| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Department of Behavioral Sciences in Medicine (P.K.G., A.F.), University of Oslo, Oslo, Norway.
Address reprint requests to: Peter Kjær Graugaard, Department of Behavioural Sciences in Medicine, Postboks 1111, Blindern, N-0317 Oslo, Norway.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
METHODS: Twenty students with low trait anxiety and 21 students with high trait anxiety each had a single consultation with a physician who performed the consultation using either a patient-centered or doctor-centered style of communication. Questionnaires about emotional state were completed by the students before and after the consultation, and a questionnaire about satisfaction was completed after the consultation.
RESULTS: Students with low trait anxiety were significantly more satisfied with a patient-centered than a doctor-centered style of communication. There were no significant differences in emotional response to the two styles of communication. Students with high trait anxiety reacted emotionally more positively to a doctor-centered communication style, with significant and nearly significant change scores for the emotions of tension/anxiety and vigor/activity, respectively. No significant difference was found between satisfaction scores.
CONCLUSIONS: Data indicate that differences between subjects’ emotional traits may be of importance for a differentiated response to patient-centered and doctor-centered communication styles. Subjects’ trait anxiety seems to be a significant factor that should be taken into account when assessing the effects of different communication styles.
Key Words: physician-patient relation • communication skills • patient-centered approach • satisfaction • emotional reactions
Abbreviations: HADC = students with high trait anxiety subjected todoctor-centered intervention; HAPC = students with high traitanxiety subjected to patient-centered intervention; LADC =students with low trait anxiety subjected to doctor-centeredintervention; LAPC = students with low trait anxiety subjected topatient-centered intervention; POMS = Profile of Mood States; STAI = Spielberger State-Trait Anxiety Inventory.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
Several authors have described the skills considered to be important to promote good outcomes in a model of communication called "patient-centered" communication (10–12). Although this model is not defined unambiguously, certain key elements stand out and are often contrasted to a so-called "doctor-centered" style of communication. In the patient-centered style of communication, there is a move away from the ideas in the physician’s mind toward those in the patient’s mind as the central topics for the dialog between the two participants. In addition, social, psychological, and emotional aspects of the disorders are generally assumed to be as equally important as the somatic aspects. This view has obvious implications for the manner in which the physician performs interviews with patients. Skills that facilitate the patient’s expression of expectations, opinions, and feelings are considered crucial, as is the physician’s ability to negotiate with the patient on the different possibilities of examination and treatment.
Evidence of the effectiveness of the patient-centered model in relation to psychological health has mostly been derived from studies that have not specifically been designed to evaluate this model but that nevertheless have been interpreted as supporting one or a number of its elements (4). In two studies, however, a more extensive examination of the model was performed with psychological outcome as a dependent variable. One study applied a computer-based interaction analysis to evaluate the relation between patient-centered communication style and psychological adjustment for patients with cancer (13). In this study, in which the analysis of communication was approached from both a micro and a macro perspective, only one indicator provided support for the notion that patient-centered consultations would contribute to an improved psychological outcome. The patients who had their questions answered showed better psychological adjustment 3 weeks after the consultation. Another thorough study examined the effects of training physicians in specific emotion-handling and problem-defining skills (7). In this study, it was found that patients with high distress at baseline who consulted trained physicians showed significantly less distress 2 weeks after the consultation than distressed patients consulting physicians without special training. The authors themselves do not specifically use the term patient-centered to designate the skills taught, but the similarities of both emotion-handling and problem-defining skills to the elements of the patient-centered model are so striking that it seems justified to consider the study a test of the model.
The limited research specifically designed for testing the patient-centered model leaves several questions unanswered. Because most studies cited in support of the patient-centered model evaluate health outcome 2 to 6 weeks after a single consultation, little is known about patients’ reactions immediately after the consultation or about the long-term effects of ongoing treatment, which may allow a stronger relationship between the physician and patient to evolve. In addition, patient characteristics are often restricted to demographic and diagnostic data, whereas one could expect emotional and psychological factors to be highly important.
Reports from patients indicate sensitivity not only to the information they receive from the physician but also to the manner in which the physician performs the consultation (14). Even though feedback is not necessarily given to the physician during the consultation, patients will often later express strong emotions about the physician’s style of communication. We suspect that these emotional experiences are of considerable importance for different outcome measures, particularly emotional and psychological adjustment to disease.
