately 11 minutes to
approximately 40 minutes of speaking time).

For these analyses, four new undergraduate coders were trained, with an
intercoder reliability for coding agreement .90 being achieved. Again, these coders
remained blind to the hypotheses, but were given basic instruction in interaction
analysis. The transcripts were prepared in the same manner as in the pilot study;
divided into three segments of interview time. Each of the hypotheses was tested then
using a Pearson product-moment correlation (Pearson-r). Two correlation matrixes
were originally created. The first one was for the patient and physician behaviors
exhibited during the introduction-history segment only. This was used specifically to
test the first hypothesis where the timing of certain behaviors was a variable. The .
second correlation matrix included all the physician and patient behaviors hypothesized

in the second, third and fourth hypotheses.

CHAPTER THREE
RESULTS
This study attempted to find relationships among a variety of patient
communicative behaviors indicative of satisfaction/dissatisfaction, and physician
behaviors. Eight specific relationships were tested in four hypotheses, with significant
results found for two of the relationships. In this section, descriptive statistics for each
of the behaviors are presented, results for each of the hypotheses are reviewed, and
relationships discovered beyond those hypothesized are also discussed.
Descriptive Statistics
Twelve patient behavior variables and 12 physician behavior variables were
identified in the present study. Table 2 illustrates the mean occurrence of each

variable, the range, and the standard deviation across all transcripts.

35

36

Table 2
Descriptive Statistics for Patient and Physician Behaviors

 

Behavior Variable Mean Range Std Dev Total

 

Patient Behaviors

 

 

 

asks questions 14.82 0 - 59 13.80 252
shows disagreement or criticism 0.53 0 - 4 1.01 9
summarizes/repeats self 1.06 0 - 3 1.09 18
expresses non-disease tension 3.35 0 - 9 2.74 57
expresses disease tension 1.94 0 - 7 2.36 33
requests opiniom/sug gestions 1.06 0 - 3 1.09 18
summarizes other 1.41 0 - 4 1.18 24
personal remarks or digressiom 5.76 0 - 17 4.52 98
responds directly to a question 23.00 2 - 53 13.80 391
gives volrmtary orientation 30.12 7 - 69 15.44 512
shows agreement or tmderstanding 22.18 6 - 58 14.39 377
offers solidarity 0.65 0 - 6 1.54 11
Physician Behaviors
shows disagreement or criticism 0.35 0 - 1 0.49 6
gives direction/instruction 7.24 r - 20 5.49 123
summarizes/repeats self 2.94 0 - 16 4.02 50
personal remarks or digressions 6.65 l - 20 4.91 113
gives information or opinion 43.94 5 - 106 29.48 747
indicates listening 19.53 4 - 47 12.50 332
summarizes other 2.76 0 - 8 2.49 47
requests questions 0.41 0 - 2 0.62 7
requests opinions or suggestions 0.35 0 - 2 0.61 6
asks questiom 26.53 4 - 56 14.25 451
shows agreement or understanding 9.24 0 - 21 6.05 157
shows solidarity 3.65 0 - 7 2.23 62

 

Note: N - l7 transcripts. Means represent average occurrence in each transcript.

Range, Standard Deviations and Total Utterances represent distribution of behaviors

across all tramcripts.

Tension Release
The first hypothesis posited a positive relationship between patient tension .
release and physician expressions of: (1) agreement and understanding; and (2)

solidarity. The first part of this was confirmed. The Pearson-r correlation between

37
patient tension release and physician agreement/understanding was .47 (p s .05).
Thus, as the incidence of a patient expressing disease-related tension increases or
decreases, the rate of the physician expressing agreement or understanding likewise
increases or decreases.

The second part of the first hypothesis, the relationship between patient tension
release and physician expressions of solidarity, was not confirmed. The Pearson r
correlation was .24, and not statistically significant. Table 3 illustrates the correlations
between variables in the first hypothesis. Significant relationships are distinguished

with an asterisk, and the significance level is indicated below the table.

Table 3
Correlation Matrix of Patient Tension Release
and Physician Expressions of Agreement/Understanding and Solidarity

 

Patient Tension Physician Agreement] Physician

 

Release Understanding Solidarity
Patient Temion 1.0000 .4677* .2369
Release
Physician Agreement] 1.0000 .2687
Understanding
Physician Solidarity 1.0000

 

Note: N - 17. Correlations represent only behaviors in the introduction—history
segment of the interview.
* p s .05
Non-Disease Tension
The second hypothesized relationship concerned patient expressions of non-
disease related tension. According to the literature, these statements indicate

dissatisfaction on the patient’s behalf (Carter et al., 1982). They were hypothesized to

be negatively correlated with the positive behaviors of a physician showing agreement

38
or understanding and expressing solidarity. However, neither relationship was
supported. The Pearson-r correlation for non-disease tension and physician
agreement/understanding was r = .10 (n.s.), while the correlation for non-disease
tension and physician solidarity was r = .41 (n.s.; see Table 4). This would indicate
that the expression of concerns or tensions which are not related to a physical
condition, but rather to unrelated circumstances (such as complaints regarding the
length of time the patient waited before being seen by the physician, or difficulty in
finding transportation or obtaining the appointment), are not associated with these

particular physician behaviors.

