- Email: [email protected]
Is GP–patient communication related to their perceptions of illness severity, coping and social support?
Social Science & Medicine 55 (2002) 1245–1253
Is GP–patient communication related to their perceptions of illness severity, coping and social support? M. Deveugele*, A. Derese, J. De Maeseneer Department of General Practice and Primary Health Care, University of Ghent, UZ-1K3, De Pintelaan 185, 9000 Gent, Belgium
Abstract The aim of the study was to explore the relationship between the communicative behaviour of general practitioner and patient on the one hand and the perception of the coping behaviour of the patient, the severity of the complaint and the presence of social support on the other hand. From 20 general practitioners (GP), 15 consultations per GP were videotaped and analysed using the Roter Interaction Analysis System. Doctors and patients rated their perceptions on questionnaires. The finding was that doctors and patients used predominantly task-oriented (instrumental) behaviour, with some exceptions. With older patients and patients with low social support the GPs used more affective communication, mainly consisting of social talk and mutual agreement. In the case of complex problems, the GP paid special attention to the relationship with the patient. Within the domain of instrumental communication, some differences between doctor and patient were observed. Although doctors and patients exchanged a lot of information about medical issues, patients gave information about their lifestyle and emotions, which the doctors did not verbally explore. In consultations where the patient perceived the complaint as severe, he or she was more focussed on the medical content. When the GP considered psychosocial issues important, doctor and patient communicated about lifestyle, emotions and social relations. This doctor–patient correlation was not found when patients perceived their problem as psychosocial. r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Doctor–patient communication; Belgium; General practice
Introduction Research on doctor–patient communication reveals a number of positive and negative effects of the general practitioner’s (GP) communication style on outcomes such as satisfaction (Bertakis, Roter, & Putnam, 1991; Hall, Roter, & Katz, 1988; Suchman, Roter, Green, & Lipkin, 1993), health (Stewart, 1995) and compliance (Meeuwesen, Schaap, & Staak van der, 1991; Wartman, Morlock, Malitz, & Palm, 1983; Williams, Weinman, Dale, & Newman, 1995). Prevention of somatisation (Grol, De Maeseneer, Whitfield, & Mokkink, 1990a; *Corresponding author. Fax: +32-9-2404967. E-mail address: [email protected] (M. Deveugele).
Salmon, Peters & Stanley, 1999), recognition of mental disorders (Roter et al., 1997; Verhaak, 1988) and referral and prescription rates (Butler, 1998; Kaplan, Gandek, Greenfield, Rogers, & Ware, 1995; Ong, De Haes, Hoos, & Lammes, 1995; Simpson et al., 1991; Winefield, Murell, & Clifford, 1995) are also strongly related to the doctor–patient communication. As a consequence, researchers predominantly looked for what might explain variations in the communicative behaviour of both doctors and patients. One type of research has demonstrated that communication style might be determined by personal, social and cultural factors like age, sex, education and ethnicity of doctors and patients (Gulbrandsen, Fugelli, Sandvik, & Hjortdahl, 1998; Makoul, Arntson, & Schofield, 1995). Other researches focused on mutual interactions between doctor and
0277-9536/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 2 7 7 - 9 5 3 6 ( 0 1 ) 0 0 2 4 1 - 6
1246
M. Deveugele et al. / Social Science & Medicine 55 (2002) 1245–1253
patient (Charon, Greene, & Adelman, 1994, Street, 1991). The doctors’ communicative behaviour and strategies (like questioning or giving advice) were explained as responses to the communicative behaviour of patients. When the results of this research are combined, doctor–patient communication is viewed as a process of personal and mutual influence that unfolds according to the characteristics of both participants and the interactive processes (Street, 1991,1992). A typical feature of consultations in general practice is that doctors deal with somatic, psychological and social problems. Behavioural and medical research shows that even somatic problems are embedded in psychological and social factors (Grol et al., 1990a, b). Nevertheless, doctors are still trained to focus on medical problems and tend to underestimate the patients’ own perception of their health problems as a valid source of information (Bensing, 1992; Bensing & Dronkers, 1992; Verhaak, 1988). Patients consulting a doctor are able to give useful information on both medical and psychosocial aspects of their lives and they do have intentions and expectations about the outcome of the consultation (Salmon et al., 1999). Physicians, however, are often unaware of the cognitions and emotions of their patients. They tend to underestimate the relevance of psychosocial issues during the consultation (Verhaak, 1988). They also underestimate the severity of the complaint as perceived by the patient. Nevertheless, GPs may make judgements about the coping abilities of their patients, classifying them as copers (Kahn, Steeves, & Benoliel, 1994) and non-copers (Jones & Morrell, 1995). Copers were identified as ‘‘very capable’’, ‘‘a normal straightforward person’’ and ‘‘easy, having normal reactions’’. Non-copers were referred to as ‘‘worriers’’ and as ‘‘not good at coping’’. When GPs perceived their patients as good copers they tended to underestimate the psychosocial aspects of the problem and overestimate the medical issues. For non-copers the opposite was found. Doctors gave information about medical issues and about lifestyle to both types of patients, but they believed that the information was more valuable for copers. Doctors also estimated the support their patients received from their environment, and whether the environment was stressful (Ell, 1996). In our research, the main focus was whether these perceptions affect the communicative behaviour in consultations. More precisely, the aim of our study was to answer the following questions: What are the patterns of communication in general practice in Flanders? Do GPs’ and patients’ perceptions of the coping behaviour of the patients, the severity of the complaint and available social support influence their communication pattern? If so, what is the nature of these influences? Our research was a part of the Eurocommunication study (Brink-Muinen et al., 1999). The main objective of the larger study was to investigate
how the characteristics of various health care systems affect doctor–patient communication in general practice.
Method Participants To obtain a group of 20 GPs representative of those in Flanders, we sent a letter to a random sample of 150 GPs from a Flemish data base. The letter contained information about the aim, the background and the methodology of the study. Of the 20 GPs who responded, seven agreed to participate. Phone calls to non-responders revealed that the videotaping was the main obstacle. After some explanation four more agreed to participate. The group was increased to 20 by asking nine GPs amongst the co-workers of the Department of General Practice and Primary Care of Ghent University. Finally, our study group consisted of 12 men and eight women, with a mean age of 43 years. Thirteen GPs worked in single-handed practices, two worked in a duo, and five in group practices. Three had not followed vocational training; they graduated before vocational training was introduced in Flanders. Nineteen doctors had previously participated in research; 12 doctors had already been videotaped during their consultations and 12 were trainers. For each GP, consecutive patients were asked to participate and to give written consent, until 20 contacts could be videotaped. Home visits were excluded for logistic reasons. Twenty-five per cent of the patients refused to participate. Refusals were accepted without any attempt to persuade the patient. The mean age of the patients was 43 years; 39% were male and 61% female. Procedures Patients completed questionnaires before the consultation. The GPs were asked to complete a log sheet after each consultation. Of the 20 consultations videotaped per doctor only 15 were included in the analysis. Generally the first three consultations were not coded to avoid bias, possibly caused by the doctor’s adaptation to the video camera. Some videos could not be rated due to technical problems (e.g. unforeseen damage, a hardly audible conversation or a partially recorded consultation). The camera was fixed in a position enabling observation of both patient and doctor. The face of the doctor needed to be visible. Due to the fixed position of the camera, physical examinations were not in view, but the recorder was left on to enable analysis of the whole conversation and measure the total length of the consultation and the length of the physical examination.