In the study described here, we applied an experimental design to explore the predictive value of one stable psychological characteristic, trait anxiety, for different emotional reactions to patient-centered and doctor-centered communication styles. Based on the results of existing research and clinical experience, the following hypotheses were formulated: 1) Patient-centered and doctor-centered communication styles affect subjects’ emotional reactions to the consultation in significantly different ways as measured immediately after the consultation. The differences will be more marked for subjects with high than low trait anxiety. 2) Subjects undergoing a patient-centered consultation will generally feel an increase in positively experienced emotions and a decrease in negatively experienced emotions immediately after the consultation (both of these effects are henceforth referred to as positive emotional reactions). In contrast, a doctor-centered communication style promotes a decrease in positively experienced emotions and an increase in negatively experienced emotions (referred to as negative emotional reactions). 3) Subjects undergoing a patient-centered style of communication will generally be more satisfied immediately after the consultation than those receiving a doctor-centered style of communication.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
Participants
Students.
Medical, dental, and psychology students in their first or second year of study were invited at lectures to participate as "patients" in a study examining aspects of communication between physician and patient. Students who agreed to participate were asked to complete the trait part of the STAI during the lecture. To maximize the homogeneity of the sample and because more female than male students were recruited, the experiment was limited to the female sex. Of the 82 women who wished to participate, the 25 with the lowest and the 26 with the highest trait anxiety scores were selected for participation. Subjects in the two subsamples were labeled as low and high anxiety students, respectively. Of these 51 students, 10 dropped out for reasons such as lack of time and problems of availability. Forty-one students participated in the experiment, of whom 16 studied medicine; 3, odontology; and 22, psychology. Mean age was 21.4 years (range, 19–38 years; SD, 3.14 years).
Physician.
One experienced physician conducted all 41 consultations with the students.
Procedures
Students’ instructions.
Appointments with the physician were made for the students by telephone. Instructions about their participation were given by mail 2 days before the consultation. They were asked to participate as "patients" in a consultation with a physician whom they had never before consulted. They were asked to imagine the consultation as an ordinary one, although no somatic examination was undertaken. They were encouraged to present actual symptoms, but if this was not possible, the reason for seeing the physician was to be selected from the following possibilities: 1) a former symptom or disease presented as a current one or 2) concern because of a relative’s current disease. Students were asked not to role play but to present real symptoms or worries except in the case of presenting a former symptom or disease. In this case, they were to present the former concern as a current one.
Physician’s instructions.
The physician was trained to differentiate his communication style between patient-centered and doctor-centered as summarized in Table 1. Biomedical aspects of the presented concerns were to be treated with equal care without respect to style of communication. To achieve this, the physician was instructed to integrate the required doctor-centered elements in the patient-centered consultations (15).
|
Consultations.
The experiment was performed in a university communication laboratory. The laboratory was sparsely furnished with equipment belonging to a physician’s consulting room. All students attended their consultation between 9 AM and noon.
Before the consultations, an assistant who was blind to the students’ anxiety scores and the physician’s communication style administered a 65-item questionnaire (POMS) measuring multiple emotional dimensions. Informed consent was obtained from all students.
The physician communicated with the students in either a patient-centered or a doctor-centered manner, complying as much as possible with the given guidelines (see Table 1) without respect to student’s problems, personality, and responsiveness to the chosen communication style. Consultations were videotaped.
After the consultations, the same assistant administered the POMS questionnaire along with a questionnaire measuring satisfaction with the consultation. Students were given an opportunity to comment on the consultation.
Questionnaires
Trait anxiety.
Trait anxiety was measured using the trait part of the Norwegian adaptation of the STAI (form Y), which consists of 20 items (Håseth K, Hagtvet K, Spielberger CD. Manual for Norsk Tilstands-Trekk Angst Inventorium, form Y [unpublished]. Oslo: Institute of Psychology, University of Oslo). Scores for each item ranged from 1 to 4 (1, not at all; 2, somewhat; 3, moderately so; and 4, very much so).
Emotional state.
