A practical assessment of physician biopsychosocial performance

Abstract

Background: A biopsychosocial approach to care seems to improve patient satisfaction and health outcomes. Nevertheless, this approach is not widely practiced, possibly because its precepts have not been translated into observable skills.

Aim: To identify the skill components of a biopsychosocial consultation and develop an tool for their evaluation.

Methods: We approached three e-mail discussion groups of family physicians and pooled their responses to the question “what types of observed physician behavior would characterize a biopsychosocial consultation?” We received 35 responses describing 37 types of behavior, all of which seemed to cluster around one of three aspects: patient-centered interview; system-centered and family-centered approach to care; or problem-solving orientation. Using these categories, we developed a nine-item evaluation tool. We used the evaluation tool to score videotaped encounters of patients with two types of doctors: family physicians who were identified by peer ratings to have a highly biopsychosocial orientation (n = 9) or a highly biomedical approach (n = 4); and 44 general practitioners, before and after they had participated in a program that taught a biopsychosocial approach to care.

Results: The evaluation tool was found to demonstrate high reliability (α = 0.90) and acceptable interobserver variability. The average scores of the physicians with a highly biopsychosocial orientation were significantly higher than those of physicians with a highly biomedical approach. There were significant differences between the scores of the teaching-program participants before and after the program.

Conclusions: A biopsychosocial approach to patient care can be characterized using a valid and easy-to-apply evaluation tool.

Introduction

The recognition that some risk factors of disease are psychosocial rather than biomedical (Holmes & Rahe 1967; Pappas et al. 1993), and that some non-pharmacologic and non-surgical treatment modalities have a therapeutic effect (Stewart 1995), has led to the acceptance of Engel's (1977) biopsychosocial model in clinical practice. The premise of this model is that the patient's complaints cannot be considered in isolation from their psychosocial causes and consequences, and that the patient's disease cannot be divorced from his or her personality and surroundings. The clinical application of the biopsychosocial model is mainly via an interview that is both doctor-centered and patient-centered (Engel 1980; Smith & Hoppe 1991; Engel 1992), which attempts to focus on patients, rather than on the doctor's agenda, and to share with patients the control of the consultation and of the decisions about management (Eshet et al. 1993; Lewin et al. 2001).

A biopsychosocial orientation does not only include the patient-centered interviewing style. It also attempts to provide appropriate support and treatment of both the biomedical and psychosocial components of the patient's predicament. For example, a patient-centered approach to a person with headache would include an empathic attitude, an insight into the patient's beliefs about the cause of the headache and its treatment, and a management plan adapted to his or her preferences. On the other hand, a biopsychosocial consultation would also attempt to identify the patient's concerns and sources of anxiety, assess whether or not the headache is related to these concerns, arrange, if necessary, a joint meeting with the patient and his or her spouse, seek the spouse's opinion about the patient's complaint and its significance for others in the family, develop with the patient and his or her spouse a plan for diagnostic investigation and medication, and offer advice for relaxation and exercise.

Table

Practice points
•Family physicians can describe the main components of a bio-psycho-social (BPS) approach to patient care.
•This approach may be implemented even within the time constraints of a short encounter with a patient.
•Observers can identify and evaluate a biopsychosocial consultation using a simple 9 item evaluation tool.
•This evaluation tool may be used to teach a biopsychosocial approach to patient care and to remedy deficiencies.

Such an approach has been shown to enhance patient satisfaction and reduce health-care expenditure, without markedly changing the duration of the encounter (Margalit et al. 2004). Nevertheless, while remaining influential in medical education (Smith et al. 1994; Waldstein et al. 2001), the biopsychosocial model has been criticized for ambiguity in addressing the practical dimensions of clinical work (Sadler & Hulgus 1990). Attempts to address these dimensions have been based on conceptual models of patient interviewing and management (Marvel et al. 1993; Marvel et al. 1994). However, while interviewing skills have been integrated into practice (Roter 1977; Smith et al. 2000), we know of no attempts to translate the remaining precepts of a biopsychosocial consultation into a list of observable clinical skills that permits their quantitative evaluation.

It would seem that a list of such clinical skills, and a reliable and valid tool for their evaluation, is the cornerstone of the teaching of a biopsychosocial consultation and of an effective feedback to the learners. The objective of this paper is to define the skill components of a biopsychosocial consultation beyond those of the patient-centered interview, and develop an easy-to-use tool for their measurement.

