We studied the efficacy and feasibility of using computer-based instruction to provide medication information to hospitalized patients with acute psychotic conditions. Patients were randomly assigned to receive computer-based (n = 21) or personal instruction (n = 21); for the final analyses the computer group was expanded to include 13 patients from a pilot study. Outcome measures were knowledge retention (indicated by changes in test scores) and compliance with medication regimens after discharge (indicated by telephone follow-up at one week, one month, and three months). The subjects reacted positively to the computer program. Knowledge retention and compliance were similar in the computer and control groups. We conclude that psychiatric inpatients admitted for acute care can participate in, and learn from, computerized medication instruction.