OBJECTIVE: The authors examined the refusal of antipsychotic medications and associated outcomes prospectively in a group of 348 psychiatric patients admitted to three acute inpatient units in a state- operated mental health facility in Virginia where psychiatrists have the discretionary power to administer treatment over patients' objections. METHODS: Newly admitted patients were administered both a questionnaire to ascertain their attitudes toward admission and the Brief Psychiatric Rating Scale (BPRS). Patients who refused antipsychotic medication were identified, and data were collected on the length of refusal and whether the refusal episode was terminated voluntarily or involuntarily. A group of patients compliant with antipsychotic medication was selected for comparison on outcome measures, including the rate of seclusion and restraint and length of hospitalization. RESULTS: Patients who refused treatment were found to have significantly higher BPRS scores than were patients who complied with antipsychotic treatment and more negative attitudes toward hospitalization and past, present, and future treatment. Refusal episodes were brief, on average 2.8 days, and all patients, who refused treatment were treated. When compared with the compliant patients, patients who refused treatment were more likely to be assaultive, were more likely to require seclusion and restraint, and had longer hospitalizations. CONCLUSIONS: Psychiatrists exercised their discretion to promptly treat all patients who refused treatment. Nonetheless, these patients suffered more morbidity than compliant patients. This study suggests that the negative sequelae of in-hospital treatment refusal cannot be eliminated by rapid treatment. The policy implications are discussed.Abstract Teaser