Objective: Ample research has examined physicians' evidence-based medicine (EBM) knowledge and skills; however, previous research has not linked EBM knowledge to objective measures of process of care.
Design: A cross-sectional study of quality of care measures extracted from electronic medical records and EBM knowledge assessed via a validated questionnaire.
Setting: One region of the largest Health Maintenance Organization in Israel.
Participants: Seventy-four physicians and their 8334 diabetic patients, 7092 coronary heart disease patients and 17 132 hypertensive patients.
Main outcome measures: Outcome measures were four diabetes quality of care indicators (LDL tests, microalbumin tests, hemoglobin A1C tests, eye examination referrals), and two drug prescription indicators (statin prescription for coronary heart disease patients, and thiazide prescription for hypertensive patients). Independent variables were total EBM knowledge and its components: critical appraisal and information retrieval.
Results: Total EBM knowledge was independently and significantly associated with LDL testing (b = 0.13; P = 0.036), microalbumin testing (b = 0.33; P = 0.001), hemoglobin A1C testing (b = 0.17; P = 0.036), eye examination referrals (b = 0.16; P = 0.021) and statin prescriptions (b = 0.18; P = 0.025). Critical appraisal was independently associated with microalbumin tests (b = 0.46; P = 0.002) and eye examination referrals (b = 0.20; P = 0.048). Information retrieval was only independently associated with hemoglobin A1C testing (b = 0.43; P = 0.004). Thiazide prescription was not associated with EBM knowledge scores.
Conclusions: Physicians' higher total EBM knowledge primarily correlates with better quality of care; however, correlations were modest and explained only a small portion in the variance of clinical performance. Results indicate that there might be a need to focus on teaching all the components of EBM rather than EBM microskills.