M. Deveugele et al. / Social Science & Medicine 55 (2002) 1245–1253
Measurement instruments The doctors’ questionnaire After every consultation, the GPs registered the gender and the year of birth of the patient, how long the patient had been in his/her practice and the number of contacts between doctor and patient during the last year. In addition, they registered the reason for the encounter as expressed by the patient, the medical diagnosis and the perceived psychosocial background of the problem. For the perceptions of doctors and patients, we constructed a questionnaire of 11 items (Deveugele, 1998), each to be rated on a 5-point Likert scale to assess: severity of the problem, coping ability and social support (Jones & Morrell, 1995; Van Dulmen, Fennis, Mokkink, Velden van der, & Bleijenberg, 1994). To ascertain content validity, an expert group of doctors and psychologists looked at the terms and discussed the words used. The patients’ questionnaire The patients recorded their year of birth, sex, living situation, highest level of education, and the reason for the present consultation. They then completed 12 of the 42 items from the patient request form (Valori, Woloshynowych, Bellenger, Aluvihare, & Salmon, 1996) on which they rated on 5-point scales the importance they attached to somatic and psychosocial aspects of this consultation. Factor analysis revealed two subscales: a biomedical scale of six items and a psychosocial scale of four items. Cronbach’s alpha for biomedical aspects was 0.76 and for psychosocial aspects 0.64. We used the mean scores of the somatic and the psychosocial subscales. Health problems were coded according to the International Classification of Primary Care (ICPC) (Lamberts & Woods, 1987). Ratings of coping, severity and
1247
social support items (10 of the 11 items rated also by GPs) were made on 5-point Likert scales before the consultation. The doctor did not fill in the questionnaire before, because even if he knew the patient, the reason for the encounter was only revealed in the consultation. The timing of the patients’ ratings does not seem to matter. A formal check with an independent sample, consisting of 96 patients of 10 GPs, showed no difference between patients’ answers before and after the consultation (coping t ¼ 0:82; po0:001; severity t ¼ 0:72; po0:001; and social support t ¼ 0:95; po0:001). Consequently, patients made their ratings before. Observation protocol The videotapes were then analysed and scored for time management and communication. Time management variables were measured by means of a stopwatch: the length of the consultation, the length of the physical examination and interruptions, e.g. when somebody entered the room, when the GP left the room or when there was a phone call. These interruptions were subtracted from the total length of the consultation. Affective and instrumental behaviours were scored according to the Roter Interaction Analysis System (RIAS) (Roter, 1991). This system is well documented and widely used in the United States (Roter et al., 1995; Roter & Russel, 1994) and has been validated for use in Dutch settings (Bensing, 1991; Brink-Muinen et al., 1999). The system is designed to code the communicative behaviour of both doctors and patients. It distinguishes affective (socioemotional) and instrumental (task-oriented) behaviour, reflecting the care–cure distinction. The unit of analysis is the smallest meaningful string of words. All utterances are assigned to mutually exclusive categories. The categories are merged into 16 clusters, seven for affective and nine for instrumental behaviour (see Table 1).
Table 1 Roter interaction analysis system, affective and instrumental clusters Affective behaviour
Instrumental behaviour
Social behaviour: gives personal remarks, tells jokes/ laughs, shows approvalFdirect, gives complimentF general Agreement: shows agreement or understanding Paraphrase: paraphrases, checks for understanding
Giving directions: makes a transition, gives orientation, instruction
Verbal attention: shows empathy, legitimises, shows partnership and support Showing concern: showing concern or worry Reassurance: reassures, encourages or shows optimism Disagreement: shows disapprovalFdirect, shows criticismFgeneral
Asking clarification: asks for repetition, for understanding, for an opinion Asks questions: about the medical condition, the therapeutic regimen, requests for services Asks questions: on lifestyle/social context; psychosocial situation/feelings Gives information: about the medical condition, the therapeutic regimen Gives information: on lifestyle/social context, psychosocial situation/ feelings Counsels or directs behaviour: about the medical condition, the therapeutic regimen Counsels or directs behaviour: on lifestyle/social context, psychosocial situation/feelings Other instrumental or unintelligible utterances.
1248
M. Deveugele et al. / Social Science & Medicine 55 (2002) 1245–1253
Instrumental behaviour was further subdivided, according to Roter et al. (1997), into: psychosocial talk, e.g. asking questions, giving information and counselling (only GP) about issues of lifestyle, social context, psychosocial situation and feelings, and medical talk, e.g. asking questions, giving information and counselling (only GP) about medical and therapeutic issues. Four observers were trained until they reached almost identical ratings of the videotaped consultations. The interrater reliability was measured by calculating Pearson’s correlation coefficients for the number of utterances observed in each category per patient, by pairs of observers. Twenty patients, two of each GP, were observed by two observers. Interrater reliability was not calculated for clusters occurring less frequently than 2%, because for under-utilised clusters it could not be done confidently due to the skewing effect. The average interrater reliability was 0.82 (range 0.68–0.91).
did not differ significantly concerning the total amount of affective (F ¼ 0:51; df 2,295) or instrumental behaviour (F ¼ 0:59; df 2,295). Factor analysis of the doctor’s questionnaire revealed three factors (see Table 2): coping with a Cronbach’s alpha of 0.88, severity of the complaint 0.72 and social support 0.78. Factor analysis of the patients’ questionnaire also revealed three factors: coping with a Cronbach’s alpha of 0.75, severity with 0.71 and social support with 0.37. Despite the low alpha, patients’ ratings of social support were included in subsequent analyses because the two items correlated highly with the third factor and this factor was similar to the GPs’. Likewise, in order to keep GP and patient questionnaires comparable, doctors’ item 6 was excluded. Consultation time The consultations had a mean length of 13.6 min (SD 5.8). The physical examination took approximately 3.4 min (SD 3.2).
Statistical analysis Doctor–patient communication The factor analysis of the questionnaires about perception was performed looking for factors with eigenvalue >1 using a Varimax rotation. The Wilcoxon Signed Ranks test was used to compare the numbers of utterances of doctors and patients. Finally, eight linear regression analyses were conducted to determine predictors of affective and instrumental communications, as well as the two subgroups of instrumental, namely, psychosocial and medical talk.
Results Preliminary analyses In order to know the extent to which the GPs were representative of the entire GP population, a comparison was made with the study population of the Task Profile Study, an international study in Europe with representative samples of GPs (N ¼ 511) of different European countries (Boerma, Zee van der, & Fleming, 1997). Some bias was found with respect to age (our GPs were some years older), workload (they saw fewer patients a day) and vocational training (two of our GPs did not have vocational training). Compared with the research of De Maeseneer (1989), the most recent in the Flemish area, our study group of doctors and patients was very similar with regard to gender, age distribution, health problem, and consultation length. Since our study group consisted of randomly selected doctors as well as co-workers of the department, a comparison between both groups was made. The groups
Affective and instrumental behaviour as classified by the RIAS can be seen in Table 3. The mean number of utterances of doctors was 143.39 (SD 75.12) and of patients was 119.75 (SD 70.28). A Wilcoxon Signed Ranks test revealed a significant difference between the total number of utterances of doctor and patient (instrumental behaviour Z ¼ 7:08; po0:000; affective behaviour Z ¼ 6:83; po0:001). Most of the affective behaviour of patients and doctors had to do with obtaining agreement and social talk. Paraphrasing was the third most frequent type of affective behaviour of the doctors. Verbal attention, showing concern and reassurance counted for only 2.3% of the total utterances. Regarding instrumental behaviour GPs gave directions and information about medical issues. Patients gave information on the medical issues but also on their lifestyle and feelings. The Wilcoxon Signed Ranks test was conducted on absolute numbers within categories containing more than 2% of the total utterances of doctors or patients. There was a significant difference between the total number of utterances of doctor and patient. Doctors talked more in the categories social talk (Z ¼ 5:73), paraphrase (Z ¼ 13:03), gives directions (Z ¼ 14:66), asks questions about medical issues (Z ¼ 12:54), asks questions about lifestyle (Z ¼ 11:42) and counsels about medical issues (Z ¼ 12:89). Patients talked more in the categories gives information about lifestyle (Z=12.33), and other (Z=3.94). There was no significant difference for the categories agreement and gives information on medical issues.
1249
M. Deveugele et al. / Social Science & Medicine 55 (2002) 1245–1253 Table 2 Principal component coefficients of the doctors’ and patients’ responses to coping, severity and social support questions Doctors’ questionnaire
Factor 1
Factor 2
Factor 3
The patient is able to solve the problem The patient can cope with the problem in an emotional adequate way The patient will have social support from his environment for this problem The problem burdens the patient’s work or living situation The patient has physical pain from this complaint. The patient exaggerates the problem This problem is serious In general, the patient has sufficient skills to solve problems In general, the patient is able to cope in an emotional adequate way Global assessment of the coping devices of the patient Global assessment of the social support the patient gets from the environment.