Table 4
Correlation Matrix of Patient Tension, Question Asking and Summarizing
with Related Physician Behaviors

 

 

Physician Behaviors

 

 

Patient

Behaviors Agreement] Shows Asks Smnmarizes Smnmarizes
Understanding Solidarity Questions Self Other

Non-Disease .1062 .41 10 -. 1568 -.0992 -.0984

Temion

Question .0867 .1800 -.5916* .6828“ -.2781

Asking

Summarize] .3872 .2147 .1984 -.0991 -.2485

Repeat Self

 

My; N - l7. Correlations represent behaviors from all segments of the interview.
*ps.01,**p s.001
Question Asking
For the third hypothesis, question asking behaviors of the patient were
investigated. The ”frequent asking of questions” was operationalized as the ratio of
questions to other behaviors. Included in this ratio are all questioning behaviors. For

patients this would include questions under the first, broad category of ”asks

39

questions” as well as those under the sixth category, ”requests opinions and
suggestions.” For the physician, question asking behavior incorporates three
categories: ”requests questions,” ”requests opinions/suggestions,” and ”asks questions."

The first part of the hypothesis stated that the frequency of patient questions
should be negatively correlated with the frequency of physician questions. This
relationship was strongly supported. The Pearson-r correlation was -.59 (p s .01; see
Table 4). Therefore, as a physician asks more questions of the patient, the patient
asks fewer questions of the physician.

The second part of the hypothesized relationship stated that patient questions
should be negatively correlated with the physician summarizing her/himself. This
relationship was not supported. In fact, where a negative relationship was anticipated,
a strong positive relationship was discovered. The Pearson-r correlation for this
relationship was .68, p s .001 (see Table 4).

Repetitiveness

The final hypothesis dealt with repetitive patient behaviors. It must be noted
that there were too few instances of repetitiveness in the sampled transcripts to be able
to accurately test this hypothesis (see Table 2). However, the findings are reported
here. Negative relationships were hypothesized between repetitive utterances by the
patient and two physician behaviors: (1) summarizing the patient; and (2) showing
signs of agreement and understanding. Although not statistically significant, the
correlation between repetitive utterances and physician summarizing her/himself was in

the predicted direction (r = -.25, n.s.). The second part of the hypothesis was not

40

supported. The Pearson-r correlation for patient repetition and physician statements of
agreement/understanding was .39, (n.s.; see Table 4).
Other Relationships

Although not directly hypothesized, several other significant relationships
emerged between patient and physician verbal behaviors (see Table 5). A strong
positive correlation was found between the patient offering commitment and the
physician directly requesting opinions or suggestions from the patient (r = .54, p s
.01). Likewise, patient solidarity was strongly correlated with the physician requesting
questions from the patient (r = .49, p s .05). Although no causal ordering can be
established, it would appear that the more the physician elicited either opinions and
suggestions or questions from the patient, the more the patient demonstrated a
commitment to the discussed medical regimen.

The positive behavior of patients giving voluntary orientation was also found to
be highly correlated with two physician variables. The first relationship is with
physicians showing agreement or understanding (r = .83, p s .001). The second is
with physicians giving information or opinions (r = .65, p s .01). This would suggest
that as physicians express agreement with and understanding of the patient, the more
the patient is willing to offer further information. Additionally, the more information
the physician offers, the more the patient offers as well. Again, no causal ordering
can be inferred from the data.

Another interesting finding regarding patient orientation giving is that is was
negatively correlated with physician question asking (r = -.42, p s .05). As physicians

ask more questions, patients volunteer less information. In fact, patients disclose more

41

information when the physician provides information, than when the physician asks

excessive questions.

Table 5 summarizes the relationships found between variables which were not

specified in the original hypotheses. The implications of these findings are discussed

in the final chapter.