Methods

Identification of the behavioral components of a biopsychosocial orientation

We approached three e-mail discussion groups of family physicians with the question “what types of observed physician behavior would characterize a biopsychosocial consultation?” Each group was approached only once. The responders were not asked any follow-up questions, and there were no discussions within the groups. In total, 35 responses were received: 19 from the USA; 6 from Canada; 5 from Europe; 4 from Israel; and 1 from Australia. The responses detailed 37 types of behavior. After combining and collapsing similar items we reduced the number of behavior types to 22, which appeared to cluster around the following three categories (Appendix 1).

The therapeutic (patient-centered) interview Bird and Cohen-Cole (1990) have defined the goals of the patient-centered interview as data gathering, responding to the patient's emotions (empathy), providing relief of the patient's anxiety, and laying the foundation for a relationship based on trust and cooperation. The e-mail panel emphasized two additional behaviors: acknowledgement and exploration of the patient's sources of distress and provision of support to the patient in dealing with his or her concerns.

A systems orientation and family-centered approach to care A systems orientation is based on Engel's (1977) premise that nothing exists in isolation. Just as it is impossible to understand cardiac dysfunction without considering its effect on the respiratory system, so also is it impossible to understand heart disease without considering the depression and anxiety it causes, the effect of these issues on the patient's performance at home and at work, and how all of these factors loop back to affect the patient's compromised cardiac function. A family-centered approach is based on the assumption that patient care requires a plan of management that is acceptable to both the patient and his or her family, in order to ensure their cooperation and involvement in treatment.

An integrated approach to biomedical and psychosocial problem solving This approach is an attempt to resolve both the biomedical and the psychosocial components of the patient's disease, while creating realistic expectations. Doing so necessitates knowledge of up-to-date, evidence-based clinical guidelines, addressing the immediate patient's problem or source of concern within the time limits of the encounter, making appropriate decisions regarding an economic use of time and health resources, and assessing the need for immediate versus delayed diagnostic work-up.

Development of an evaluation tool

Appendix 1 lists 22 types of desirable doctor behavior during a biopsychosocial consultation. They are presented in the respondents' own words and do not always reflect formal, behaviorally defined skills, such as agenda setting, emotion seeking, and empathy. Accordingly, the behaviors have been defined more succinctly, in order to produce an evaluation tool that guides evaluators without overwhelming them with or restricting them to an overly detailed checklist. An expert panel (consisting of a convenience sample of 12 instructors of the psychosocial aspects of disease in Israeli family medicine residency programs) rated the resulting 16-item questionnaire as to the degree to which each item reflected a biopsychosocial consultation, using a 0–100 scale. The eight items that received the highest ratings (Appendix 2) clustered around the following three aspects of the biopsychosocial consultation: communication skills (items 1–3); treatment (items 4–7); and response to patient expectations (item 8). Neither item 8 nor item 9 were among the 22 types of behavior proposed by the respondents; both were added at the suggestion of the panel. Item 8 assesses the degree to which the physician responded to the patient's needs beyond the patient's expectations (e.g. whether the physician discussed the possibility that distress contributed to a complaint of headache and offered advice for tension reduction, rather than suggesting analgesics only), and item 9 assesses global or overall impression of the physician's performance.

Evaluation of the tool

The tool was evaluated in three studies. The first compared the performance of physicians with a strongly biopsychosocial orientation with the performance of physicians with a strongly biomedical approach. The orientations of these physicians were identified by 51 peers chosen at random from the list of 102 senior residents and board-certified family physicians in Israel. The 51 physicians were asked to name those of their colleagues who had a biomedical approach and those who had a biopsychosocial approach. Approximately half of the physicians named in each category agreed to be videotaped during patient encounters. From these, we randomly selected nine physicians who had reputations of a highly biopsychosocial orientation and four physicians who had reputations of a highly biomedical approach. We deliberately over-sampled physicians with a highly biopsychosocial orientation, because we anticipated a higher variation within their behavior during patient encounters, compared with physicians with high biomedical orientation. All were videotaped in a primary-care clinic while caring for a randomly chosen walk-in patient for the first time. One of the nine physicians with a highly biopsychosocial orientation actually taught a biopsychosocial approach to primary care. He was videotaped during three additional encounters with patients, which increased the number of videotaped encounters to 16.