0.83 0.74 0.10 0.07 0.01 0.50 0.23 0.85 0.83 0.84 0.21
0.08 0.33 0.00 0.83 0.66 0.28 0.74 0.00 0.16 0.05 0.07
0.17 0.11 0.92 0.06 0.00 0.16 0.02 0.13 0.03 0.12 0.89
Explained variance
41%
19%
18%
Patient’s questionnaire. I am able to solve this problem I can cope in an emotional way with this problem My environment will help me to solve this problem This problem is a burden for my work or my living situation I have physical pain I think this problem is real serious In general, I have skills to solve problems In general, I can emotionally deal with problems Global assessment of my coping behaviour Global assessment of the social support of my environment
0.38 0.48 0.00 0.25 0.03 0.10 0.84 0.88 0.76 0.23
0.28 0.24 0.04 0.78 0.79 0.75 0.04 0.06 0.04 0.05
0.33 0.34 0.86 0.08 0.02 0.08 0.08 0.08 0.03 0.66
Explained variance
25%
20%
16%
Table 3 Mean percentages and SDs of affective and instrumental behaviour (according to RIAS) of doctors and patients as a percentage of the total count of utterances in 299 consultations Utterances
General practitioner (%) (SD)
Patient (%) (SD)
Affective behaviour Social behaviour Agreement Paraphrase Verbal attention Shows concern Reassurance Disagreement Total affective behaviour
8.8 14.2 4.4 0.9 0.2 1.2 0.1 29.8
(8.8) (8.8) (3.3) (1.4) (0.5) (1.9) (0.6)
8.9 18.2 1.5 0.0 0.5 0.2 0.1 29.4
(10.2) (9.5) (1.6) (0.4) (1.2) (0.7) (0.4)
Instrumental behaviour Gives directions Asks clarifications Asks questionsFmedical Asks questionsFlifestyle Gives informationFmedical Gives informationFlifestyle CounselsFmedical CounselsFlifestyle Other
10.5 0.4 9.5 3.8 27.0 6.3 4.1 0.8 7.8
(5.6) (0.9) (6.0) (3.6) (12.9) (7.6) (4.7) (2.1) (6.7)
0.7 0.3 3.9 1.1 33.5 22.7 0.0 0.0 8.4
(1.9) (0.6) (3.5) (2.3) (16.1) (17.3) (0.3) (1.0) (7.6)
Total instrumental behaviour
70.2
70.6
1250
M. Deveugele et al. / Social Science & Medicine 55 (2002) 1245–1253
Eight regression analyses were conducted. Older patients used more instrumental communication but they obtained more affective talk from the doctor. Doctors who perceived more social support to be available for the patient provided less affective and instrumental talk; likewise, these patients used less affective talk. When doctors perceived better coping in the patient, the patient used more affective speech. Yet patients used more instrumental communication if they considered their complaint as severe. Patients who rated their social support as high received and gave more instrumental talk (Table 4). Further analysis of the instrumental communication according to medical and psychosocial talk showed no significant predictors of medical talk. For psychosocial talk, the following were significant: when the doctor perceived the complaint as having more psychosocial impact, he talked more about the lifestyle and emotions of the patient. The same applied to the patient. Concerning severity, there was a positive correlation between the patient’s perception of severity and psychosocial talk by the doctor. As for social support, doctors who perceived their patients as having low social support gave more psychosocial talk; for the patient the opposite was found. Factors such as the sex of the doctor and patient, acquaintance with the patient (the number of years the patient was known and the number of visits in the last year) and the importance of medical and psychosocial aspects as mentioned by the patient did not influence the communicative behaviour of the doctor or the patient (Table 5)
Discussion The main finding of the study was that the doctor and patient used predominantly task-oriented (instrumental) communication. The ratio of affective to instrumental behaviour for both was 1 to 2.5. This ratio was also found in other studies and other countries (Brinck-Muinen et al., 1999). Although the general pattern seemed similar, some specific differences within the instrumental pattern could be observed. First, the doctor asked the questions, the patient generally did not. This is in agreement with other research findings (Roter et al., 1991; Street, 1991). Table 5 Beta coefficients for predictors of psychosocial talk in 299 doctor–patient consultationsa F (df)
Psychosocial GP 3.129 (9.93)*
Psychosocial P 2.483(13,99)*
Predictors Coping GP Severity GP Support GP Psychosoc GP Coping P Severity P Support P Imp psychosoc P Imp medical P
0.036 0.037 0.327* 0.277* 0.134 0.137* 0.259 0.076 0.184
0.028 0.071* 0.179 0.324* 0.113 0.100 0.235* 0.091 0.155
a Note: N (299) po0:5; GP=general practitioner, P=patient.
Table 4 Beta coefficients for predictors of affective and instrumental behaviour in 299 doctor–patient consultationsa GP communication
Patient communication
F (df)
Aff. beh. 1.997 (14,87)*
Instr. beh. 2.285 (14,87)*
Aff. beh. 2.129 (14,87)*
Instr. beh. 2.285 (14,87)*
Independent var. Sex GP Sex P Age P Visits last year Years patient Importance of somatic to P Importance of psychosoc. to P Coping GP Severity GP Support GP Psychosoc GP Coping P Severity P Support P
0.013 0.034 0.342* 0.183 0.127 0.151 0.109 0.004 0.082 0.343* 0.139 0.148 0.155 0.214
0.153 0.171 0.217 0.149 0.124 0.136 0.115 0.063 0.081 0.413* 0.045 0.111 0.169 0.401*
0.182 0.128 0.197 0.115 0.184 0.062 0.024 0.275* 0.038 0.476* 0.131 0.118 0.120 0.119
0.032 0.086 0.327* 0.157 0.082 0.073 0.029 0.004 0.099 0.196 0.174 0.162 0.235* 0.235*
a
Note: N (299) po0:05; GP=general practitioner, P=patient.
M. Deveugele et al. / Social Science & Medicine 55 (2002) 1245–1253
Patients are rather passive and accept that the doctor is leading the consultation. This was not different in our study, although the doctors presumably were more interested in communication. Second, the doctor gave a lot of information on medical issues but less on lifestyle. Patients gave information on both. Moreover, the doctors did not verbally react to patients’ utterances about lifestyle, social context, psychosocial situations and feelings. Perhaps the doctor was listening to and reflecting on this contextual information without making specific statements. However, the patient may not always know that the message had been received and understood (Arborelius, Bremberg, & Timpka, 1991). Third, both doctors and patients seldom asked for clarification. Since most communication contains a lot of ambiguities, not asking for clarification can easily lead to misunderstanding. The affective behaviour that doctors and patients used in our study was limited to social talk and agreement (the ‘‘relationship’’ categories) and to some extent to paraphrase. It is obvious that creating a good relationship is of major importance and a quality indicator of general practice (Bensing, 1991). One might wonder why in this study the affective behaviour was limited to two categories, and why concern and reassurance were not more common. Maybe social talk and agreement contribute to the necessary client binding required to maintain the relationship (Grol, Whitfield, & De Maeseneer, 1990b). With certain categories of patients, the communication pattern of doctors and patients was somewhat special. Older patients used more instrumental talk during the consultation, despite the fact that doctors used more affective communication to them. A possible explanation is that doctors compensate for their inability to really solve the complex medical problems of their older patients by caring for the relationship. Thinking of the adage ‘‘seldom to cure, often to comfort, always to care’’, the positive relation between affective behaviour and age could be a sign of good care for the elderly. On the other hand, patients might have been frustrated that they were not able to elicit instrumental talk from their GP. If patients were perceived as having low social support, the doctors used more affective communication (social talk and agreement). This kind of talk may have been offered to bolster the patient’s feeling of support. These same patients themselves used more affective talk. In this case there was good symmetry. Likewise, patients who rated their social support highly gave and received more instrumental talk. As far as the doctor’s perception of coping is concerned we cannot draw the same conclusion. Although Jones and Morrell (1995) revealed that perception of coping behaviour influences the commu-
1251
nicative behaviour of the doctor, our study did not confirm this conclusion. Moreover, patients perceived as good copers by their doctors showed more affective behaviour. It is not clear whether the physicians had identified copers before the consultation based on prior experience or whether the patient’s affective communication led to the good coping designation. The more the patients perceived their complaints as severe, the more instrumental behaviour they used. This is consistent with the literature showing that ill people talk more about medical issues and give more psychosocial information (instrumental behaviour). Another finding from the literature, namely that severely ill patients get less social talk from their doctors (Hall, Irish, Roter, Ehrlich, & Miller, 1994) was not confirmed in our study. We did not see any relation between psychosocial problems and communicative behaviour in the multiple regression analysis. However, bivariate correlations between perceptions of the importance of psychosocial factors and both categories of communication were positive and significant for GP and patient (available on request from the authors). Both may have lost some predictive power to patients’ sex and age in the multiple regression analyses. Limitations of the study The way we used the observation instrument RIAS explains some of the conclusions. The Roter instrument has its roots in instrumental tradition. It is an adaptation of Bales’ Interaction Analysis System that was based on problem-solving theories, focusing on verbal task-related behaviour. Not surprisingly, the instrument detects mainly instrumental behaviour. Affective behaviour has two main objectives: creating a good atmosphere and relationship (observed in the categories social talk and agreement) and providing information about the interaction-aspects of general practice (categories paraphrase, verbal attention, concern, reassurance and disagreement). These objectives are mainly communicated by non-verbal behaviour, which the RIAS does not explore. The RIAS does not register the interaction between doctor and patient, which is another limitation. Nothing can be said about the communicative influence of doctor and patient on each other. The system splits the communication into separate strings of words, which we did not put on a timetable. As a result, we do not have the possibility of looking at the contextual communicative development within the consultation. For example, social talk could have a different meaning at the beginning of the consultation (warming up, building relation) than at the end (client binding). Furthermore, there are limitations to the generalisation of our study. Although our patient group was
1252
M. Deveugele et al. / Social Science & Medicine 55 (2002) 1245–1253
representative of the patients consulting GPs in Flanders, our group of GPs was not. They were more interested and better-trained in communication skills than the average GP. They were older and had a lighter workload. Finally, a multilevel analysis would have provided more information. Our data were characterised by dependency of observational units; patients were nested within practices. Hierarchical analysis would have allowed determining the variance due to the doctor– patient interaction term. Multilevel analysis assumes sufficient units at each level of measuring. In this study, 20 GPs participated, which is a rather limited number for performing hierarchical analyses.
