Performance-based assessment in continuing medical education for general practitioners: construct validity

The use of performance-based assessment has been extended to postgraduate education and practising doctors, despite criticism of validity. While differences in expertise at this level are easily reflected in scores on a written test, these differences are relatively small on performance-based tests. However, scores on written tests and performance-based tests of clinical competence generally show moderate correlations. A study was designed to evaluate construct validity of a performance-based test for technical clinical skills in continuing medical education for general practitioners, and to explore the correlation between performance and knowledge of specific skills. A 1-day skills training was given to 71 general practitioners, covering four different technical clinical skills. The effect of the training on performance was measured with a performance-based test using a randomized controlled trial design, while the effect on knowledge was measured with a written test administered 1 month before and directly after the training. A training effect could be shown by the performance-based test for all four clinical skills. The written test also demonstrated a training effect for all but one skill. However, correlations between scores on the written test and on the performance based test were low for all skills. It is concluded that construct validity of a performance-based test for technical clinical skills of general practitioners was demonstrated, while the knowledge test score was shown to be a poor predictor of competence for specific technical skills.     

Educational assessment of general practice experience for psychiatric trainees

Eighteen psychiatric trainees spent 6 months each as general practice trainees. The educational impact of the experience was assessed by a self-assessment questionnaire, a semi-structured interview and a videotaped interview with a psychiatric patient. Each assessment was conducted at baseline and after 12 months. A control group of 14 trainees was recruited from the same rotation. On the self-assessment questionnaire, the study registrars rated their abilities to solve general medical problems significantly improved compared to controls. They had also acquired greater understanding of the limitations of their knowledge and their legal responsibilities towards their patients. The semi-structured interview failed to distinguish between the two groups. Videotapes for rating at baseline and follow-up were available for only 17 of the trainees. Assessment of the tapes used the Maguire Scale and the Interview Behaviour Scale. Neither scale demonstrated any intervention effect. The interviews were all characterized by a preponderance of 'closed psychological' and 'checking-out' questions. It appears that psychiatric trainees' interviewing styles had not been influenced by the experience. This study suggests that psychiatric trainees gain greater confidence in their role as a doctor and greater understanding of the scope and nature of general practice by such an attachment. It is unclear whether or not supplementary interviewing skills had been acquired which were not utilized in the taped interview, which conforms very much to traditional psychiatric examination behaviour. Trainees were reassured that they had increased their knowledge without losing any of their specific professional skills.     

Growth of Self-Perceived Clinical Competence in Postgraduate Training for General Practice and its Relation to Potentially Influencing Factors

Objective To examine the increase in self-perceived clinical competence during a three-year postgraduate training in general practice and to explore the relation between the growth of self-perceived competence and several background variables. Design Cohort, 1995–1998. Setting Three-year Postgraduate Training for General practice in the Netherlands Participants All Dutch trainees who followed postgraduate training from September 1995 to September 1998 (N=191). Intervention We asked the trainees at the start and at the end of their postgraduate training to complete a questionnaire, which assessed their self-perceived knowledge, clinical skills and consultations skills. We collected information about potentially influencing background variables. Amongst these were variables such as: age, gender, prior medical experience, the effort someone has spent upon her/his education, insight in weak and strong areas of clinical competence and knowledge and skills levels. Main outcome measure Self-perceived competence. Results A total of 127 trainees completed both questionnaires (190 at the first administration and 128 at the second one). We found statistically significant growth of self-perceived clinical competence. Self-perceived consultation skills increased more than self-perceived knowledge and clinical skills. The afore mentioned background variables did not relate in any way with the growth of self-perceived clinical competence. Conclusion This study shows that a 3-year postgraduate training in general practice enhances self-perceived clinical competence. However, we still do not know how to explain this improvement. Further study into the theoretical concept of self-assessment in medical education and into the factors contributing to the feeling of being competent, is required.     