In the second study, we used the newly developed evaluation tool to assess the videotaped performance of 44 general practitioners 1 month before and 6 months after they had participated in a course teaching a biopsychosocial orientation to primary care. The course, which consisted of weekly 4–6 hour meetings over 12 weeks with two preceptors, has been described previously in detail (Margalit et al. 2004). Briefly, participants were allocated to one of two groups (22 in each group). In the ‘didactic group’, the teaching method consisted of lectures, reading assignments, and discussions. The other, ‘interactive group’, method utilized a smaller proportion of lectures and reading assignments, but used the following three additional methods: (a) role-playing of common encounters with a patient or a family, followed by discussions; (b) analysis of videotaped doctor–patient encounters of the participants before the teaching intervention and analysis of their feelings and state of mind during the encounter; and (c) Balint groups, in which participants shared with others their feelings and attitudes toward patients. In both the interactive and didactic groups, participants were encouraged to implement in their practice the guidelines they had agreed upon in the previous meeting, and to discuss in the next meeting their success in implementing these guidelines.

An unselected sample of 677 consecutive walk-in patients, whose permanent physicians were on vacation, was seen in the patients’ primary-care clinics by course participants, who were temporarily replacing their physicians. Using temporary replacements for absent physicians is common practice and was authorized by the patients' health plan. All but two patients consented in writing to have their encounter videotaped for evaluation of the participant's performance. Depending on the load in the specific primary-care clinic, each participant saw an average of 10 (range 4–15) patients before and after the teaching intervention, and his or her performance was assessed via the averages of the scores of the videotaped patient encounters before and after the teaching intervention.

In the third study, we used the biopsychosocial evaluation tool to compare the videotaped performance of each of the 44 participants encountering the same simulated patient before and after the teaching intervention. In addition to real patients, each of the participants saw a professional actor who impersonated a visibly anxious patient with shoulder pain. The actor was instructed to reveal, only if asked, that he recently resumed smoking after a prolonged cessation, that he was in the midst of divorce proceedings and that he hit his wife on repeated occasions after suspecting she was seeing another man.

Scoring

In the first study, a convenience sample of 18 Israeli instructors in Family Medicine was shown the responses of the international panel (Appendix 1) and encouraged to use them as guidelines when scoring the videotaped performance of the participants with the evaluation tool. We carried out the evaluation through a multiple-matrix sampling design (Munger & Loyd 1988; Gressard & Loyd 1991), which includes a subset of test items (in this case, videotaped encounters with patients) from a large pool that are randomly administered to sub-samples of respondents (in this case, evaluators). Each evaluator scored a total of eight videotaped patient encounters: four with physicians with a highly biopsychosocial orientation; two with physicians with a highly biomedical orientation; and two with the instructor (a specialist in biopsychosocial consultations). Therefore, nine observers independently evaluated each videotaped encounter.

In the second and third studies, the videotaped patient encounters were randomly distributed to one of two observers; these observers were blinded to the teaching method used during the intervention (didactic or interactive) and timing of the encounter (before or after the teaching intervention). The two observers were trained via readings about the biopsychosocial approach, discussions, viewings of videotaped patient encounters, comparisons of their scoring with those of the 18 evaluators from the first study, and further discussions of possible differences between their scores and those of the experts.

The evaluators scored each item on a scale of 0 (minimal) to 100 (maximal). Communication skills were assessed by averaging the scores for items 1–3, treatment was assessed by averaging the scores for items 4–7, the participant's response to the patient's expectations was assessed by the score of item 8, and item 9 was the global rating. The final scores are presented as means and standard errors of the means.

Analysis In the first study, we compared the scores of the two groups of physicians by applying the mixed model, designating the physicians and evaluators as two random effects and the two groups as fixed effects. In the second and third studies, we applied the mixed model with the physicians as a random effect, the group (didactic versus interactive) and time (before and after the teaching intervention) and group x time interaction as fixed effects. The SAS® software (PROC MIXED) was used for the statistical analysis.

