What constructs do GPs use when diagnosing psychological problems?

BACKGROUND The mismatch between general practice and psychiatric diagnosis of psychological problems has been frequently reported. AIMS To identify which items from the 28-item general health questionnaire (GHQ-28) best predicted general practitioners' (GPs') own assessments of morbidity and the proportion of time spent in consultations on psychological problems. DESIGN OF STUDY Cross-sectional survey. SETTING General practice in southeast London. METHOD Eight hundred and five consultations were carried out by 47 GPs, during which patients completed the 28-item GHQ, and doctors independently assessed the degree of psychological disturbance and the proportion of the consultation spent on psychological problems. Data from the consultations were entered into a stepwise multiple regression to determine the best GHQ-item predictors of GP judgements. RESULTS GPs' assessments of the degree of psychological disturbance were best predicted using only seven GHQ items, and their perceptions of the proportion of time spent on psychological problems were predicted by only four items. Items were drawn predominantly from the 'anxiety and insomnia' and 'severe depression' sub-scales, ignoring the 'somatic' and 'social dysfunction' dimensions. CONCLUSION In diagnosing psychological disturbance GPs ignore major symptom areas that psychiatrists judge important.