Conclusions This research revealed some important patterns in the communicative behaviour of doctors and patients. Doctors were task oriented; doing a good medical job was their main objective. In the encounter between GP and patient, the communicative behaviour of both was very similar: patients and doctors focused on medical issues. Nevertheless, 30% of the utterances were affective in nature. Spending time on social talk allowed doctors to estimate important psychosocial aspects of the presented complaint. It is known that the satisfaction of doctors and patients is more related to affective than to instrumental behaviour. The same counts for compliance of patients. The doctors in this research tended to have more affective reactions with some groups of patients, like the elderly and the patients perceived to have low social support.
Acknowledgements This study has been made possible by funding from the BIOMED-II research programme of the European Union (Contract No. BMH4-CT96-1515) and by the FWO-Belgium (F9885). The authors thank the central co-ordinators for the BIOMED program from NIVEL (The Netherlands): A. Van Den Brink-Muinen, P. Verhaak, J. Bensing and the national co-ordinators: L Gask, N Mead (UK), O. Bahr (Germany), A. Perez (Spain), V. Messerli, M. Peltenburg, L. Oppizzi (Switzerland). The authors thank the GPs who participated in this study.
References Arborelius, E., Bremberg, S., & Timpka, T. (1991). What is going on when the general practitioner doesn’t grasp the situation? Family Practice, 8(1), 3–9.
Bensing, J. M. (1991). Doctor–patient communication and the quality of care. An observation study into affective and instrumental behaviour in general practice. Utrecht: Boekbinderij Post. Bensing, J. M. (1992). Instrumental and affective behaviour in physician behaviour. Medical Care, 30(4), 283–298. Bensing, J. M., & Dronkers, J. (1992). Instrumental and affective aspects of physician behaviour. Medical Care, 30(4), 283–297. Bertakis, K. D., Roter, D., & Putnam, S. M. (1991). The relationship of physician medical interview style to patient satisfaction. The Journal of Family Practice, 32(2), 175–181. Boerma, W. G. W., Zee van der, J., & Fleming, D. M. (1997). Services profiles of general practitioners in Europe. British Journal of General Practice, 37, 481–486. Brink-Muinen van den, A., Verhaak, P. F. M., Bensing, J. M., Bahrs, O., Deveugele, M., Gask, L., Mead, N., LeivaFernandes, F., Perez, A., Messerli, V., Oppizzi, L., & Peltenburg, M. (1999). The Eurocommunication study. An international comparative study in sex European countries on doctor–patient communication in general practice. Utrecht: Drukkerij Anraad Nieuwegein. Butler, C. C. (1998). Understanding the culture of prescribing: Qualitative study of general practitioners’ and patients’ perceptions of antibiotics for sore throats. British Medical Journal, 317, 637–642. Charon, R., Greene, M. G., & Adelman, R. D. (1994). Multidimensional interaction analysis: A collaborative approach to the study of medical discourse. Social Science and Medicine, 39(7), 955–965. De Maeseneer, J. (1989). Huisartsgeneeskunde: een verkenning. Een explorerend onderzoek bij huisartsen-stageleiders aan de R.U. Gent. Proefschrift tot het verkrijgen van de graad van Geaggregeerde van het Hoger Onderwijs. Universiteit Gent. Deveugele, M. (1998). Percepties van huisartsen. Constructie van een vragenlijst. Interne publicatie. Ell, K. (1996). Social networks, social support and coping with serious illness: The family connection. Social Science and Medicine, 42(2), 173–183. Grol, R., De Maeseneer, J., Whitfield, M., & Mokkink, H. (1990a). Disease-centred versus patient-centred attitudes: Comparison of general practitioners in Belgium, Britain and The Netherlands. Family Practice, 7(2), 100–136. Grol, R., Whitfield, W., De Maeseneer, J., & Mokkink, M. (1990b). Attitudes to risk taking in medical decision making of British, Dutch and Belgian General Practitioners. The British Journal of General Practice, 40, 134–136. Gulbrandsen, P., Fugelli, P., Sandvik, L., & Hjortdahl, P. (1998). Influence of social problems on management in general practice: Multipractice questionnaire survey. British Medical Journal, 317, 2830. Hall, J. A., Irish, J. T., Roter, D. J., Ehrlich, C. M., & Miller, L. H. (1994). Satisfaction, gender and communication in medical visits. Medical Care, 32(12), 1216–1231. Hall, J. A., Roter, D. L., & Katz, N. C. (1988). Meta-analysis of correlations of provider behavior in medical encounters. Medical Care, 26(7), 657–675. Jones, I., & Morrell, D. (1995). General practitioners’ background knowledge of their patients. Family Practice, 12(1), 49–53.
M. Deveugele et al. / Social Science & Medicine 55 (2002) 1245–1253 Kahn, D. L., Steeves, R. H., & Benoliel, J. Q. (1994). Nurses’views of the coping of patients. Social Science and Medicine, 38(10), 1423–1430. Kaplan, S. H., Gandek, B., Greenfield, S., Rogers, W., & Ware, J. E. (1995). Patient and visit characteristics related to physicians’ participatory decision making style. Medical Care, 33(12), 1176–1187. Lamberts, H., & Woods, M. (1987). International classification of primary care (ICPC). Oxford: Oxford University Press. Makoul, G., Arntson, P., & Schofield, T. (1995). Health promotion in primary care: Physician–patient communication and decision making about prescription medications. Social Science and Medicine, 41(9), 1241–1254. Meeuwesen, L., Schaap, C., & Staak van der, C. (1991). Verbal analysis of doctor–patient communication. Social Science and Medicine, 32(10), 1143–1150. Ong, L. M. L., De Haes, J. C. J. M., Hoos, A. M., & Lammes, F. B. (1995). Doctor–patient communication: A review of the literature. Social Science and Medicine, 40(7), 903–918. Roter, D. L. (1991). The roter method of interaction process analysis. RIAS Manual. Baltimore: John Hopkins University. Roter, D. L., Hall, J. A., Kern, D., Barker, L. R., Cole, K. A., & Roca, R. P. (1995). Improving physicians’ interviewing skills and reducing patients’ emotional distress. Archives of Internal Medicine, 155, 1877–1884. Roter, D. L., & Russel, N. K. (1994). Validity of physician selfreport in tracking patient education objectives. Health Education Quarterly, 21(1), 27–38. Roter, D. L., Stewart, M., Putnam, S., Lipkin, M., Stiles, W., & Inui, T. (1997). Communication patterns of primary health care. The Journal of the American Medical Association, 277(4), 350–356. Salmon, P., Peters, S., & Stanley, I. (1999). Patients’ perceptions of medical explanation for somatisation disorders: Qualitative analysis. British Medical Journal, 318, 372–376. Simpson, M., Buckman, R., Stewart, M., Maguire, P., Lipkin, M., Novack, D., & Till, J. (1991). Doctor–patient commu-
1253
nication: The toronto consensus statement. British Medical Journal, 303, 1385–1387. Stewart, M. (1995). Effective physician–patient communication and health outcomes: A review. Canadian Medical Association Journal, 152(9), 1423–1433. Street, R. L. (1991). Information giving in medical consultations: The influence of patients’ communicative styles and personal characteristics. Social Science and Medicine, 32(5), 541–548. Street, R. L. (1992). Communicative styles and adaptations in physician–parent consultations. Social Science and Medicine, 34(10), 1155–1163. Suchman, A. L., Roter, D. L., Green, M., & Lipkin, M. (1993). Physician satisfaction with primary core office visits. Medical Care, 31(12), 1083–1092. Valori, R., Woloshynowych, M., Bellenger, N., Aluvihare, V., & Salmon, P. (1996). The patient request form: A way of measuring what patients want from their general practitioner. Van Dulmen, A. M., Fennis, J. F. M., Mokkink, G. H. A., Velden van der, H. G. M., & Bleijenberg, G. (1994). Doctors’ perception of patients’ cognitions and complaints in irritable bowel syndrome at an outpatient clinic. Journal of Psychosomatic Research, 38(6), 581–590. Verhaak, P. F. M. (1988). Detection of psychological complaints by general practitioners. Medical Care, 26(10), 1009– 1019. Wartman, S. A., Morlock, L. L., Malitz, F. E., & Palm, E. A. (1983). Patient understanding and satisfaction as predictors of compliance. Medical Care, 21(9), 886–891. Williams, S., Weinman, J., Dale, J., & Newman, S. (1995). Patient expectations: What do primary care patients want from the GP and how far does meeting expectations affect patient satisfaction? Family Practice, 12(2), 193–201. Winefield, H. R., Murell, T. G., & Clifford, J. (1995). Process and outcomes in general practice consultations: Problems in defining high quality care. Social Science and Medicine, 41(7), 969–975.