The model includes three variance components: variance among physicians; variance among evaluators (interobserver variability); and unexplained, or error, variance. The interobserver variability (Tables 1 and 2) is the difference between evaluators’ scores of the same physician that is beyond the error variance (i.e. beyond the possible differences in the scores that would be given by the same observer on the same physician).

Table 1.  Ratings of clinical performance during the patient encounters of nine biopsychosocially oriented physicians and four biomedically oriented physicians

Variable	Average score (SE) of biopsychosocially oriented physicians (n = 9)	Average score (SE) of biomedically oriented physicians (n = 4)	P value	Interobserver   variance
Communication skills	79 (2.5)	17 (3.5)	<0.0001	29.3
Treatment	73 (2.3)	16 (3.2)	<0.0001	23.2
Response to patients’ expectations	76 (1.8)	7 (2.6)	<0.0001	59.3

Abbreviation: SE, standard error. Note: The P value was evaluated using the MIXED model, which takes into account variability between physicians and evaluators.

Results

Reliability

The overall reliability of the evaluation tool (Cronbach α) was 0.90, and it discriminated between the evaluated physicians (Table 1). Each of the eight items in the evaluation tool discriminated between physicians with a highly biopsychosocial orientation and those with a highly biomedical approach (Table 1). Item 8 (“offers help beyond patient's expectation”) had the highest discriminatory power (largest t-value), but there were practically no differences between the discriminating power of the items: for each of them, the associated p value of the t statistic was <0.0001.

Table 2.  Scores of the various components of clinical performance during encounters between 44 Israeli general practitioners and real patients, before and after a teaching course that promoted a biopsychosocial orientation

	Average score (SE) of interactive-group members (n = 22)		P value	Average score (SE) of didactic-group members (n = 22)			P value of differences between the interactive-group and didactic-group average scores
Variable^a	Before intervention	After intervention		Before intervention	After intervention	P value	Before intervention	After intervention
Number of patient encounters	197	145	-	166	167	-	-	-
Communication skills	7 (1.3)	71 (1.3)	<0.0001	7 (2.0)	53 (1.9)	<0.0001	0.61	0.0002
Treatment	5 (1.3)	68 (1.4)	<0.0001	5 (1.8)	48 (1.8)	<0.0001	0.29	<0.0001
Response to patients’ expectations	4 (1.3)	75 (1.4)	<0.0001	5 (2.0)	45 (1.9)	<0.0001	0.31	<0.0001
Overall evaluation of the patient encounter	10 (2.3)	76 (2.4)	<0.0001	10 (1.7)	60 (1.7)	<0.0001	0.97	<0.0001

Abbreviation: SE, standard error. Notes: ^aThe nominal and operational definitions of the variables are detailed in the methods section.

There was moderate interobserver variability in the scores on communication skills and treatment, but high in the scores on the participant's response to patient expectations (Table 1). It was assumed that each variance component is normally distributed. Thus, given a physician who scores 70 on a 1–100 scale, and given an interobserver variance of 25, then two observers may differ by as much as 10 points (SD = 5, times two) or even more; given an interobserver variance of 64, two observers may differ by as much as 16 points.

Content validity

The evaluation tool was developed after consultation with an international panel of expert discussion groups and repeated reviews by a panel of Israeli experts, who selected the items that best reflect a biopsychosocial orientation in the context of primary care in Israel.

Construct validity

The evaluation tool discriminated between physicians with a highly biopsychosocial orientation and physicians with a highly biomedical approach (Table 1). The scores of the instructor (a specialist in biopsychosocial consultations) were even (although nonsignificantly) higher than those of the eight physicians with a highly biopsychosocial orientation (data not shown). In the second and third studies, the participants’ average scores were significantly higher after the course than before it, and significantly higher for the interactive group than for the didactic one (Tables 2 and 3). As has already been reported, Margalit et al. 2004 the average duration of the encounters after the teaching intervention increased by about 40 seconds compared with encounters before the intervention.