Table 5
Correlation Matrix of Patient Solidarity and Orientation
with Related Physician Behaviors

 

 

 

Patient Physician Behaviors
Behaviors
Agreement] Request Opinion] Request Give Information] Ask
Understanding Suggestions Questions Opinion Questions
Solidarity 0.3924 0.5438“' 0.4907* 0.0946 -0.0543
Voluntary 0.8309*** -0.1315 -0.0708 0.6471** -0.4218*
Orientation

 

Note: N - 17. Correlations represent behaviors from all segments of the interview.
*p$.05, **p 5.01, ***p s .001

CHAPTER FOUR
DISCUSSION

Perhaps Rowland-Morin and Carroll (1990) expressed it best: ”If it is true that
patient satisfaction is significantly affected by the interviewing style of doctors, then it
is important to identify specific variables in the interviewer’s style that produce such
an effect” (p. 173). This study has attempted to do just that. The ability of a
physician to discern dissatisfaction during interaction with a patient is particularly
valuable. When a physician is aware that the patient may be unhappy with the visit,
s/he can take immediate steps to turn those feelings around, and thus decrease the
likelihood of malpractice suits and increase the probability of the patient complying
with medical advice.

This paper identified patient and physician behaviors which have been
associated with satisfaction as an outcome of the medical interview. It then proposed
a specially adapted translation of three coding procedures in order to tackle the
challenge of identifying satisfaction and dissatisfaction as a process which develops in
the interview, and which can be encouraged by specific verbal behaviors from the
physician. This process is revealed by examining the correlations between physician
behaviors and patient behaviors which are known to indicate satisfaction or
dissatisfaction.

The developed coding system is a combination of content categories from
Bales’ Interaction Process Analysis, Roter’s modified IPA, and Gottman’s Couples.
Interaction Scoring System. This combination, although not completely free from the

limitations of other systems, does provide for two things. The system presents a

42

43

comprehensive set of categories which are mutually exclusive and ones which are
particular to the individual interactants’ roles (i.e., physician and patient). It also
incorporates behavior categories which have been identified by different researchers
and linked to patient satisfaction. Thus, it synthesizes the findings of numerous
studies.

Significant findings included a positive correlation between patients relieving
tension through the expression of concerns regarding their illness or treatment and
physicians showing agreement or understanding during the initial introduction and
history taking segment of the medical interview. The data also revealed two
significant findings regarding patient question asking behaviors. In particular, the
frequency of patient questions was negatively correlated with physician questions, and
positively correlated with the physician summarizing her/himself.

Several significant relationships which were not hypothesized, but which
emerged in the data were relevant to patient solidarity and voluntary orientation
giving. Specifically, the physician behaviors of showing agreement or understanding
and giving information or opinions were positively correlated with patients freely
offering information, while physicians asking questions was negatively correlated with
patients giving additional information (information which was not directly requested
by the physician, but which may aid in diagnosis or treatment). Furthermore, the
physician behavior of requesting opinions or suggestions from patients was positively

correlated with patients expressing solidarity or offering commitment.

44
Implications of the Findings
Dim-Related Tension Release

The first hypothesis was developed based upon research by Carter et al. (1982)
which found that patient expressions of tension release are positively related to
satisfaction. In the research, tension release was operationalized as statements of
disease-related tension, whereby the patient shared concerns about their medical or
socioemotional condition with the physician. It was posited that when these behaviors
are met or encouraged by statements of agreement or understanding from the physician
or statements of solidarity, they will result in higher levels of satisfaction for the
patient.

This hypothesis was partially supported, with a strong correlation between
physician agreement and patient tension release. This finding is significant for two
reasons. First, it confirms how important it is that the physician is aware of the
impact that her/his communication has on the satisfaction of the patients, and
potentially on long term health behaviors. In order to make the patient feel more
comfortable and to relieve their anxieties the physician must demonstrate empathy and
agreement, letting the patient know that her/his concerns are valid, and that they can
be dealt with effectively. Second, given that patient tension release and physician
behaviors which demonstrate support and empathy were both identified in the
literature as strongly correlated with patient satisfaction (Carter et al., 1982; Feletti et
al., 1986; Hall et al., 1981) this findings lends support to the idea that satisfaction can
be identified not only as an outcome of interaction, but as it develops in the medical

interview.

45
(ligation Asking

The literature suggests that excessive questioning behaviors by the patient are
indicative of dissatisfaction with the interview, because they force the patient to
extract information which the physician should be offering freely. Given this premise,
patient question asking should be negatively correlated with physician behaviors which
are strongly correlated with satisfaction. One of these behaviors, according to
Comstock et al. (1982) and Woolley et al. (1978) is physician question asking.