Is GP–patient communication related to their perceptions of illness severity, coping and social support? M. Deveugele*, A. Derese, J. De Maeseneer Department of General Practice and Primary Health Care, University of Ghent, UZ-1K3, De Pintelaan 185, 9000 Gent, Belgium
Abstract The aim of the study was to explore the relationship between the communicative behaviour of general practitioner and patient on the one hand and the perception of the coping behaviour of the patient, the severity of the complaint and the presence of social support on the other hand. From 20 general practitioners (GP), 15 consultations per GP were videotaped and analysed using the Roter Interaction Analysis System. Doctors and patients rated their perceptions on questionnaires. The finding was that doctors and patients used predominantly task-oriented (instrumental) behaviour, with some exceptions. With older patients and patients with low social support the GPs used more affective communication, mainly consisting of social talk and mutual agreement. In the case of complex problems, the GP paid special attention to the relationship with the patient. Within the domain of instrumental communication, some differences between doctor and patient were observed. Although doctors and patients exchanged a lot of information about medical issues, patients gave information about their lifestyle and emotions, which the doctors did not verbally explore. In consultations where the patient perceived the complaint as severe, he or she was more focussed on the medical content. When the GP considered psychosocial issues important, doctor and patient communicated about lifestyle, emotions and social relations. This doctor–patient correlation was not found when patients perceived their problem as psychosocial. r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Doctor–patient communication; Belgium; General practice
Introduction Research on doctor–patient communication reveals a number of positive and negative effects of the general practitioner’s (GP) communication style on outcomes such as satisfaction (Bertakis, Roter, & Putnam, 1991; Hall, Roter, & Katz, 1988; Suchman, Roter, Green, & Lipkin, 1993), health (Stewart, 1995) and compliance (Meeuwesen, Schaap, & Staak van der, 1991; Wartman, Morlock, Malitz, & Palm, 1983; Williams, Weinman, Dale, & Newman, 1995). Prevention of somatisation (Grol, De Maeseneer, Whitfield, & Mokkink, 1990a; *Corresponding author. Fax: +32-9-2404967. E-mail address: [email protected] (M. Deveugele).
Salmon, Peters & Stanley, 1999), recognition of mental disorders (Roter et al., 1997; Verhaak, 1988) and referral and prescription rates (Butler, 1998; Kaplan, Gandek, Greenfield, Rogers, & Ware, 1995; Ong, De Haes, Hoos, & Lammes, 1995; Simpson et al., 1991; Winefield, Murell, & Clifford, 1995) are also strongly related to the doctor–patient communication. As a consequence, researchers predominantly looked for what might explain variations in the communicative behaviour of both doctors and patients. One type of research has demonstrated that communication style might be determined by personal, social and cultural factors like age, sex, education and ethnicity of doctors and patients (Gulbrandsen, Fugelli, Sandvik, & Hjortdahl, 1998; Makoul, Arntson, & Schofield, 1995). Other researches focused on mutual interactions between doctor and
0277-9536/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 2 7 7 - 9 5 3 6 ( 0 1 ) 0 0 2 4 1 - 6
1246
M. Deveugele et al. / Social Science & Medicine 55 (2002) 1245–1253
patient (Charon, Greene, & Adelman, 1994, Street, 1991). The doctors’ communicative behaviour and strategies (like questioning or giving advice) were explained as responses to the communicative behaviour of patients. When the results of this research are combined, doctor–patient communication is viewed as a process of personal and mutual influence that unfolds according to the characteristics of both participants and the interactive processes (Street, 1991,1992). A typical feature of consultations in general practice is that doctors deal with somatic, psychological and social problems. Behavioural and medical research shows that even somatic problems are embedded in psychological and social factors (Grol et al., 1990a, b). Nevertheless, doctors are still trained to focus on medical problems and tend to underestimate the patients’ own perception of their health problems as a valid source of information (Bensing, 1992; Bensing & Dronkers, 1992; Verhaak, 1988). Patients consulting a doctor are able to give useful information on both medical and psychosocial aspects of their lives and they do have intentions and expectations about the outcome of the consultation (Salmon et al., 1999). Physicians, however, are often unaware of the cognitions and emotions of their patients. They tend to underestimate the relevance of psychosocial issues during the consultation (Verhaak, 1988). They also underestimate the severity of the complaint as perceived by the patient. Nevertheless, GPs may make judgements about the coping abilities of their patients, classifying them as copers (Kahn, Steeves, & Benoliel, 1994) and non-copers (Jones & Morrell, 1995). Copers were identified as ‘‘very capable’’, ‘‘a normal straightforward person’’ and ‘‘easy, having normal reactions’’. Non-copers were referred to as ‘‘worriers’’ and as ‘‘not good at coping’’. When GPs perceived their patients as good copers they tended to underestimate the psychosocial aspects of the problem and overestimate the medical issues. For non-copers the opposite was found. Doctors gave information about medical issues and about lifestyle to both types of patients, but they believed that the information was more valuable for copers. Doctors also estimated the support their patients received from their environment, and whether the environment was stressful (Ell, 1996). In our research, the main focus was whether these perceptions affect the communicative behaviour in consultations. More precisely, the aim of our study was to answer the following questions: What are the patterns of communication in general practice in Flanders? Do GPs’ and patients’ perceptions of the coping behaviour of the patients, the severity of the complaint and available social support influence their communication pattern? If so, what is the nature of these influences? Our research was a part of the Eurocommunication study (Brink-Muinen et al., 1999). The main objective of the larger study was to investigate
how the characteristics of various health care systems affect doctor–patient communication in general practice.
Method Participants To obtain a group of 20 GPs representative of those in Flanders, we sent a letter to a random sample of 150 GPs from a Flemish data base. The letter contained information about the aim, the background and the methodology of the study. Of the 20 GPs who responded, seven agreed to participate. Phone calls to non-responders revealed that the videotaping was the main obstacle. After some explanation four more agreed to participate. The group was increased to 20 by asking nine GPs amongst the co-workers of the Department of General Practice and Primary Care of Ghent University. Finally, our study group consisted of 12 men and eight women, with a mean age of 43 years. Thirteen GPs worked in single-handed practices, two worked in a duo, and five in group practices. Three had not followed vocational training; they graduated before vocational training was introduced in Flanders. Nineteen doctors had previously participated in research; 12 doctors had already been videotaped during their consultations and 12 were trainers. For each GP, consecutive patients were asked to participate and to give written consent, until 20 contacts could be videotaped. Home visits were excluded for logistic reasons. Twenty-five per cent of the patients refused to participate. Refusals were accepted without any attempt to persuade the patient. The mean age of the patients was 43 years; 39% were male and 61% female. Procedures Patients completed questionnaires before the consultation. The GPs were asked to complete a log sheet after each consultation. Of the 20 consultations videotaped per doctor only 15 were included in the analysis. Generally the first three consultations were not coded to avoid bias, possibly caused by the doctor’s adaptation to the video camera. Some videos could not be rated due to technical problems (e.g. unforeseen damage, a hardly audible conversation or a partially recorded consultation). The camera was fixed in a position enabling observation of both patient and doctor. The face of the doctor needed to be visible. Due to the fixed position of the camera, physical examinations were not in view, but the recorder was left on to enable analysis of the whole conversation and measure the total length of the consultation and the length of the physical examination.
M. Deveugele et al. / Social Science & Medicine 55 (2002) 1245–1253
Measurement instruments The doctors’ questionnaire After every consultation, the GPs registered the gender and the year of birth of the patient, how long the patient had been in his/her practice and the number of contacts between doctor and patient during the last year. In addition, they registered the reason for the encounter as expressed by the patient, the medical diagnosis and the perceived psychosocial background of the problem. For the perceptions of doctors and patients, we constructed a questionnaire of 11 items (Deveugele, 1998), each to be rated on a 5-point Likert scale to assess: severity of the problem, coping ability and social support (Jones & Morrell, 1995; Van Dulmen, Fennis, Mokkink, Velden van der, & Bleijenberg, 1994). To ascertain content validity, an expert group of doctors and psychologists looked at the terms and discussed the words used. The patients’ questionnaire The patients recorded their year of birth, sex, living situation, highest level of education, and the reason for the present consultation. They then completed 12 of the 42 items from the patient request form (Valori, Woloshynowych, Bellenger, Aluvihare, & Salmon, 1996) on which they rated on 5-point scales the importance they attached to somatic and psychosocial aspects of this consultation. Factor analysis revealed two subscales: a biomedical scale of six items and a psychosocial scale of four items. Cronbach’s alpha for biomedical aspects was 0.76 and for psychosocial aspects 0.64. We used the mean scores of the somatic and the psychosocial subscales. Health problems were coded according to the International Classification of Primary Care (ICPC) (Lamberts & Woods, 1987). Ratings of coping, severity and
1247
social support items (10 of the 11 items rated also by GPs) were made on 5-point Likert scales before the consultation. The doctor did not fill in the questionnaire before, because even if he knew the patient, the reason for the encounter was only revealed in the consultation. The timing of the patients’ ratings does not seem to matter. A formal check with an independent sample, consisting of 96 patients of 10 GPs, showed no difference between patients’ answers before and after the consultation (coping t ¼ 0:82; po0:001; severity t ¼ 0:72; po0:001; and social support t ¼ 0:95; po0:001). Consequently, patients made their ratings before. Observation protocol The videotapes were then analysed and scored for time management and communication. Time management variables were measured by means of a stopwatch: the length of the consultation, the length of the physical examination and interruptions, e.g. when somebody entered the room, when the GP left the room or when there was a phone call. These interruptions were subtracted from the total length of the consultation. Affective and instrumental behaviours were scored according to the Roter Interaction Analysis System (RIAS) (Roter, 1991). This system is well documented and widely used in the United States (Roter et al., 1995; Roter & Russel, 1994) and has been validated for use in Dutch settings (Bensing, 1991; Brink-Muinen et al., 1999). The system is designed to code the communicative behaviour of both doctors and patients. It distinguishes affective (socioemotional) and instrumental (task-oriented) behaviour, reflecting the care–cure distinction. The unit of analysis is the smallest meaningful string of words. All utterances are assigned to mutually exclusive categories. The categories are merged into 16 clusters, seven for affective and nine for instrumental behaviour (see Table 1).
