Validity of the 28-item General Health Questionnaire in hospitalized gastroenterology patients.

The usefulness of the 28-item General Health Questionnaire (GHQ) in detecting psychiatric comorbidity in a sample of hospitalized gastroenterology patients was tested. Validity coefficients were determined by comparing the results of the GHQ with findings from a psychiatric interview. At the best discriminating score (8/9), the GHQ yielded a sensitivity of 77%, a specificity of 75%, a positive predictive value of 67%, and a misclassification rate of 24%. The GHQ did not accurately identify alcohol dependence and abuse and did not discriminate between patients with anxiety disorders and those with depressive disorders. Despite its limitations, the GHQ, if supplemented with specific questions about alcohol abuse, can effectively screen medical inpatients for psychiatric disorders.