A short computer interview for obtaining psychiatric diagnoses.

Abstract

To determine whether a short computer interview could be used in place of a full diagnostic interview to obtain psychiatric diagnoses, the authors examined the short interview's sensitivity, specificity, and diagnostic agreement with the full interview. Patients recently discharged from a university psychiatric service were interviewed in two back-to-back sessions, one in which a full diagnostic interview was used and the other in which a short computer interview was used. Based on diagnoses derived from both interviews, the sensitivity and specificity of the short interview and kappa values reflecting the diagnostic agreement of the two interviews were calculated. The short interview had high sensitivity and specificity and excellent diagnostic agreement with the full interview for most disorders. It also had a significantly shorter administration time. However, it missed a substantial percentage of cases of generalized anxiety disorder and misclassified as in remission a substantial proportion of patients with active cases of post-traumatic stress disorder. With few exceptions, the short interview may be substituted for the full interview when missing an active case is not important or when a count of individual symptoms and subtyping of disorders are not needed. Such uses include screening subjects for inclusion in a study and obtaining an overview of a patient's lifetime psychiatric status.