Table 1 Roter interaction analysis system, affective and instrumental clusters Affective behaviour
Instrumental behaviour
Social behaviour: gives personal remarks, tells jokes/ laughs, shows approvalFdirect, gives complimentF general Agreement: shows agreement or understanding Paraphrase: paraphrases, checks for understanding
Giving directions: makes a transition, gives orientation, instruction
Verbal attention: shows empathy, legitimises, shows partnership and support Showing concern: showing concern or worry Reassurance: reassures, encourages or shows optimism Disagreement: shows disapprovalFdirect, shows criticismFgeneral
Asking clarification: asks for repetition, for understanding, for an opinion Asks questions: about the medical condition, the therapeutic regimen, requests for services Asks questions: on lifestyle/social context; psychosocial situation/feelings Gives information: about the medical condition, the therapeutic regimen Gives information: on lifestyle/social context, psychosocial situation/ feelings Counsels or directs behaviour: about the medical condition, the therapeutic regimen Counsels or directs behaviour: on lifestyle/social context, psychosocial situation/feelings Other instrumental or unintelligible utterances.
1248
M. Deveugele et al. / Social Science & Medicine 55 (2002) 1245–1253
Instrumental behaviour was further subdivided, according to Roter et al. (1997), into: psychosocial talk, e.g. asking questions, giving information and counselling (only GP) about issues of lifestyle, social context, psychosocial situation and feelings, and medical talk, e.g. asking questions, giving information and counselling (only GP) about medical and therapeutic issues. Four observers were trained until they reached almost identical ratings of the videotaped consultations. The interrater reliability was measured by calculating Pearson’s correlation coefficients for the number of utterances observed in each category per patient, by pairs of observers. Twenty patients, two of each GP, were observed by two observers. Interrater reliability was not calculated for clusters occurring less frequently than 2%, because for under-utilised clusters it could not be done confidently due to the skewing effect. The average interrater reliability was 0.82 (range 0.68–0.91).
did not differ significantly concerning the total amount of affective (F ¼ 0:51; df 2,295) or instrumental behaviour (F ¼ 0:59; df 2,295). Factor analysis of the doctor’s questionnaire revealed three factors (see Table 2): coping with a Cronbach’s alpha of 0.88, severity of the complaint 0.72 and social support 0.78. Factor analysis of the patients’ questionnaire also revealed three factors: coping with a Cronbach’s alpha of 0.75, severity with 0.71 and social support with 0.37. Despite the low alpha, patients’ ratings of social support were included in subsequent analyses because the two items correlated highly with the third factor and this factor was similar to the GPs’. Likewise, in order to keep GP and patient questionnaires comparable, doctors’ item 6 was excluded. Consultation time The consultations had a mean length of 13.6 min (SD 5.8). The physical examination took approximately 3.4 min (SD 3.2).
Statistical analysis Doctor–patient communication The factor analysis of the questionnaires about perception was performed looking for factors with eigenvalue >1 using a Varimax rotation. The Wilcoxon Signed Ranks test was used to compare the numbers of utterances of doctors and patients. Finally, eight linear regression analyses were conducted to determine predictors of affective and instrumental communications, as well as the two subgroups of instrumental, namely, psychosocial and medical talk.
Results Preliminary analyses In order to know the extent to which the GPs were representative of the entire GP population, a comparison was made with the study population of the Task Profile Study, an international study in Europe with representative samples of GPs (N ¼ 511) of different European countries (Boerma, Zee van der, & Fleming, 1997). Some bias was found with respect to age (our GPs were some years older), workload (they saw fewer patients a day) and vocational training (two of our GPs did not have vocational training). Compared with the research of De Maeseneer (1989), the most recent in the Flemish area, our study group of doctors and patients was very similar with regard to gender, age distribution, health problem, and consultation length. Since our study group consisted of randomly selected doctors as well as co-workers of the department, a comparison between both groups was made. The groups
Affective and instrumental behaviour as classified by the RIAS can be seen in Table 3. The mean number of utterances of doctors was 143.39 (SD 75.12) and of patients was 119.75 (SD 70.28). A Wilcoxon Signed Ranks test revealed a significant difference between the total number of utterances of doctor and patient (instrumental behaviour Z ¼ 7:08; po0:000; affective behaviour Z ¼ 6:83; po0:001). Most of the affective behaviour of patients and doctors had to do with obtaining agreement and social talk. Paraphrasing was the third most frequent type of affective behaviour of the doctors. Verbal attention, showing concern and reassurance counted for only 2.3% of the total utterances. Regarding instrumental behaviour GPs gave directions and information about medical issues. Patients gave information on the medical issues but also on their lifestyle and feelings. The Wilcoxon Signed Ranks test was conducted on absolute numbers within categories containing more than 2% of the total utterances of doctors or patients. There was a significant difference between the total number of utterances of doctor and patient. Doctors talked more in the categories social talk (Z ¼ 5:73), paraphrase (Z ¼ 13:03), gives directions (Z ¼ 14:66), asks questions about medical issues (Z ¼ 12:54), asks questions about lifestyle (Z ¼ 11:42) and counsels about medical issues (Z ¼ 12:89). Patients talked more in the categories gives information about lifestyle (Z=12.33), and other (Z=3.94). There was no significant difference for the categories agreement and gives information on medical issues.
1249
M. Deveugele et al. / Social Science & Medicine 55 (2002) 1245–1253 Table 2 Principal component coefficients of the doctors’ and patients’ responses to coping, severity and social support questions Doctors’ questionnaire
Factor 1
Factor 2
Factor 3
The patient is able to solve the problem The patient can cope with the problem in an emotional adequate way The patient will have social support from his environment for this problem The problem burdens the patient’s work or living situation The patient has physical pain from this complaint. The patient exaggerates the problem This problem is serious In general, the patient has sufficient skills to solve problems In general, the patient is able to cope in an emotional adequate way Global assessment of the coping devices of the patient Global assessment of the social support the patient gets from the environment.