Validation of core medical knowledge by postgraduates and specialists

BACKGROUND: Curriculum constructors and teachers must decide on the content and level of objectives and materials included in the medical curriculum. At University Medical Centre Utrecht it was decided to test relatively detailed knowledge at a regular level in study blocks and to design a progress test aimed at the medical core knowledge that every graduating doctor should possess. This study was conducted to validate the level of knowledge tested in this progress test. AIM: We designed a questionnaire to investigate whether postgraduate trainees and experienced specialists agree with item writers on what is required core knowledge. METHODS: Postgraduates and specialists received a questionnaire with 80 items designed to test core knowledge. Respondents were asked to indicate to what extent the items actually represented the core knowledge required of a recently graduated medical student. RESULTS: Of the clinical questions, 82.4% were judged to reflect core knowledge, whereas only 42.4% of the basic science questions were judged to reflect core knowledge. There was a strikingly high correlation on the mean judgements per item of postgraduate trainees versus medical specialists (r = 0.975). CONCLUSION: Many items, written to reflect core knowledge, appear to be judged by postgraduates and clinicians as pertaining to non-core knowledge. Postgraduate trainees appear to be as capable as experienced specialists of making judgements regarding core knowledge. Fewer basic science items are regarded as core knowledge than clinical items. This may suggest that, specifically, basic science teachers do not agree with physicians on what is to be considered medical core knowledge for graduating doctors.     

Acquisition of communication skills in postgraduate training for general practice

PURPOSE: The evidence suggests that a longitudinal training of communication skills embedded in a rich clinical context is most effective. In this study we evaluated the acquisition of communication skills under such conditions. METHODS: In a longitudinal design the communication skills of a randomly selected sample of 25 trainees of a three-year postgraduate training programme for general practice were assessed at the start and at the end of training. Eight videotaped real life consultations were rated per measurement and per trainee, using the MAAS-Global scoring list. The results were compared with each other and with those of a reference group of 94 experienced GPs. RESULTS: The mean score of the MAAS-Global was slightly increased at the end of training (2.4) compared with the start (2.2). No significant difference was found between the final results of the trainees and the reference group. According to the criteria of the rating scale the performance of both trainees and GPs was unsatisfactory. CONCLUSION: The results of this study indicate that communication skills do not improve in a three-year postgraduate training comprising both a rich clinical context and a longitudinal training of communication skills, and that an unsatisfactory level still exists at the end of training. Moreover, GPs do not acquire communication skills during independent practice as they perform comparably to the trainees. Further research into the measurement of communication skills, the teaching procedures, the role of the GP-trainer as a model and the influence of rotations through hospitals and the like, is required.     

Predictive validity of a written knowledge test of skills for an OSCE in postgraduate training for general practice

PURPOSE: To examine the validity of a written knowledge test of skills for performance on an OSCE in postgraduate training for general practice. METHODS: A randomly-selected sample of 47 trainees in general practice took a knowledge test of skills, a general knowledge test and an OSCE. The OSCE included technical stations and stations including complete patient encounters. Each station was checklist rated and global rated. RESULTS: The knowledge test of skills was better correlated to the OSCE than the general knowledge test. Technical stations were better correlated to the knowledge test of skills than stations including complete patient encounters. For the technical stations the rating system had no influence on the correlation. For the stations including complete patient encounters the checklist rating correlated better to the knowledge test of skills than the global rating. CONCLUSION: The results of this study support the predictive validity of the knowledge test of skills. In postgraduate training for general practice a written knowledge test of skills can be used as an instrument to estimate the level of clinical skills, especially for group evaluation, such as in studies examining the efficacy of a training programme or as a screening instrument for deciding about courses to be offered. This estimation is more accurate when the content of the test matches the skills under study. However, written testing of skills cannot replace direct observation of performance of skills.     

Effect of a short skills training course on competence and performance in general practice

OBJECTIVES: Evaluation of the efficacy of a short course of technical clinical skills to change performance in general practice. DESIGN: Subjects were self-selected general practitioners (n=59) who were unaware of the study design. They were assigned to the intervention group (n=31) or control group (n=28) according to their preference for course date. The course covered four different technical clinical skills (shoulder injection technique, PAP-smear, laboratory examination of vaginal discharge, ophthalmoscopic control in diabetes mellitus). Main outcome measures used were pre- and post-training scores on a knowledge test of skills (60 multiple choice items), and pre- and post-training performance of procedures in practice using a log-diary covering 20 days. SETTING: University of Maastricht, The Netherlands. SUBJECTS: Self-selected general practitioners. RESULTS: Competence, as measured by the knowledge test of skills, improved significantly as a result of the training and skills test scores were satisfactory after training. A significant effect on performance in practice was found for PAP-smear and shoulder injection technique, whereas no effect could be demonstrated for examination of vaginal discharge and ophthalmoscopic control in diabetes mellitus. CONCLUSIONS: A good degree of competence is a necessary but not always sufficient condition for a physician to change his performance in practice. While some skills training seems adequate to bring about desired changes, for other skills more complex interventions are probably needed.     