The hypothesized negative correlation between physician and patient question
asking was supported, and has several implications. The relationship is likely due, in
part, to an increased sense of physician ”thoroughness.” Characteristics of an ideal
physician, identified by Feletti et al., (1986) which were highly associated with general
satisfaction included competence in the physical domain and the amount of time spent
in consultation. When the physician asks many questions of the patient, this may
increase the patient’ 8 perceptions of the physician’s competence and the amount of time
the physician is spending with the patient, examining problems and trying to elicit
valuable information from the patient. This increasing satisfaction results in a
decreased need for the patient to ask questions.

These data would suggest that physicians need to be aware that their
questioning behaviors are relevant to patient satisfaction, as well as to their diagnosis.
By asking questions, physicians also fulfill patient needs. However, the literature does
suggest that questions need to be phrased carefully. Very directive questioning may
actually lead to dissatisfaction, if it communicates physician dominance (Rowland-

Morin & Carroll, 1990). Questions should be open-ended. They should also be

46

intended to clarify patient information and express empathy. Furthermore, the
physician needs to encourage questions from the patient and elicit the patient’s
concerns and expectations. There was a strong positive correlation between physician
asking the patient if s/he had any questions or concerns and the patient’s expression of
solidarity (r = .49, p s .05).

It was further hypothesized that when a physician summarizes her/himself, this
should serve to clarify and synthesize information, thereby decreasing the need for
patient questions. However, instead of the anticipated negative correlation, a strong
positive correlation was found between patient questions and physicians summarizing
themselves.

Several things may contribute to this unexpected finding. First, as the
physician summarizes things s/he has already said, further clarifying information or
instruction, the patient may think of new things they want to ask and take this
opportunity to be sure they understand. By summarizing earlier statements, the
physician demonstrates a concern that the patient understands, and perhaps an
increased willingness to take time to assist the patient. This may allow a greater
opportunity for the patient to ask questions, and may make the patient feel more
comfortable to ask questions. When physicians fail to review their comments, they
may create an impression that they are either hurried, or not very concerned about the
patient’s level of comprehension. Such an impression might actually inhibit the
question asking behaviors of patients.

This assumption is supported with an examination of the correlation between

patient question asking and total time spent in the interview. The questions-time

47

relationship has a correlation of r = .55 (p s .01). Furthermore, the physician
summarizing her/himself is also correlated with the length of interview or time spent
with patient. This relationship has a correlation of .46 (p s .05). Lastly, those
physicians who took the time to summarize themselves, are also the ones who more
frequently asked the patient if slhe had any questions. The correlation between
physician summarizing self and physician requesting questions of the patient is .49 (p
s .05). This is further evidence that those physicians who were summarizing
themselves may have imparted a greater sense of comfort on the patient, allowing the
patient more freedom and confidence to ask questions.
Solidarity

One of the most interesting findings of this study concerned the patient
expressing solidarity with the physician in overcoming difficult health situations,
maintaining a prescribed health regimen, or otherwise leading a healthy lifestyle. This
type of verbal behavior is a positive expression of commitment from the patient,
indicating a willingness to work with the physician. As such, any behavior on the part
of the physician which encourages this would be beneficial. The data revealed that
two physician behaviors were highly correlated with patient expressions of solidarity:
(1) requests for questions; and (2) requests for opinions or suggestions.

Therefore, important questioning behaviors in which physicians should engage
are asking patients if they have any questions or concerns and seeking opinions or
suggestions from patients regarding their own health. Particularly when opinions and

suggestions are elicited, they seem to be accompanied by solutions in which the

48

patient believes and for which the patient has a high level of self-efficacy, thus
increasing patient commitment. This conclusion is summarized well by Lane (1983):

”Communication is the best method for getting at patient expectations.

When patients and physicians agree about expectations regarding an

illness or treatment, patients have greater expectations for improvement.

Similarly, ...communicating about expectations is the best method of

fulfilling them and creating treatment regimen adherence” (p. 775).
Orientation Giving

The research also found a positive correlation between patients freely
volunteering information, and the physician behaviors of giving information or
opinions and showing agreement or understanding. These findings would suggest first
that the basic principle of reciprocal exchange applies to physician-patient interactions,
so that increased disclosure of information offered by one individual is reciprocated
with an increase in disclosure by the communication partner. Furthermore, it
demonstrates that this reciprocal exchange is encouraged by the expression of
agreement and understanding on the part of the physician. Note that listening
behaviors were not correlated with patient information sharing, but rather the
expression of true understanding, which differed in this coding scheme.
Demonstrating listening (i.e., ”okay,” ”yeah,” and ”I see”) appears to be insufficient. It
is important that the physician become involved in the conversation to a greater extent,
showing the patients that they are aware of the implications of their information and

that they empathize with them.