0.83 0.74 0.10 0.07 0.01 0.50 0.23 0.85 0.83 0.84 0.21
0.08 0.33 0.00 0.83 0.66 0.28 0.74 0.00 0.16 0.05 0.07
0.17 0.11 0.92 0.06 0.00 0.16 0.02 0.13 0.03 0.12 0.89
Explained variance
41%
19%
18%
Patient’s questionnaire. I am able to solve this problem I can cope in an emotional way with this problem My environment will help me to solve this problem This problem is a burden for my work or my living situation I have physical pain I think this problem is real serious In general, I have skills to solve problems In general, I can emotionally deal with problems Global assessment of my coping behaviour Global assessment of the social support of my environment
0.38 0.48 0.00 0.25 0.03 0.10 0.84 0.88 0.76 0.23
0.28 0.24 0.04 0.78 0.79 0.75 0.04 0.06 0.04 0.05
0.33 0.34 0.86 0.08 0.02 0.08 0.08 0.08 0.03 0.66
Explained variance
25%
20%
16%
Table 3 Mean percentages and SDs of affective and instrumental behaviour (according to RIAS) of doctors and patients as a percentage of the total count of utterances in 299 consultations Utterances
General practitioner (%) (SD)
Patient (%) (SD)
Affective behaviour Social behaviour Agreement Paraphrase Verbal attention Shows concern Reassurance Disagreement Total affective behaviour
8.8 14.2 4.4 0.9 0.2 1.2 0.1 29.8
(8.8) (8.8) (3.3) (1.4) (0.5) (1.9) (0.6)
8.9 18.2 1.5 0.0 0.5 0.2 0.1 29.4
(10.2) (9.5) (1.6) (0.4) (1.2) (0.7) (0.4)
Instrumental behaviour Gives directions Asks clarifications Asks questionsFmedical Asks questionsFlifestyle Gives informationFmedical Gives informationFlifestyle CounselsFmedical CounselsFlifestyle Other
10.5 0.4 9.5 3.8 27.0 6.3 4.1 0.8 7.8
(5.6) (0.9) (6.0) (3.6) (12.9) (7.6) (4.7) (2.1) (6.7)
0.7 0.3 3.9 1.1 33.5 22.7 0.0 0.0 8.4
(1.9) (0.6) (3.5) (2.3) (16.1) (17.3) (0.3) (1.0) (7.6)
Total instrumental behaviour
70.2
70.6
1250
M. Deveugele et al. / Social Science & Medicine 55 (2002) 1245–1253
Eight regression analyses were conducted. Older patients used more instrumental communication but they obtained more affective talk from the doctor. Doctors who perceived more social support to be available for the patient provided less affective and instrumental talk; likewise, these patients used less affective talk. When doctors perceived better coping in the patient, the patient used more affective speech. Yet patients used more instrumental communication if they considered their complaint as severe. Patients who rated their social support as high received and gave more instrumental talk (Table 4). Further analysis of the instrumental communication according to medical and psychosocial talk showed no significant predictors of medical talk. For psychosocial talk, the following were significant: when the doctor perceived the complaint as having more psychosocial impact, he talked more about the lifestyle and emotions of the patient. The same applied to the patient. Concerning severity, there was a positive correlation between the patient’s perception of severity and psychosocial talk by the doctor. As for social support, doctors who perceived their patients as having low social support gave more psychosocial talk; for the patient the opposite was found. Factors such as the sex of the doctor and patient, acquaintance with the patient (the number of years the patient was known and the number of visits in the last year) and the importance of medical and psychosocial aspects as mentioned by the patient did not influence the communicative behaviour of the doctor or the patient (Table 5)
Discussion The main finding of the study was that the doctor and patient used predominantly task-oriented (instrumental) communication. The ratio of affective to instrumental behaviour for both was 1 to 2.5. This ratio was also found in other studies and other countries (Brinck-Muinen et al., 1999). Although the general pattern seemed similar, some specific differences within the instrumental pattern could be observed. First, the doctor asked the questions, the patient generally did not. This is in agreement with other research findings (Roter et al., 1991; Street, 1991). Table 5 Beta coefficients for predictors of psychosocial talk in 299 doctor–patient consultationsa F (df)
Psychosocial GP 3.129 (9.93)*
Psychosocial P 2.483(13,99)*
Predictors Coping GP Severity GP Support GP Psychosoc GP Coping P Severity P Support P Imp psychosoc P Imp medical P
0.036 0.037 0.327* 0.277* 0.134 0.137* 0.259 0.076 0.184
0.028 0.071* 0.179 0.324* 0.113 0.100 0.235* 0.091 0.155
a Note: N (299) po0:5; GP=general practitioner, P=patient.
Table 4 Beta coefficients for predictors of affective and instrumental behaviour in 299 doctor–patient consultationsa GP communication
Patient communication
F (df)
Aff. beh. 1.997 (14,87)*
Instr. beh. 2.285 (14,87)*
Aff. beh. 2.129 (14,87)*
Instr. beh. 2.285 (14,87)*
Independent var. Sex GP Sex P Age P Visits last year Years patient Importance of somatic to P Importance of psychosoc. to P Coping GP Severity GP Support GP Psychosoc GP Coping P Severity P Support P
0.013 0.034 0.342* 0.183 0.127 0.151 0.109 0.004 0.082 0.343* 0.139 0.148 0.155 0.214
0.153 0.171 0.217 0.149 0.124 0.136 0.115 0.063 0.081 0.413* 0.045 0.111 0.169 0.401*
0.182 0.128 0.197 0.115 0.184 0.062 0.024 0.275* 0.038 0.476* 0.131 0.118 0.120 0.119
0.032 0.086 0.327* 0.157 0.082 0.073 0.029 0.004 0.099 0.196 0.174 0.162 0.235* 0.235*
a
Note: N (299) po0:05; GP=general practitioner, P=patient.
M. Deveugele et al. / Social Science & Medicine 55 (2002) 1245–1253
Patients are rather passive and accept that the doctor is leading the consultation. This was not different in our study, although the doctors presumably were more interested in communication. Second, the doctor gave a lot of information on medical issues but less on lifestyle. Patients gave information on both. Moreover, the doctors did not verbally react to patients’ utterances about lifestyle, social context, psychosocial situations and feelings. Perhaps the doctor was listening to and reflecting on this contextual information without making specific statements. However, the patient may not always know that the message had been received and understood (Arborelius, Bremberg, & Timpka, 1991). Third, both doctors and patients seldom asked for clarification. Since most communication contains a lot of ambiguities, not asking for clarification can easily lead to misunderstanding. The affective behaviour that doctors and patients used in our study was limited to social talk and agreement (the ‘‘relationship’’ categories) and to some extent to paraphrase. It is obvious that creating a good relationship is of major importance and a quality indicator of general practice (Bensing, 1991). One might wonder why in this study the affective behaviour was limited to two categories, and why concern and reassurance were not more common. Maybe social talk and agreement contribute to the necessary client binding required to maintain the relationship (Grol, Whitfield, & De Maeseneer, 1990b). With certain categories of patients, the communication pattern of doctors and patients was somewhat special. Older patients used more instrumental talk during the consultation, despite the fact that doctors used more affective communication to them. A possible explanation is that doctors compensate for their inability to really solve the complex medical problems of their older patients by caring for the relationship. Thinking of the adage ‘‘seldom to cure, often to comfort, always to care’’, the positive relation between affective behaviour and age could be a sign of good care for the elderly. On the other hand, patients might have been frustrated that they were not able to elicit instrumental talk from their GP. If patients were perceived as having low social support, the doctors used more affective communication (social talk and agreement). This kind of talk may have been offered to bolster the patient’s feeling of support. These same patients themselves used more affective talk. In this case there was good symmetry. Likewise, patients who rated their social support highly gave and received more instrumental talk. As far as the doctor’s perception of coping is concerned we cannot draw the same conclusion. Although Jones and Morrell (1995) revealed that perception of coping behaviour influences the commu-
1251
nicative behaviour of the doctor, our study did not confirm this conclusion. Moreover, patients perceived as good copers by their doctors showed more affective behaviour. It is not clear whether the physicians had identified copers before the consultation based on prior experience or whether the patient’s affective communication led to the good coping designation. The more the patients perceived their complaints as severe, the more instrumental behaviour they used. This is consistent with the literature showing that ill people talk more about medical issues and give more psychosocial information (instrumental behaviour). Another finding from the literature, namely that severely ill patients get less social talk from their doctors (Hall, Irish, Roter, Ehrlich, & Miller, 1994) was not confirmed in our study. We did not see any relation between psychosocial problems and communicative behaviour in the multiple regression analysis. However, bivariate correlations between perceptions of the importance of psychosocial factors and both categories of communication were positive and significant for GP and patient (available on request from the authors). Both may have lost some predictive power to patients’ sex and age in the multiple regression analyses. Limitations of the study The way we used the observation instrument RIAS explains some of the conclusions. The Roter instrument has its roots in instrumental tradition. It is an adaptation of Bales’ Interaction Analysis System that was based on problem-solving theories, focusing on verbal task-related behaviour. Not surprisingly, the instrument detects mainly instrumental behaviour. Affective behaviour has two main objectives: creating a good atmosphere and relationship (observed in the categories social talk and agreement) and providing information about the interaction-aspects of general practice (categories paraphrase, verbal attention, concern, reassurance and disagreement). These objectives are mainly communicated by non-verbal behaviour, which the RIAS does not explore. The RIAS does not register the interaction between doctor and patient, which is another limitation. Nothing can be said about the communicative influence of doctor and patient on each other. The system splits the communication into separate strings of words, which we did not put on a timetable. As a result, we do not have the possibility of looking at the contextual communicative development within the consultation. For example, social talk could have a different meaning at the beginning of the consultation (warming up, building relation) than at the end (client binding). Furthermore, there are limitations to the generalisation of our study. Although our patient group was
1252
M. Deveugele et al. / Social Science & Medicine 55 (2002) 1245–1253
representative of the patients consulting GPs in Flanders, our group of GPs was not. They were more interested and better-trained in communication skills than the average GP. They were older and had a lighter workload. Finally, a multilevel analysis would have provided more information. Our data were characterised by dependency of observational units; patients were nested within practices. Hierarchical analysis would have allowed determining the variance due to the doctor– patient interaction term. Multilevel analysis assumes sufficient units at each level of measuring. In this study, 20 GPs participated, which is a rather limited number for performing hierarchical analyses.