Students’ emotional state was measured immediately before and after the consultation by administering a Norwegian translation of the POMS (17). The POMS was originally developed to measure fluctuating emotional states in people undergoing counseling and psychotherapy, but the instrument has been used in many other investigations to assess transient, distinct emotional states. The questionnaire is composed of 65 items scored by subjects on a five-point scale. A score ranging from 1 (not at all) to 5 (extremely) was applied. The instrument contains six subscales originally derived from factor analytic studies (depression/dejection, tension/anxiety, anger/hostility, confusion/bewilderment, fatigue/inertia, and vigor/activity). The "right now" response set (eg, "How are you feeling right now?") was used in this study. The test-retest reliability of the POMS has been reported to range from 0.65 (vigor/activity) to 0.74 (depression/dejection) (17).
Satisfaction.
Seven items from a questionnaire originally consisting of 11 items were used to measure subjects’ satisfaction (18). Four items dealing with satisfaction with the examination part of the consultation were removed from the questionnaire. Scores for each item ranged from 1 to 6; low scores indicated a high degree of satisfaction.
Statistical Analysis
Time effects (from baseline to after consultation) were evaluated by using a paired samples t test. The effects of communication style and anxiety on affective response and satisfaction were assessed by using two-way analyses of variance. Moreover, differences between groups were further evaluated by applying post hoc analyses of the least significant differences.
Statistical analyses were performed by using SPSS software.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
The scores of medical compared with psychology students for the two subsamples showed significant differences, with more medical students displaying low anxiety and more psychology students displaying high anxiety.
Students’ Symptoms and Duration of Consultations
The complaints for which the students consulted the physician were quite similar to the kind of problems young women generally present to a general practitioner. Nine students complained of symptoms related to the musculoskeletal system, with vascular and respiratory system complaints being the second most common choice. The students seemed to behave to a high degree as if the consultation was for a real clinical situation, very likely because the great majority were actually consulting the physician because of current symptoms of their own. However, it is hard to determine the degree to which the students in fact experienced the consultation as a normal consultation. In three instances, the tapes revealed that the students were presenting former symptoms as current ones. Five students presented concerns about a relative’s disease.
The mean duration of consultations conducted in a patient-centered manner was 21.00 minutes (range, 12.22–31.50 minutes; SD = 4.40 minutes); the mean duration of doctor-centered consultations was 13.34 minutes (range, 7.36–22.42 minutes; SD = 3.38 minutes). Differences in duration between the two communication styles were highly significant (p < .001). However, no significant difference was found in the duration of consultations between low- and high-anxiety students, nor did the duration of the consultations predict affect change scores or satisfaction scores when logistic regression was applied.
Emotional Reactions
Table 2 presents mean POMS scores for the four sample subgroups. The internal consistency of each POMS subscale was computed, producing Cronbach’s 
values ranging from 0.67 at baseline to 0.71 after consultation for confusion/bewilderment to 0.88 at both time points for depression/dejection.
|
In the LAPC subgroup, the reduction in confusion/bewilderment was even more marked, with a reduction from 1.76 to 1.43 (t = 4.12, p = .003). In the HADC subgroup, there was a significant reduction in tension/anxiety from 2.04 to 1.76 (t = 2.92, p = .014), whereas a significant reduction in vigor/activity from 3.11 to 2.85 in mean score per item (t = 2.33, p = .048) was found in the HAPC subgroup.
Main and Interaction Effects of Trait Anxiety and Communication Style
Both the main and interaction effects of the two variables defining the experimental conditions (level of trait anxiety and communication style) on change scores were computed by applying two-way analysis of variance.
There were no significant main effects of trait anxiety. The one main effect of communication style concerned anger (F = 5.674, p = .022), with more anger in the patient-centered subgroups and less anger in the doctor-centered subgroups.
For one subscale, tension/anxiety, there was a significant interaction effect between trait anxiety and communication style (F = 4.461, p = .041). In the low-anxiety subgroups, tension/anxiety tended toward reduction in the patient-centered intervention condition, whereas in high-anxiety students, tension/anxiety reduction was significantly associated with the doctor-centered style (Figure 1). A parallel, nearly significant interaction effect was observed for the vigor/activity subscale.
|
Satisfaction.