49

This last assumption relates to a hypothesis by Freemon et al. (1971) that
patients perceive longer history taking as ineffective communication. Perhaps it is not
necessarily the length of history taking time, but the nature of the conversation which
prompts perceptions of ineffectiveness. If excessive question asking is perceived
negatively by patients, then this would be consistent with the finding that increased
question asking is negatively correlated with voluntary information giving. As
discussed previously, Freemon et al. (1971) found that patients more favorably rate a
physician who freely offers information than one who requires the patient to extract
information through extensive question asking. Again, this voluntary offer of
information from the physician seems to be responded to favorably with more
information from the patient, increasing the length of time spent in history taking, but
doing so in a productive and mutually satisfying manner.

In summary, increased disclosure of information and increased commitment
from the patient are behaviors which are critical to the promotion of health as well as
to the development of satisfaction. Physician should be able to promote these
behaviors by expressing more understanding and agreement with patients, as well as
by taking time to elicit the opinions, suggestions and questions of patients. These
findings offer valuable insight into improving the physician-patient interaction.

Limitations and Directions for Future Research

Several limitations of the present research must be acknowledged. Each of
these limitations, however, provide opportunities for future research which can
strengthen our understanding of communication in the health care setting and help in

the development of more effective medical interviews.

50

Methodological Issues

First, several hypothesized relationships were not supported. Weaknesses in
the data set and methodology could account for the lack of significant findings. In
this section, a number of problems common to several of the unsupported hypotheses
will be discussed, and potential solutions presented.

Consistently, when hypotheses were not supported, one problem was too few
instances of a behavior. If there is an insufficient amount of data for any variable,
then it becomes difficult to correlate that variable with other variables, and the
problem is defined as a ”restriction of range.” This was the case for a number of
variables in this study. In particular, there were very few utterances of patient tension,
patient repetitiveness, and physician solidarity. (The average transcript had fewer than
two utterances of patient repetition, while no single transcript had more than three
occurrence of this behavior. Refer to Table 2).

A second problem that emerged in this research regards the method of analysis
itself. For at least one hypothesis, the use of correlations may not be appropriate. For
instance, the hypothesis for patient expressions of non-disease tension was originally
formulated with the assumption that if such behaviors continued, they would lead to
greater overall dissatisfaction with the physician. In order to circumvent such tension
and avoid the development of dissatisfaction, it was suggested that physicians would
need to demonstrate understanding of the expressed concerns and offer solidarity to
the patient. However, these relationships were not found. It is suggested that since
understanding and solidarity may be expressed in conjunction with any number of

patient behaviors (and not just tension statements) they will often occur in the absence

5 1
of any tension. Thus, an observed correlation with tension may not necessarily
indicate a direct relationship between the two behaviors.
Potential Solutions

There are at least two potential solutions to these problems. The first would be
to gather more data. With 30 or more transcripts one might be able to find sufficient
levels of solidarity, tension and repetition to test the hypotheses and find some
support. It is interesting, however, to observe that the variables which occurred most
infrequently on the patient’s behalf are those which are most highly correlated with
dissatisfaction. This points to a potential problem with the sample. It may be that the
patients in the transcripts obtained for this research were largely satisfied with their
physicians. Hopefully, a larger sample would correct for this as well, providing a
more even distribution of satisfied and dissatisfied patients.

The second potential solution to the problems encountered with the research,
would be the utilization of different methods of analysis. For instance, correlations are
problematic when behaviors may occur in response to numerous variables. This was
the case described with physician expressions of agreement or understanding and
statements of solidarity. Sequential analysis techniques such as phase mapping allow
the researcher to examine series of conversation and determine what types of
behaviors accompany other behaviors. That is, for each patient expression of non-
disease tension, one would be able to see what physician verbal behaviors immediately
proceeded and followed the patient’s utterance. Another potential advantage to this
technique is that it would allow the researcher to make conclusions about causation,

which can not be inferred from correlations.

52
@nceptual Issues

Another limitation of the research is its focus on verbal interaction.
Communication is a highly complex process which naturally involves more than verbal
behaviors. The coding scheme developed in this research was designed for and
applied only to transcribed interactions. Thus, all non-verbal behaviors were
overlooked. It is possible that non-verbal cues, such as facial expressions and body
movement, as well as vocal qualities such as tone of voice, rate of speech, and pitch,
would play a significant role in demonstrating patient satisfaction and dissatisfaction.
To some extent physicians may already be able to interpret these cues and any verbal
behaviors which they exhibit in response to non-verbal messages, would be
misinterpreted in this research. It is important that further research is undertaken
which specifically identifies the relationship between various non—verbal or vocal cues
and patient satisfaction, and that these behaviors are examined in conjunction with
physician verbal and non-verbal behaviors. This would provide a bigger picture of the
medical interview as well as more precise tools for physician communication training.