Conclusions This research revealed some important patterns in the communicative behaviour of doctors and patients. Doctors were task oriented; doing a good medical job was their main objective. In the encounter between GP and patient, the communicative behaviour of both was very similar: patients and doctors focused on medical issues. Nevertheless, 30% of the utterances were affective in nature. Spending time on social talk allowed doctors to estimate important psychosocial aspects of the presented complaint. It is known that the satisfaction of doctors and patients is more related to affective than to instrumental behaviour. The same counts for compliance of patients. The doctors in this research tended to have more affective reactions with some groups of patients, like the elderly and the patients perceived to have low social support.
Acknowledgements This study has been made possible by funding from the BIOMED-II research programme of the European Union (Contract No. BMH4-CT96-1515) and by the FWO-Belgium (F9885). The authors thank the central co-ordinators for the BIOMED program from NIVEL (The Netherlands): A. Van Den Brink-Muinen, P. Verhaak, J. Bensing and the national co-ordinators: L Gask, N Mead (UK), O. Bahr (Germany), A. Perez (Spain), V. Messerli, M. Peltenburg, L. Oppizzi (Switzerland). The authors thank the GPs who participated in this study.
References Arborelius, E., Bremberg, S., & Timpka, T. (1991). What is going on when the general practitioner doesn’t grasp the situation? Family Practice, 8(1), 3–9.
Bensing, J. M. (1991). Doctor–patient communication and the quality of care. An observation study into affective and instrumental behaviour in general practice. Utrecht: Boekbinderij Post. Bensing, J. M. (1992). Instrumental and affective behaviour in physician behaviour. Medical Care, 30(4), 283–298. Bensing, J. M., & Dronkers, J. (1992). Instrumental and affective aspects of physician behaviour. Medical Care, 30(4), 283–297. Bertakis, K. D., Roter, D., & Putnam, S. M. (1991). The relationship of physician medical interview style to patient satisfaction. The Journal of Family Practice, 32(2), 175–181. Boerma, W. G. W., Zee van der, J., & Fleming, D. M. (1997). Services profiles of general practitioners in Europe. British Journal of General Practice, 37, 481–486. Brink-Muinen van den, A., Verhaak, P. F. M., Bensing, J. M., Bahrs, O., Deveugele, M., Gask, L., Mead, N., LeivaFernandes, F., Perez, A., Messerli, V., Oppizzi, L., & Peltenburg, M. (1999). The Eurocommunication study. An international comparative study in sex European countries on doctor–patient communication in general practice. Utrecht: Drukkerij Anraad Nieuwegein. Butler, C. C. (1998). Understanding the culture of prescribing: Qualitative study of general practitioners’ and patients’ perceptions of antibiotics for sore throats. British Medical Journal, 317, 637–642. Charon, R., Greene, M. G., & Adelman, R. D. (1994). Multidimensional interaction analysis: A collaborative approach to the study of medical discourse. Social Science and Medicine, 39(7), 955–965. De Maeseneer, J. (1989). Huisartsgeneeskunde: een verkenning. Een explorerend onderzoek bij huisartsen-stageleiders aan de R.U. Gent. Proefschrift tot het verkrijgen van de graad van Geaggregeerde van het Hoger Onderwijs. Universiteit Gent. Deveugele, M. (1998). Percepties van huisartsen. Constructie van een vragenlijst. Interne publicatie. Ell, K. (1996). Social networks, social support and coping with serious illness: The family connection. Social Science and Medicine, 42(2), 173–183. Grol, R., De Maeseneer, J., Whitfield, M., & Mokkink, H. (1990a). Disease-centred versus patient-centred attitudes: Comparison of general practitioners in Belgium, Britain and The Netherlands. Family Practice, 7(2), 100–136. Grol, R., Whitfield, W., De Maeseneer, J., & Mokkink, M. (1990b). Attitudes to risk taking in medical decision making of British, Dutch and Belgian General Practitioners. The British Journal of General Practice, 40, 134–136. Gulbrandsen, P., Fugelli, P., Sandvik, L., & Hjortdahl, P. (1998). Influence of social problems on management in general practice: Multipractice questionnaire survey. British Medical Journal, 317, 2830. Hall, J. A., Irish, J. T., Roter, D. J., Ehrlich, C. M., & Miller, L. H. (1994). Satisfaction, gender and communication in medical visits. Medical Care, 32(12), 1216–1231. Hall, J. A., Roter, D. L., & Katz, N. C. (1988). Meta-analysis of correlations of provider behavior in medical encounters. Medical Care, 26(7), 657–675. Jones, I., & Morrell, D. (1995). General practitioners’ background knowledge of their patients. Family Practice, 12(1), 49–53.
M. Deveugele et al. / Social Science & Medicine 55 (2002) 1245–1253 Kahn, D. L., Steeves, R. H., & Benoliel, J. Q. (1994). Nurses’views of the coping of patients. Social Science and Medicine, 38(10), 1423–1430. Kaplan, S. H., Gandek, B., Greenfield, S., Rogers, W., & Ware, J. E. (1995). Patient and visit characteristics related to physicians’ participatory decision making style. Medical Care, 33(12), 1176–1187. Lamberts, H., & Woods, M. (1987). International classification of primary care (ICPC). Oxford: Oxford University Press. Makoul, G., Arntson, P., & Schofield, T. (1995). Health promotion in primary care: Physician–patient communication and decision making about prescription medications. Social Science and Medicine, 41(9), 1241–1254. Meeuwesen, L., Schaap, C., & Staak van der, C. (1991). Verbal analysis of doctor–patient communication. Social Science and Medicine, 32(10), 1143–1150. Ong, L. M. L., De Haes, J. C. J. M., Hoos, A. M., & Lammes, F. B. (1995). Doctor–patient communication: A review of the literature. Social Science and Medicine, 40(7), 903–918. Roter, D. L. (1991). The roter method of interaction process analysis. RIAS Manual. Baltimore: John Hopkins University. Roter, D. L., Hall, J. A., Kern, D., Barker, L. R., Cole, K. A., & Roca, R. P. (1995). Improving physicians’ interviewing skills and reducing patients’ emotional distress. Archives of Internal Medicine, 155, 1877–1884. Roter, D. L., & Russel, N. K. (1994). Validity of physician selfreport in tracking patient education objectives. Health Education Quarterly, 21(1), 27–38. Roter, D. L., Stewart, M., Putnam, S., Lipkin, M., Stiles, W., & Inui, T. (1997). Communication patterns of primary health care. The Journal of the American Medical Association, 277(4), 350–356. Salmon, P., Peters, S., & Stanley, I. (1999). Patients’ perceptions of medical explanation for somatisation disorders: Qualitative analysis. British Medical Journal, 318, 372–376. Simpson, M., Buckman, R., Stewart, M., Maguire, P., Lipkin, M., Novack, D., & Till, J. (1991). Doctor–patient commu-
1253
nication: The toronto consensus statement. British Medical Journal, 303, 1385–1387. Stewart, M. (1995). Effective physician–patient communication and health outcomes: A review. Canadian Medical Association Journal, 152(9), 1423–1433. Street, R. L. (1991). Information giving in medical consultations: The influence of patients’ communicative styles and personal characteristics. Social Science and Medicine, 32(5), 541–548. Street, R. L. (1992). Communicative styles and adaptations in physician–parent consultations. Social Science and Medicine, 34(10), 1155–1163. Suchman, A. L., Roter, D. L., Green, M., & Lipkin, M. (1993). Physician satisfaction with primary core office visits. Medical Care, 31(12), 1083–1092. Valori, R., Woloshynowych, M., Bellenger, N., Aluvihare, V., & Salmon, P. (1996). The patient request form: A way of measuring what patients want from their general practitioner. Van Dulmen, A. M., Fennis, J. F. M., Mokkink, G. H. A., Velden van der, H. G. M., & Bleijenberg, G. (1994). Doctors’ perception of patients’ cognitions and complaints in irritable bowel syndrome at an outpatient clinic. Journal of Psychosomatic Research, 38(6), 581–590. Verhaak, P. F. M. (1988). Detection of psychological complaints by general practitioners. Medical Care, 26(10), 1009– 1019. Wartman, S. A., Morlock, L. L., Malitz, F. E., & Palm, E. A. (1983). Patient understanding and satisfaction as predictors of compliance. Medical Care, 21(9), 886–891. Williams, S., Weinman, J., Dale, J., & Newman, S. (1995). Patient expectations: What do primary care patients want from the GP and how far does meeting expectations affect patient satisfaction? Family Practice, 12(2), 193–201. Winefield, H. R., Murell, T. G., & Clifford, J. (1995). Process and outcomes in general practice consultations: Problems in defining high quality care. Social Science and Medicine, 41(7), 969–975.