Assessment in general practice: the predictive value of written-knowledge tests and a multiple-station examination for actual medical performance in daily practice

This study compares the predictive values of written-knowledge tests and a standardized multiple-station examination for the actual medical performance of general practitioners (GPs) in order to select effective assessment methods to be used in quality-improvement activities. A comprehensive assessment was performed in four phases. First, 100 GPs from the southern part of the Netherlands were assessed by a general medical knowledge test and by a knowledge test on technical skills. Second, in order to check for time-order effects, participants were randomly divided into two groups of 50 each, comparable on scores of both knowledge tests and on professional characteristics. Finally, both groups went through a multiple station examination using standardized patients and a practice video assessment of real surgery, but in opposite orders. Consultations were videotaped and assessed by well-trained peer observers. The drop-out rate was 10%. In both groups the predictive value of medical knowledge tests, ranging from 0.43 to 0.56 (Pearson correlation disattenuated), proved to be comparable with the predictive value of the multiple-station examination for actual performance (0.33-0.59). The overall explained variance of scores of the practice video assessment, measured by multiple regression analysis with performance scores as dependent variables and scores on the knowledge tests and the multiple-station examination as independent variables was moderate (19%). A time-order effect showed in only one direction: from practice video assessment to the multiple-station examination. The GP's professional characteristics did not contribute to the explanation of variation in performance. Medical knowledge tests can predict actual clinical performance to the same extent as a multiple-station examination. Compared with a station examination, a knowledge test may be a good alternative method for assessment the procedures of a large number of practising GPs.     

Does knowledge of drug prescribing predict drug management of standardized patients in office practice?

Drug prescribing for the elderly is an important area of medical knowledge since inappropriate prescribing may lead to significant adverse drug events. The objective of this study was to evaluate the association between knowledge of drug use and quality of drug management by general practitioners in practice. A cross-sectional study design was used to evaluate a sample of 37 GPs in practice. A set of common musculoskeletal problems was chosen to evaluate doctors' knowledge of non-steroidal anti-inflammatory drug use, while performance in practice was assessed using elderly standardized patients. The reliability of knowledge test scores was evaluated using measures of internal consistency. The relationship between knowledge of drug use and quality of therapeutic management in practice was evaluated by rank order and linear correlation analysis. Demographic characteristics of participating and non-participating doctors were similar. The reliability of the knowledge test was 0.55. The reliability of performance scores was 0.66. The correlation between overall performance on the standardized patient cases and total knowledge test score was 0.22 (95% confidence interval = 0–0.63). Knowledge test scores were poorly correlated with quality of therapeutic management in office practice. This indicates that knowledge tests alone cannot predict quality of performance with regard to drug prescribing for the elderly in primary care office practice.     

Quality assessment in general practice trainers

INTRODUCTION: General practice trainers hold a key position in general practice training, especially through their provision of a role model. Their own competence in general practice care is important in this regard. The purpose of the study was to evaluate whether a quality assessment programme could identify the strengths and weaknesses of GP trainers in four main domains of general practice care. METHODS: The quality assessment programme comprised validated tests on four domains of general practice: general medical knowledge, knowledge of medical-technical skills, consultation skills and practice management. The criterion for the identification of relative strengths and weaknesses of GP trainers was a variation in the scores of trainers indicating higher and lower scores (strengths and weaknesses) within each domain. RESULTS: GP trainers (n=105) were invited to participate in the study and 90% (n=94) did so. The variation in scores allowed the indication of strengths and weaknesses. Main strengths were: general medical knowledge of the digestive system; knowledge of medical skills relating to the skin; consultation skills concerning empathy; practice management with regard to accessibility. Main weaknesses were: general medical knowledge of the neurological system; knowledge of the medical/technical skills relating to the endocrine metabolic and nutritional system; consultation skills regarding shared decision making; practice management involving cooperation with staff and other care providers. DISCUSSION: This first systematic evaluation of GP trainers identified their strengths and weaknesses. The weaknesses identified will be used in the improvement process as topics for collective improvement in the GP trainers' general curriculum and in individual learning plans.     

