[Evidence-based health care in practice. The scientific basis of clinical management is better than reported].

In a series of 197 consecutive patients admitted to a department of medicine, the scientific basis of major interventions was evaluated by comparison with corresponding data in published reports of randomised controlled trials (RCTs) or, if no such data were available, according to the consensus of a panel of experienced consultants in internal medicine. Of the 369 major interventions (including expectant management) in the series, 50 per cent were judged according to RCTs and 34 per cent according to panel consensus, no authoritative support being available in the remaining 16 percent of cases. Intervention consistent with corresponding RCT data was somewhat more common in the under 70 than in the over 70-year-old age subgroup of patients, but there was no such difference between male and female subgroups. Among the various specialties and subspecialties, support from RCTs was available for the majority of interventions in cardiology, stroke/angiology and gastroenterology, where-as support from the panel was available for the majority of interventions in endocrinology, nephrology, haematology, and oncology. Regarding the management of patients presenting with symptoms only, without a specific diagnosis, no support was available from RCTs and little consensus among the panel.