Abstract
The interrater reliability of the Psychiatric Research Interview for Substance and Mental Disorders (PRISM) was assessed in a multicentre study. Four sites of the National NeuroAIDS Tissue Consortium performed blinded reratings of audiotaped PRISM interviews of 63 HIV-infected patients. Diagnostic modules for substance-use disorders and major depression were evaluated. Seventy-six per cent of the patient sample displayed one or more substance-use disorder diagnoses and 54% had major depression. Kappa coefficients for lifetime histories of substance abuse or dependence (cocaine, opiates, alcohol, cannabis, sedative, stimulant, hallucinogen) and major depression ranged from 0.66 to 1.00. Overall the PRISM was reliable in assessing both past and current disorders except for current cannabis disorders when patients had concomitant cannabinoid prescriptions for medical therapy. The reliability of substance-induced depression was poor to fair although there was a low prevalence of this diagnosis in our group. We conclude that the PRISM yields reliable diagnoses in a multicentre study of substance-experienced, HIV-infected individuals.