Comparison between medical students' experience,confidence and competence

OBJECTIVES: This study was undertaken to determine whether or not breadth of clinical experience and student levels of confidence were indicators of competency on standardized simulator performance-based assessments. METHODS: All students (n=144) attending an educational session were asked to complete a 25-point questionnaire regarding specific clinical experiences and levels of confidence in their ability to manage patient problems. For enumeration of clinical experiences, students were asked to estimate the number of times a situation had been encountered or a skill had been performed. For level of confidence, each response was based on a 5-point Likert scale where 1=novice and 5=expert. Students then participated in a standardized simulated performance test. Median and range were calculated and data analysed using Spearman rank correlations. A P-value <0.05 was considered significant. Level of confidence data were compared to performance during clinical rotation and to marks in the anaesthesia final examination. RESULTS: A total of 144 students attended the session, completed the questionnaire and participated in the standardized test. There were wide ranges of experience and confidence in the 25 listed items. Analysis of data showed good correlation between clinical experience and level of confidence. There was no correlation between clinical experience, level of confidence and performance in a standardized simulation test. Neither was there any correlation between level of confidence and clinical grades or written examination marks. CONCLUSIONS: Clinical experience and level of confidence have no predictive value in performance assessments when using standardized anaesthesia simulation scenarios.     

卫生应急管理人员现场培训模式效果评价

目的 评价卫生应急管理人员现场培训模式的实施效果.方法 在世界卫生组织的资助下,对来自中国省级卫生应急管理部门的9个学员进行6个月的现场培训,采用答辩汇报、试卷测试、问卷调查、访谈和座谈会等方法对现场培训的效果进行评价.结果 全部学员的毕业答辩总体表现均在良好以上(79.58 ～ 88.49分);学员在培训后的卫生应急能力笔试测试平均成绩为(38.25±2.712)分,明显高于培训前的平均成绩(21.50±3.817)分,差异有统计学意义(P＜0.001);培训后学员的卫生应急能力自评总得分为(59.40±5.605)分,明显高于培训前自评总得分(52.00±10.198)分;其中分析评价能力、理解执行政策能力、交流沟通能力、工作实践技能、基础知识技能和领导思考能力都得到明显提高(均有P＜0.05);所有学员均认为参加培训后,卫生应急8方面的能力都有不同程度的提高.结论 卫生应急管理人员现场培训模式效果满意,值得推广.     

Validation of MSAT: an instrument to measure medical students' self-assessed confidence in musculoskeletal examination skills

CONTEXT: Self-assessment promotes reflective practice, helps students identify gaps in their learning and is used in curricular evaluations. Currently, there is a dearth of validated self-assessment tools in rheumatology. We present a new musculoskeletal self-assessment tool (MSAT) that allows students to assess their confidence in their skills in and knowledge of knee and shoulder examination. OBJECTIVES: We aimed to validate the 15-item MSAT, addressing its construct validity, internal consistency, responsiveness, repeatability and relationship with competence. METHODS: Participants were 241 Year 3 students in Newcastle upon Tyne and 113 Year 3 students at University College London, who were starting their musculoskeletal skills placement. Factor analysis explored the construct validity of the MSAT; Cronbach's alpha assessed its internal consistency; standardised response mean (SRM) evaluated its responsiveness, and test-retest, before and after a pathology lecture, assessed its repeatability. Its relationship with competence was explored by evaluating its correlation with shoulder and knee objective structured clinical examinations (OSCEs). Results The MSAT was valid in distinguishing the 5 domains it intended to measure: clinical examination of the knee; clinical examination of the shoulder; clinical anatomy of the knee and shoulder; history taking, and generic musculoskeletal anatomical and clinical terms. It was internally consistent (alpha = 0.93), responsive (SRM 0.6 in Newcastle and 2.2 in London) and repeatable (intraclass correlation coefficient 0.97). Correlations between MSAT scores and OSCE scores were weak (r < 0.2). CONCLUSIONS: The MSAT has strong psychometric properties, thereby offering a valid approach to evaluating the self-assessment of confidence in examination skills by students. Confidence does not necessarily reflect competence; future research should clarify what underpins confidence.     

