Content of command hallucinations predicts self-harm but not violence in a medium secure unit

Evidence to date has supported negative relationships, a null relationship and a positive relationship between command hallucinations and violence or self-harm. This study was designed to determine the relationship between command hallucinations with violent or self-harm content and incidents of violence and self-harm in forensic inpatients. Patients with (n = 56) and without (n = 54) a lifetime history of command hallucinations and resident in a medium-security hospital were identified through clinical and legal records over 51 months. Measures included: staff-observed violence and self-harm; presence and content of command hallucinations; paranoid delusions; previous violent convictions; length of stay; gender; history of alcohol or illicit drug abuse. Statistical analyses used negative binomial regression. Violent command hallucinations and inpatient violence were unrelated. Self-harming command hallucinations and an absence of paranoid delusions were positively associated with self-harm. The processes that determine compliance with command hallucinations remain unclear.