An additional concern involves contradictions in the literature and the present
findings regarding patient questioning behaviors. Given that frequent question asking
(which is purported to indicate dissatisfaction) was negatively correlated with one
positive physician behavior and positively correlated with another, it is not clear if
excessive patient questioning necessarily indicates dissatisfaction. There may be a
problem with how the coding scheme grouped directive, closed-ended questions with
probing and open-ended questions for both the patient and the physician. There are

known differences in how physician questions affect patient satisfaction, but the

53

coding scheme does not adequately address these differences. However, there are
other possibilities which may account for the inconsistencies. For example, one or
more of the variables believed to be associated with satisfaction, particularly patient
question asking, may be poorly specified. There may also be a curvilinear relationship
between the frequency of patient questions and satisfaction. This points to a need for
further research.
flgge Rflch

Research is still needed at several levels. One, researchers must continue to
identify patient and physician variables which are associated with satisfaction. This
will require that outcome measures such as self-reports and third-party observations of
satisfaction continue to be used. Second, relationships between physician and patient
behaviors need to be explored further. By refining the coding scheme and applying it
to a wider sample of interactions, significant relationships may still be found, and
hypotheses which were not supported here may be substantiated. Third, as research
continues to identify variables associated with satisfaction as an outcome, it may be
beneficial to merge methodologies. That is, when researchers collect outcome data
(with self-report or third-party observation measures), they should collect recordings of
the interaction and use interaction analysis as well to obtain process measures of
satisfaction. This may help identify where there are differences, and allow researchers
to better understand some of the weaknesses of each methodology, while permitting a
broader perspective on patient satisfaction. Potentially, such a perspective may also

help to more clearly define patient satisfaction.

54
Pragmatic Relevance

”Descriptions or even predictions of those aspects of communication which are
related to positive effects are not highly meaningful if communication cannot be
improved as a result” (Roter, Hall, & Katz, 1988, p. 112). While the coding scheme
presented here requires some additional refining, it does have practical implications for
physician training in communication and psychosocial medicine. By using this system
to link physician behaviors and patient behaviors which contribute to satisfaction, we
can confirm the ways in which the physician can recognize the development of
satisfaction in the process of interaction, and thus intercede in the development to
facilitate higher levels of satisfaction. This research has identified a number of
physician behaviors which may have a positive impact on patient satisfaction. It is the
responsibility of communication scholars to continue to identify such behaviors and
then to incorporate them into training programs designed to improve physician
communication.

Summ ary

Sir William Osler’s words about listening to the patient ring true today as they
did 90 years ago. However, the present research has demonstrated that listening is not
enough. Physicians must take an active role in the medical encounter, recognizing
patient cues and communicating effectively using a number of techniques in order to
encourage the development of patient satisfaction. Furthermore, they must be acutely
aware of their own communicative behaviors and how they impact patient satisfaction.
Although listening is certainly an important technique, asking questions of and

specifically eliciting questions and opinions from the patient are critical, as are freely

55
offering information and expressing comrrritrnent to the patient. With greater attention

to these issues, improvements can be made in the health care system.

APPENDICES

APPENDICES

Appendix A: Characteristics of an "Ideal" Physician3

 

dramcteristic

Description

 

l.

10.

Competence in a physical domain

Competence in the emotional
domain‘

Competence-social awareness‘
The physician as a model
Amormt of time for consultation‘
Perceived amormt of continuity of
care by the physician

Mutual rmderstanding in the
doctor-patient relatiomhip‘
Patients’ perceptions of their
individuality

Information transfer"

Competence-Physical examination

Perception of the physic ian's ability to understand and
cure their illness.

Perceptiom of the physician’s understanding of the
importance of their emotional status.

Perceptiom of the physician’s understanding of the
interaction between patient’s health and their relationship
with significant others.

Perceptions of whether or not the physician modeled a
desirable style of living.

Perceptions of the physician giving adequate time and
attention to their problem.

Perceptions of whether or not they felt the doctor was
providing adequate follow-up care.

Perceptions of whether or not they felt they were treated
as an equal partner in decision regarding their health
care.

Perceptions of whether they were treated as unique
individuals or as ”yet another” medical complaint.

Perceptions of the value of information supplied to them
by the doctor and their perceptions of the style with
which such transfer of knowledge occurred

Perceptions of the thoroughness, care, and gentleness
with which the physician conducted the examination.