The seven items of the questionnaire relating to satisfaction together with the mean scores for each of the four groups are presented in Table 3. Because item 7 requests information about the degree of general satisfaction with the consultation, responses to this item were examined. Two-way analysis of variance was applied to compute main and interaction effects. No significant main effect of either communication style or trait anxiety was found. However, a significant interaction effect was identified between trait anxiety and communication style (F = 7.20, p = .011). In the low-anxiety subgroups, students consulted in a patient-centered manner were the most satisfied, whereas in the high-anxiety subgroups, students consulted in a doctor-centered manner were the most satisfied. Post hoc analyses showed a significant difference between the LADC and LAPC subgroups (p = .012), whereas no such difference was found between the HADC and HAPC subgroups.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
Hypotheses
Our first hypothesis was that patient-centered and doctor-centered communication styles would affect subjects’ emotional reactions to the consultation in significantly different ways as measured immediately after the consultation. It was hypothesized that the differences would be more marked for subjects with high than low trait anxiety. Communication style was found to have a main effect only for anger/hostility. However, for tension/anxiety (significant) and vigor/activity (nearly significant), interaction effects emerged between trait anxiety and style of communication. This stresses the importance of taking subject characteristics into account when evaluating communication style. As expected, the differences in emotional responses were most marked for students with high trait anxiety.
Our second hypothesis was that subjects undergoing a patient-centered consultation would generally feel an increase in positively experienced emotions and a decrease in negatively experienced emotions immediately after the consultation. In contrast, it was hypothesized that a doctor-centered communication style would promote a decrease in positively experienced emotions and an increase in negatively experienced emotions. Unexpectedly, we did not find a general tendency for emotional change in a positive direction after a consultation conducted in a patient-centered style, nor did we find a trend toward emotional change in a negative direction when the consultation was conducted in a doctor-centered manner. Results were to some extent in accordance with our hypotheses for the students with low anxiety but in disagreement with our hypotheses for students with high anxiety.
Our third hypothesis was that subjects undergoing a patient-centered style of communication would generally be more satisfied immediately after the consultation than those receiving a doctor-centered style of communication. The pattern of students’ satisfaction scores to some degree paralleled their emotional scores, leaving this hypothesis only partly confirmed. Students with low trait anxiety were significantly more satisfied with a patient-centered than a doctor-centered communication style, whereas among the students with high trait anxiety, no significant difference was found. The trend for these latter subgroups, however, tended more toward satisfaction with a doctor-centered communication style.
Main Findings
The results of this study emphasize the importance of subjects’ emotional and psychological characteristics as determinants of their responses to a physicians’ communication style. The only two studies that have previously related a patient-centered style of communication to satisfaction did not find any significant association (20, 21). In those studies, however, no emotional or psychological patient characteristics were taken into account.
Unexpectedly, students with high trait anxiety responded more positively to a doctor-centered than a patient-centered communication style. One explanation for this could be that anxious subjects prefer a firmer consultation structure and perhaps a more paternalistic physician; a patient-centered communication style could perhaps be experienced by these subjects as an abdication of the physician’s responsibility (22). Perhaps the physician’s facilitation of emotional expression and focus on psychosocial aspects of the disease evoke difficult and therefore unpleasant feelings in students with high trait anxiety immediately after the consultation, but the effect of this communication style turns out to be more positive in the long run.
When considering this surprising finding, one must also take into account that the study was not performed in a real clinical situation and that the problems presented by the students generally were minor somatic complaints. Consequently, students entered the consultations with relatively little illness-related state anxiety. In real clinical situations, one might expect subjects suffering from severe illnesses and a high degree of related anxiety to generally respond more positively to patient-centered communication regardless of the level of trait anxiety. Trait anxiety, however, could turn out to be a good predictor in subjects with minor concerns.
In our study, patient-centered consultations were roughly one-third longer than doctor-centered consultations. Other studies have found either a similar result (20) or no significant difference (7). The longer duration of patient-centered consultations in this study likely reflects the fact that the physician spent extra time exploring illness experience and psychosocial aspects.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
Received for publication August 17, 1998.
Revision received July 9, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  C A Dowson, R G Kuijer, and R T Mulder Anxiety and self-management behaviour in chronic obstructive pulmonary disease: what has been learned? Chronic Respiratory Disease, October 1, 2004; 1(4): 213 - 220. [Abstract] [PDF]
|
|
|
|
 |
|
 J. Parker and E. Coiera Improving Clinical Communication: A View from Psychology J. Am. Med. Inform. Assoc., September 1, 2000; 7(5): 453 - 461. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