Predicting communication skills with a paper-and-pencil test

AIM: This study was conducted to investigate the value of a written knowledge test of communication skills for predicting scores on a performance test of communication skills. METHOD: A paper-and-pencil test of knowledge about communication skills and a performance test of communication skills, consisting of four stations with standardised patients, were administered to students of two classes of the medical schools of Maastricht and Leiden, the Netherlands. The results on these tests were compared. RESULTS: From the results of both instruments, the classes of the participating students could be recognised equally well: 60% correct qualifications of the classes by the knowledge test and 64% by the multiple station examination. Between the two tests an overall, disattenuated correlation of 0.60 was found (N=133, P < 0.01), suggesting moderate predictive value of the knowledge test for the performance test of communication skills. The correlation is stronger for students from Maastricht medical school than for their colleagues in Leiden. Correlation between the knowledge of communication skills test and other available test results of the participating Maastricht students is close to zero, suggesting that the test measures a distinct quality of students' competence. DISCUSSION: The paper-and-pencil test of knowledge of communication skills has predictive value for the performance of these skills, but this value seems to be less pronounced than similar findings for clinical procedural skills. The stronger relationship between 'knowing how' and 'showing' in the Maastricht student group might be indicative of an effect of the training format.     

Combined student ratings and self-assessment provide useful feedback for clinical teachers

Many evaluation instruments have been developed to provide feedback to physicians on their clinical teaching but written feedback alone is not always effective. We explored whether feedback effectiveness improved when teachers’ self-assessment was added to written feedback based on student ratings. 37 physicians (10 residents, 27 attending physicians) from different specialties (Internal Medicine, Surgery, Obstetrics/Gynecology, Pediatrics, Neurology, Dermatology, Ophthalmology, ENT, and Psychiatry) were invited to fill out a self-assessment questionnaire on their teaching skills. Students completed an almost identical questionnaire to evaluate the same teachers based on their experiences during clerkships. After receiving written feedback incorporating their self-assessment and the student ratings, the teachers indicated their perceptions of the self-assessment exercise and the written feedback in a questionnaire (five-point Likert scale items) and next, in more detail, in semi-structured interviews with a purposive sample of 12 of the participating teachers. 25 physicians participated (67%). The results showed that self-assessment and student feedback were both perceived as useful (3.7, SD 1.0) but the latter was considered more effective. The physicians we interviewed considered the combination of self-assessment with student ratings more effective than either self-assessment or written feedback alone. Notably, discrepancies between student ratings and self-assessment were deemed a strong incentive for change. We conclude that self-assessment can be a useful tool to stimulate improvement of clinical teaching when it is combined with written feedback based on student ratings. Future research among larger groups is needed to confirm our findings and examine whether these combined tools actually lead to improved teaching.     

Guidelines for assessing postnatal problems: introducing evidence-based guidelines in Australian general practice

BACKGROUND: Postnatal morbidity is high, and many GPs lack the confidence and knowledge to deal with common postnatal problems. There is a high consultation rate, but few women disclose common health problems. OBJECTIVE: The aim of the present study was to increase the knowledge and skills of GPs to enable them to identify and manage common health problems experienced by women in the year following childbirth. METHODS: An educational programme [Guidelines for Assessing Postnatal Problems (GAPP)] embedded within a large randomized community intervention trial [Program of Resources, Information and Support for Mothers (PRISM)] with a before/after evaluation was undergone by Australian GPs working in four metropolitan and four rural communities. The programme comprised audit, interactive workshops, role-play and evidence-based guidelines, and was evaluated at baseline and 6 months through written questionnaires and a surgery consultation with a trained simulated patient evaluator. RESULTS: A total of 68 (86%) GPs took part in the full GAPP programme. The odds of a GP improving on the knowledge items ranged from 1.0 to 16, with the greatest change occurring in knowledge about the effectiveness of cognitive behavioural therapy for maternal depression. Of the GPs with an incorrect response at baseline, the percentage demonstrating improved knowledge at follow-up ranged from 22 to 100%. Around half of the GPs demonstrated excellent communication skills at baseline. Of the remaining GPs, more than half demonstrated greatly improved skills to detect common postnatal problems at follow-up. At baseline simulated patient visit, 70% of GPs inquired about sexual problems yet none inquired about the possibility of abuse, whereas at follow-up 94% inquired about sexual problems and 51.5% facilitated the disclosure of physical and emotional abuse. Anonymous feedback on the programme by participating GPs showed that 89% believed the programme positively influenced their actual practice. Interestingly, GPs demonstrated greater knowledge and skills in the simulated setting than on the written questionnaire. CONCLUSIONS: This relatively brief multifaceted educational programme assisted many participants in improving their knowledge and the skills required to improve both physical and emotional health after birth. Despite being experienced clinicians and participating actively in a programme on interviewing skills, half of the GPs did not facilitate disclosure of the underlying sensitive issue (abuse) during the follow-up consultation and could benefit from further in-depth training in effective communication skills.     