Discussion

There are two main approaches to the assessment of clinical skills (Neufeld & Norman 1985; Arnold 2002). The first is based on behavioral coding using detailed checklists (Roter 1977; Smith et al. 2000). The second consists of qualitative ratings by expert observers, guided by a relatively small number of evaluation items (Cohen et al. 1991; Rothman & Cusimano 2000). The main advantage of the first approach is the quantification of coded behaviors and a higher degree of interobserver reliability (Lang et al. 2004). However, evaluators may be overwhelmed by the multiplicity of the items in the checklist, which in turn may not capture all components of a desirable performance. Furthermore, the use of checklists requires an investment of time in the training of the evaluators. On the other hand, the advantages of the second approach are low cost, flexibility, and the purported ability of global ratings to measure domains not amenable to evaluation by objective coding. Global ratings have been reported to be valid in the context of objective, structured clinical examinations (Cohen et al. 1991). They correlate with checklist scores (Rothman & Cusimano 2000; Lang et al. 2004; Sloane et al. 2004) and have been shown to be relatively stable over a 2-year interval (Keynan et al. 1987). However, they also require experienced evaluators, who may not be readily available.

Table 3.  Scores of the various components of clinical performance during encounters between 44 Israeli general practitioners and a simulated patient before and after a teaching course that promoted a biopsychosocial orientation

	Average score (SE) of interactive-group members (n = 22)			Average score (SE) of didactic-group members (n = 22)			P value of differences betweent he interactive-group and didactic-group average scores
Variable^a	Before intervention	After intervention	P value	Before intervention	After intervention	P value	Before intervention	After intervention
Number of encounters	22	22	-	21	22	-	-	-
Communication skills	8 (1.7)	75 (1.7)	<0.0001	9 (2.3)	55 (2.2)	<0.0001	0.88	<0.0001
Treatment	8 (2.6)	72 (2.6)	<0.0001	8 (2.5)	50 (2.5)	<0.0001	0.92	<0.0001
Response to patients’ expectations	5 (2.2)	86 (2.2)	<0.0001	5 (3.5)	50 (3.5)	<0.0001	0.81	<0.0001
Overall evaluation of the patient encounter	12 (2.0)	88 (2.0)	<0.0001	12 (1.8)	70 (1.7)	<0.0001	0.95	<0.0001

Abbreviation: SE, standard error. Notes: ^aThe nominal and operational definitions of the variables are detailed in the methods section.

Our findings support the assertion that expert observers, when guided by a small number of broad categories, can reliably evaluate the components of a biopsychosocial consultation. The evaluation tool that we derived from the skill components of a biopsychosocial consultation discriminated effectively between physicians with a highly biopsychosocial orientation and physicians with a highly biomedical approach. The tool detected a difference between the eight physicians with a reputation of a high biopsychosocial orientation and an expert instructor in a biopsychosocial approach to care. The tool detected before-and-after-training differences in the performance of general practitioners participating in an intensive course that taught a biopsychosocial orientation to patient care, and differentiated between participants in the interactive group and those in the didactic group. Finally, the evaluators reported that the evaluation tool was easy to understand and apply, and that their evaluations took approximately 1.5 times longer than the videotaped encounter they were assessing.

The teaching intervention improved the particpants’ performance during encounters with a simulated patient more than with real patients. This was probably due to the differences in the type of patients. The simulated patient was a prototype of a somatic presentation of severe emotional distress. He was designed to challenge the participant's ability to diagnose and manage the type of problems that were the focus of the course. After the course, most learners succeeded in correctly identifying and managing this problem as evidenced by the before-and-after differences in communication skills and biopsychosocial management. On the other hand, the real patients were representative of the general population, running the gamut of disorders and with varying degrees of emotional and social concerns. Not all of them needed an extensive psychosocial approach, thereby ‘diluting’ the effect of the course. Different encounters with patients necessitate different encounter goals, and not all patients' visits can and should be handled in the same way. For example, a patient who returns for suture removal may be irritated if his physician attempts to do anything beyond popping the stitch, while a patient in severe emotional distress will probably appreciate an expression of empathy.

The main limitation of this study is that it cannot rule out with certainty a possible bias among the evaluators. This bias was probably minimal in the evaluation of the videotaped patient encounters of the 44 general practitioners, who participated in the teaching intervention, because the evaluators were blinded to the teaching method (didactic or interactive) and timing of the encounter (before or after the teaching intervention). Furthermore, it is highly unlikely that the evaluators were acquainted with the participants in the teaching intervention, and had any preconceived ideas about either of them. However, the evaluators all knew the biopsychosocial expert who was videotaped three times, and it is very likely that they were acquainted with the family physicians who were identified by peer ratings as having a highly biopsychosocial orientation or a highly biomedical approach, and this may have biased their scores.

