Variability in classification of consultations by content-based descriptors.

Abstract

OBJECTIVE: To determine the reliability with which general practitioners use content-based descriptors of consultations. Our hypothesis was that there would be considerable variety in the opinion of general practitioners on the classification of consultations by descriptors. DESIGN: General practitioners in Newcastle, New South Wales, assessed videotaped consultations and the written consultation records by 12 general practitioners. SETTING: The medical faculty of the University of Newcastle and the general practitioners' own homes and offices. PARTICIPANTS: Twelve general practitioner participants were selected from consultation tapes recorded in 1983 by the Primary Care Research Group at the University of Newcastle. Thirty-nine observer general practitioners were randomly selected from a list of general practitioners practising in the Newcastle area maintained by the Hunter Postgraduate Medical Institute. Those whose videotapes had been selected were excluded. RESULTS: There were high levels of variation between assessments made from the original records and from viewing the tapes. Doctors varied greatly in their use of the descriptors. This variation included estimation of the duration of consultations and of the individual components of the consultation (history, physical examination, problems dealt with). No systematic relationships were found between any characteristics of the doctors and their ratings of consultations. CONCLUSIONS: There is potential for enormous variation in the use of content-based descriptors currently available in Australia. Observer general practitioners tended to underestimate the content of the consultations and overestimate the duration. It would be inappropriate to use the general practitioner patient record for audit of the content or difficulty of a consultation.