TO WHAT EXTENT DO CLINICAL NOTES BY GENERAL-PRACTITIONERS REFLECT ACTUAL MEDICAL PERFORMANCE - A STUDY USING SIMULATED PATIENTS

Background Review of clinical notes is used extensively as an indirect method of assessing doctors' performance. However, to be acceptable i t must be valid.Aim. This study set out to examine the extent to which clinical notes in medical records of general practice consultations r eflected doctors' actual performance during consultations. Method Thir ty nine general practitioners in the Netherlands were consulted by fou r simulated patients who were indistinguishable from real patients and who reported on the consultations. The complaints presented by the si mulated patients were tension headache, acute diarrhoea and pain in th e shoulder, and one presented for a check up for non-insulin dependent diabetes. Later, the doctors forwarded their medical records of these patients to the researchers. Content of consultations was measured ag ainst accepted standards for general practice and then compared with c ontent of clinical notes. An index, or content score, was calculated a s the measure of agreement between actions which had actually been rec orded and actions which could have been recorded in the clinical notes . A high content score reflected a consultation which had been recorde d well in the medical record. The correlation between number of action s across the four complaints recorded in the clinical notes and number of actions taken during the consultations was also calculated. Result s. The mean content score (interquartile range) for the four types of complaint was 0.32 (0.27-0.37), indicating that of all actions underta ken, only 32% had been recorded. However, mean content scores for the categories 'medication and therapy' and 'laboratory examination' were much higher than for the categories 'history' and 'guidance and advice ' (0.68 and 0.64, respectively versus 0.29 and 0.22, respectively). Th e correlation between number of actions across the four complaints rec orded in the clinical notes and number of actions taken during the con sultations was 0.54 (P<0.05). Conclusion. The use of clinical notes to audit doctors' performance in Dutch general practice is invalid. Howe ver, the use of clinical notes to rank doctors according to those who perform many or a few actions in a consultation may be justified.