 

56

Appendix B: Bales' Interaction Process Analysis Categories’

 

 

Category Description

1. Showe solidarity raises other’s status, gives help, rewards

2. Shows torsion release jokes, laughs, shows satisfaction

3. Agrees shows passive acceptance, understands, concurs, complies
4. Gives suggestion direction, implies autonomy for other

5. Gives opinion evaluation, analysis, expresses feeling, wish

6. Gives orientation information, repeats, clarifies, confirms

7. Asks for orientation information, repetition, confirmation

8. Asks for opinion evaluation, analysis, expression of feeling

9. Asks for suggestion direction, possible ways of action

10. Disagrees shows passive rejection, formality, withholds help
11. Shows tension asks for help, withdraws

12. Shows antagonism deflates other’s starts, defends or asserts self

 

57

Appendix C: Sources of Tension Statements‘s

 

 

Source Description

1. Disease pain or restriction of activities

2. Treatment urgent need for specific treatment, no desire for recommended
treatment, ineffectiveness of treatment

3. Facility organization waiting time medication dispensing

4. Other doctors withholding information, ignoring problems, incompetence

5. This interview inattentivenees, lack of understanding, demanding behavior of
physician

6. Psychosocial patients perceptiom of self, relationships to friends and family,

socioeconomic situation now or in the past

 

58

Appendix D: Roter's Modified Interaction Process Analysis Categories’

 

Physician

Patient

 

P995099.“

Personal remarks

Shows approval, gives compliment
Statement, gives information, opinion
Gives direction, imtruction

Asks questions

Direct request for questiom

Shows agreement and/or understanding

Shows disagreement or criticism

1.

53°99‘99“?!"

Personal remarks

Shows approval, gives compliment
Statement, gives information, opinion
Request for medication

Bid for clarification

Asks for questions

Shows agreement and/or tmderstanding

Shows disagreement or criticism

 

59

ENDNOTES

1. The six test transcripts were obtained courtesy of Dr. Debra Roter at Johns Hopkins
University.

2. The transcripts for the full-scale research were obtained from Dr. Nancy Ainsworth-
Vaughn of Michigan State University.

3. Reproduced from Feletti, Firman, and Sanson-Fisher, 1986, p. 392.

4. Factors which are directly related to a physician’s ability to communicate effectively with
the patient.

5. Table reproduced from Inui, Carter, Kukull, and Haigh, 1982, p. 539.
6. Table reproduced from Carter, Inui, Kukull, and Haigh, 1982, p. 562.
7. Table reproduced from Wasserman and Inui, 1983, p. 287.

60

LIST OF REFERENCES

Anderson, R., Fleming, G.V., Aday, L.A., & Aguirre, M. (1979). Consumer viewpoint:
"What is health care and what do we want?” and a response (U .S. Department of
Health Education and Welfare, DHEW Publication HRA 79-632). Washington DC:
Government Printing Office.

Beckman, H., Kaplan, S., & Frankel, R. (1989). Outcome based research on doctor-patient
communication: A review. In M. Stewart & D. Roter (Eds) Communicating with
medical patients. pp. 223-227. Newbury Park, CA: Sage.

Carter W., Inui, T., Kukull, W., & Haigh, V. (1982). Outcome-based doctor-patient interaction
analysis 11. Identifying effective provider & patient behavior. Medical Care, 20, 6, 5-
50-566.

Comstock, L.M., Hooper, E.M., Goodwin, J.M., & Goodwin, IS (1982). Physician
behaviors that correlate with patient satisfaction. Journal of Medical Education, 57,
105-1 12.

Davis, MS. (1968). Variations in patients’ compliance with doctors’ advice: An empirical
analysis of patterns of communication. American Journal of Public Health, 58, 274.

Davis, MS. (1971). Variation in patients’ compliance with doctors’ orders: Medical practice
and doctor-patient interaction. Psychiatry in Medicine, 2, 31-54.

Donabedian, MS. (1982). The Criteria and Standards of Quality. Ann Arbor, MI: Health
Administration Press.

Feletti, G., Firman, D., & Sanson-Fisher, R. (1986). Patient satisfaction with primary-care
consultations. Journal of Behavioral Medicine, 9, 4, 389-399.

Freeman, B., Negrete, V., Davis, M., & Korsch, B. (1971). Gaps in doctor-patient
communication: Doctor-patient interaction analysis. Pediatric Research, 5, 298-311.

Gottman, J.M. (1979). Marital Interaction: Experimental Investigations. New York:
Academic Press.

Hall, J.A., Roter, D.L., & Rand, CS. (1981). Communication of affect between patient and
physician. Journal of Health and Social Behavior, 22, 18-30.

Hulka, B., Kupper, L., Cassel, I ., & Mayo, F. (1975). Doctor-patient communication and
outcomes among diabetic patients. Journal of Community Health, 1, 1, 15-27.

