Screening for psychiatric illness in general practice: the general practitioner versus the screening questionnaire.

This study compares the characteristics of general practitioners and a pen and paper test in the detection of psychiatric disorder in primary care settings. A psychiatrist interviewed a stratified sample of 283 patients drawn from 590 consecutive new illnesses seen in 15 general practices. Research diagnoses could be made in between one-quarter and one-third of the consecutive new illnesses. Two different research diagnostic systems agreed quite well with one another about who should be regarded as a psychiatric 'case' - although agreement between them for individual diagnoses was less impressive. Research diagnoses of psychiatric illnesses could be made in approximately 30% of new episodes of illnesses seen. If the DSM-3 system was used as a criterion of 'caseness', the specificity of the general health questionnaire was 75.4%, and the sensitivity was 87.1%. The general practitioners had fewer false positives than the questionnaire, but they were much more likely to miss psychiatric cases. Use of the general health questionnaire could increase their sensitivity from about 50% to 95%.