Comparison of two instruments for assessing communication skills in a general practice objective structured clinical examination

OBJECTIVE: In recent decades, there has been increased interest in tools for assessing and improving the communication skills of general practice trainees. Recently, experts in the field rated the older Maas Global (MG) and the newer Common Ground (CG) instruments among the better communication skills assessment tools. This report seeks to establish their cross-validity. METHODS: Eighty trainees were observed by 2 raters for each instrument in 2 standardised patient stations from the final year objective structured clinical examination for Belgian trainee general practitioners. Each instrument was assigned 6 raters. RESULTS: Trainees showed the lowest mean scores for evaluating the consultation (MG7), summarising (MG11), addressing emotions (MG9) and addressing feelings (CG5). Inter-rater kappa statistics revealed fair-to-moderate agreement for the MG and slight-to-fair agreement for the CG. Cronbach's alpha was 0.78 for the MG and 0.89 for the CG. A generalisability study was only feasible for the MG: it was more helpful to increase the number of cases than the number of raters. Agreement between the instruments was examined using kappa statistics, Bland-Altman plots and multi-level analysis. Ranking the trainees for each instrument revealed similar results for the least competent trainees. Variances between and within trainees differed between instruments, whereas case specificity was comparable. Multi-level analysis also revealed a rater-item interaction effect. CONCLUSIONS: The 2 instruments have convergent validity, but the drawbacks of the CG, which has fewer items to be scored, include lower inter-rater reliability and score variance within trainees.     

Quantitative assessment of diagnostic ability

This paper describes variables critical to diagnostic thinking that are based on research by Bordage and Grant & Marsden on the diagnostic thinking of medical students and experienced doctors. The purpose of the study is to use their findings to develop an inventory of diagnostic thinking. A 56-item diagnostic thinking inventory was initially developed; each item contains a stem followed by a 6-point, semantic differential scale. The inventory is designed to measure two aspects of diagnostic thinking: the degree of flexibility in thinking and the degree of knowledge structure in memory. The specific goal of the study is to determine which items discriminate best between weaker and stronger diagnosticians and to reduce the inventory to only those items which significantly contribute to the overall score. Thirty subjects from nine groups, each representing a distinct phase of medical education and clinical practice, participated, namely first- and third-year clinical medical students, house officers, senior house officers, registrars, senior registrars, consultants, trainees in general practice, and general practitioners, all from the UK (n = 270). Discrimination indices were calculated for each item. The revised version of the inventory contains 41 items. All the subjects found the exercise meaningful and the resulting scores showed variance and discrimination. The inventory will eventually be used to assess individual student's and clinician's diagnostic thinking and to plan ways of improving their diagnostic thinking.     

Do doctors pay attention to the religious beliefs of their patients? A survey amongst Dutch GPs

BACKGROUND: Patients' religious beliefs can offer support at times of illness and disease. Therefore religious beliefs of patients are important in doctor-patient interaction. OBJECTIVE: To assess to what extent GPs pay attention to religious beliefs of patients in their daily work. METHODS: A postal questionnaire was sent to 120 GPs. The questionnaire consisted of five clusters of items with precoded Likert-scale answer categories related to several clinical situations. RESULTS: Response rate was 72% (n = 87). Upon registration in the practice, 16% of the GPs paid attention to the religious beliefs of patients, while in situations concerning end-of-life decisions like terminal illness or requests for euthanasia most GPs pay attention to religious beliefs of patients (79%). In general GPs brought up in Protestant families tend to pay more attention to religious beliefs of patients than GPs with a Catholic background (65% vs 36%; 95% CI 5-51) and Protestant GPs pay more attention to these aspects than Catholic GPs (81% vs 47%; 95% CI 5-63). CONCLUSIONS: Most GPs tend to pay attention to religion when their medical possibilities in patient care come to an end. GPs and trainees might be conscious of these aspects in patient management. Since most GPs are familiar just with Western religions, the increasing number of non-Western religious denominations might have consequences for patient care in general practitioners' work.