Inui, T.S., & Carter, WE. (1985). Problems and prospects for health services research on
provider-patient communication. Medical Care, 23, 5, 521—532.

61

62

Inui, T., Carter, W., Kukull, W., & Haigh, V. (1982). Outcome-based doctor—patient
interaction analysis I: Comparison of Techniques. Medical Care, 20, 6, 535-549.

Lane, SD. (1983). Compliance, satisfaction, and physician-patient communication. In R.
Bostrom (Ed), Communication Yearbook 7, 772-799. Beverly Hills, CA: Sage.

Ley, P., Bradshaw, P., Kincey J ., & Atherton, S. (1976). Increasing patients’ satisfaction
with communications. British Journal of Social and Clinical Psychology, 15, 403-413.

Mishler, E., Clark, J., Ingelfinger, J., & Simon, M. (1989). The language of attentive patient
care: A comparison of two medical interviews. Journal of General Internal
Medicine, 4, 325-335.

Oksenberg, L., & Cannell, C. (1977). Some factors underlying the validity of response in
self report. Bull. Int. Stat. Inst, XLVII, 3, 325-346.

Roter, Debra. (1989). Which facets of conununication have strong effects on outcome: A
meta-analysis. In M. Stewart & D. Roter (Eds) Communicating with medical patients.
pp. 183-196. Newbury Park, CA: Sage.

Roter, D.L., & Hall, IA. (1987). Physicians’ interviewing styles and medical information
obtained from patients. Journal of General Internal Medicine, 2, 325-329.

Roter, D., Hall, J., & Katz, N. (1987). Relations between physicians’ behaviors and analogue
patients’ satisfaction, recall, and impressions. Medical Care, 25, 5, 437-451.

Roter, D., Hall, J., & Katz, N. (1988). Patient-physician communication: A descriptive
summary of the literature. Patient Education and Counseling, 12, 99-119.

Rowland-Morin, P.A., & Carroll, 1.6. (1990). Verbal communication skills and patient
satisfaction: A study of doctor-patient interviews. Evaluation and The Health
Professions, 13, 2, 168-185.

Smith, R., Osborn, G., Hoppe, R., Lyles, J., Van Egeren, L., Henry, R., Sego, 1)., Alguire, P.,
& Stoffelmayr, B. (1991). Efficacy of a one-month training block in psychosocial

medicine for residents: A controlled study. Journal of General Internal Medicine, 6,
535-543.

Smith, R.C., & Hoppe, RB. (1991). The patient’s story: Integrating the patient- and
physician-centered approaches to interviewing. Annals Internal Medicine, 115, 470-
477.

Stewart, Moira A. (1984). What is a successful doctor-patient interview? A study of
interactions and outcomes. Social Science Medicine, 19, 2, 167—175.

63

Stiles, W.B. (1979). Discourse analysis and the doctor-patient relationship. International
Journal of Psychiatry in Medicine, 9, 3, 263-273.

Stiles, W.B., Putnam, S.M., & Jacob, M. C. (1982). Verbal exchange structure of initial
medical interviews. Health Psychology, 1, 4, 315-336.

Stiles, W.B., Putnam, S.M., Wolf, M.H., & James, SA. (1979). Interaction exchange
structure and patient satisfaction with medical interviews. Medical Care, 17, 667-681.

Stone, G. (1979). Patient compliance and the role of the expert. Journal of Social Issues,
35, 1, 34-59.

Waitzkin, Howard. ( 1984). Doctor-patient communication: Clinical implications of social

scientific research. Journal of the American Medical Association, 252, 17, 2441-
2446.

Ware, J., Davies-Avery, A., & Stewart, A. (1977). The measurement and meaning of patient
satisfaction: A review of the literature. Santa Monica, CA: Rand Corporation.

Wasserman, R.C., & Inui, TS. (1983). Systematic analysis of clinician-patient interactions:

A critique of recent approaches with suggestions for future research. Medical Care,
21, 3, 279-293.

Wolf, M., Putnam, S., James, S.A., & Stiles, W. (1978). The medical interview satisfaction
scale: Development of a scale to measure patient perceptions of physician behaviour.
Journal of Behavioural Medicine, 1, 4, 391-401.

Woolley, F.R., Kane, R.L., Hughes, C.C., & Wright, DD. (1978). The effects of doctor-
patient communication on satisfaction and outcome of care. Social Science and
Medicine, 12, 123-128.

MICHIGAN sran UNIV. LIBRARIES
"HIllllllllllllllllWilliWlllllllllllllllllllllllll
31293010253916