A possible second limitation is that, even though our evaluation tool included the item “integrates the biomedical treatment plan with the psychosocial treatment”, its bulk heavily reflects on the psychosocial aspects of the patient's predicament. Consequently, even though the evaluation tool differentiated biomedically orientated physicians from biopsychosocially oriented physicians, it might have virtually ignored the biomedical aspects. Both of these limitations might explain the much lower interobserver variance in evaluating biopsychosocially oriented physicians orientation than in evaluating the performance of biomedically oriented physicians. A third limitation is that the evaluation tool may not be able to effectively distinguish between physicians who fall between the extremes of a highly biomedical orientation and highly biopsychosocial orientation.

Nevertheless, we believe that our findings are a first step towards the identification of the specific clinical skills that define a biopsychosocial orientation and the development of a tool for their evaluation. We believe that biopsychosocial skills are important for the quality of patient care and should be a standard part of undergraduate and residency training in primary care, and perhaps in other clinical specialties, too. The evaluation tool that we have described may be of use to those who teach a biopsychosocial approach to patient care to medical students, residents and physicians, and could be used to identify specific learner deficiencies and to provide appropriate feedback.

Acknowledgments

This study was submitted in partial fulfillment of the requirements for the degree of PhD in medical education at Ben–Gurion University of the Negev, under the supervision of Dr SM Glick, Dr CZ Margolis, Ben-Gurion University of the Negev and Dr Michael Katz, Haifa University.

Additional information

Notes on contributors

Alon Pa Margalit

DR MARGALIT is a board-certified family physician and Director of a biopsychosocial clinic for unexplained medical symptoms patients. He is the Chief Surgeon of School Health in Israel. He served as a Coordinator of Education, Family Health Care Department in Haifa and has coordinated and taught continuing education programs in Primary Care and Family Medicine at the University of Tel-Aviv, Ben-Gurion University and the Medical Corps of the Israeli army.

Shimon M. Glick

DR GLICK is an internist, endocrinologist and professor emeritus at Ben Gurion University of the Negev, where he formerly served as Dean and as head of the Moshe Prywes Centre of Medical Education. He currently serves as ombudsman for Israel's National Health Service.

Jochanan Benbassat

DR BENBASSAT is an internist and a Research Associate at the Myers-JDC-Brookdale Institute. Formerly, he was a Professor of Medicine and the Israel Wechsler Professor of Medical Education at the Hebrew-University Hadassah Medical School, and Head of the Department of Sociology of Health at the Ben Gurion university.

Ayala Cohen

DR AYALA COHEN is a member on the Faculty of Industrial Engineering and Management at the Technion-Israel Institute of Technology. She is Professor and Head of the Statistics Laboratory that operates within the Technion Research and Development Foundation. Her main research areas are Biostatistics and Statistical methods in the Behavioral Sciences.

Carmi Z. Margolis

DR MARGOLIS, is a pediatrist, Director of the Moshe Prywes Center for Medical Education and Co-Director of the BGU – M.D. Program in International Health and Medicine in collaboration with Columbia University Health Sciences. Professor emeritus at Ben Gurion University of the Negev. He was Head of the pediatric department.

Appendix 1 Summary of responses of an international panel of primary care physicians to the question “What types of observed physician behavior would characterize a biopsychosocial orientation during a doctor-patient encounter?” and relevant examples

