Psychiatric disorder in the community and the General Health Questionnaire.

Goodchild & Duncan-Jones (1985) have proposed a new scoring method for Goldberg's General Health Questionnaire (GHQ) in which increased weight is given to item responses considered to indicate current (but chronic) affective disorders. This report examines both their revised and the conventional scoring of the GHQ in the context of a longitudinal general population study of psychiatric disorder among women. Sensitivity estimates based upon advised cut-points were found to be higher for the revised than for the conventional scoring scheme, and this difference remained when allowance was made for the duration for which disorders had been present prior to assessment with the GHQ. However, Relative Operating Characteristic (ROC) analysis revealed that while both forms of scoring the GHQ discriminated affective conditions (with or without allowance for their duration) there was no significant difference in their ability to do so. The justification therefore for the post-hoc interpretation of certain item responses as indicators of enduring affective states remains controversial and an issue for future research.