General Principle/Behavior	Observable Skill/Example
Therapeutic (patient-centered) interview
The doctor expresses respect and interest throughout the interview	The doctor maintaining eye contact and a body posture slightly bent forward. Listens to the patient without interrupting and encourages his/her narrative by nodding
The doctor inquires about the concern(s) that made the patient seek medical help	“What made you visit me at this particular time?”
The doctor acknowledges and explores the patient's feelings and sources of distress	“How does that make you feel?” “It is my impression that your main concern is from the cost of surgery. Is this impression correct?” “Which of your problems makes you worry most?”
The doctor is empathic and aware that his or her verbal and non-verbal behavior affects the patient; aware of his own feelings and copes with them	“I can understand how difficult all this is for you” “I can imagine how upset you were”
The doctor expresses respect for the patient's health beliefs and difficulties in coping.	“You ascribe your headache to financial worries”
The doctor considers and respects the patient's views and preferences and is responsive to both patient's needs and health beliefs.	“You may be right in thinking that your financial troubles partly caused your heart attack”
The doctor's behavior implies the message: “I am here for you, my job is to help you”	“Do you have any questions regarding your condition?” “Please don't hesitate to turn to me also for any problem you want to share with me. I shall certainly try to help”
A systems orientation and family-centered approach to care
The doctor considers all factors that may affect the patient's well being, including apparently marginal issues such as waiting time	“How do you manage your new job?”, “I assume it was not easy for you to wait so long before you came in”
The doctor acquires an insight into the meaning of the patient's complaint for him/her and his/her family	“How does this affect you at home/ at work?” “How does your spouse feel about this?” “How does this affect your performance at home/ at work?”
The doctor enquires about the contribution of significant others to the patient's decision, e.g., to seek medical help	“Did you decide about this alone or with somebody else?”
The doctor enquires about opinions of other care-providers / spouse on the patient's problem	“How does your spouse feel about this?”
The doctor explores the possibility that the patient's complaints are associated with a conflict / crisis in the family	“What do you think about the coincidence of your headache and your troubles?”
The doctor involves the patient and his family in the treatment	“It is very important to test your breathing every morning and adjust the treatment according to the results. May I explain this also to your spouse and ask him/her to remind you about this?”
An integrated biopsychosocial problem-solving approach to care
The doctor attempts to address and resolve the immediate patient's problem/source of concern within the time limits of the encounter. Makes appropriate decisions regarding immediate vs delayed diagnostic work-up	“I would like to address this second problem you mentioned more thoroughly at another time. Please make an appointment to see me about this. In the mean time let us focus on your back pain”
The doctor is familiar with updated practice guidelines. Makes appropriate decisions regarding a parsimonious use of time and health resources	“Yes, nifedipine is good for high blood pressure; however, considering your tendency to palpitations, I would still suggest that you begin with hydrochlorothiazide: it is more effective in preventing complications and has little effect on the heart. If this does not work then we shall have to reconsider”
The doctor creates realistic expectations	“I am not sure we’ll find a complete cure for your pain, but let us try to ease it a little and help you cope with daily tasks”
The doctor offers not only medications, but also advice on patient's health promoting behavior, handling of emotions and on the interaction between emotional and physical well being	“I think that the chances of benefit from antibiotics do not justify their side effects. However, I hope you understand that smoking is the real cause of your cough and threatens your health. I think I can help you if you decide to stop smoking”
The doctor uses suggestive/hypnotic messages to increase the placebo effect, and refrains from increasing unnecessarily the patient's anxiety	“Let's try this medication cautiously and at the exact amount you have to take. Please let me know how you feel it affected you”
The doctor suggests a working diagnosis that includes biomedical and psychosocial components	“What you seem to have is migraine headache, which appears particularly when you are upset, and which incapacitates you at home and at work; do you agree with me?”
The doctor suggests a treatment approach that assumes an interplay between emotional, social and physical well being	“If we do not begin by resolving the stress, it may interfere with further investigations or treatment. Do you agree with me?”
The doctor suggests a management plan that encourages the patient to seek help also for emotional and behavioral problems	“I agree with you/ with your spouse that it would be a good idea to postpone further investigation of your back pain until you resolve the problems at work”
The doctor considers possible effects of the proposed management on the patient's performance at work	“Do you think this treatment will help in your work with a computer?”

Appendix 2 Instrument used to evaluate a physician's biopsychosocial orientation during an observed doctor-patient encounter

Please indicate your rating on a range between 0 (not at all) to 100 (very much). The doctor ….
(1) Addresses the psychosocial background of the patient and family (occupation, family status, family history, health beliefs and fears).	(4) Integrates the biomedical treatment plan with the psychosocial treatment.
(2) Is reassuring during the physical examination (verbal & non-verbal).	(5) Checks whether the patient agrees with the suggested treatment.
(3) Checks whether the patient's agrees with the suggested diagnosis.	(6) Offers help beyond patient's expectation.
Overall evaluation of the doctor-patient encounter (including parsimonious use of time and health resources as well as updated practice guidelines)