Varela, Jorge2020-08-172020-08-172020-082020-07-02August 202https://hdl.handle.net/20.500.11875/2839The insanity defense has been in place for centuries to address cases in which mental illness is thought to underlie criminal behavior. Those who are found not guilty by reason of insanity (NGRI) are usually involuntarily committed to an inpatient facility until they are deemed to be rehabilitated in terms of their psychiatric symptoms and are no longer considered dangerous. Data from violence risk assessments plays an important role in release decisions for NGRI acquittees. Yet, there are no standard procedures in place for assessing risk for future violence. The identified demographic, criminal, and clinical variables influencing release recommendations vary across regions and hospitals, and possibly practitioners. Additionally, the use of risk assessment instruments has also been shown to vary greatly across studies. The present study uses archival data from a Texas state hospital—specifically, information from risk assessments completed with NGRI patients between 2010 and 2018—with the goal of improving understanding of practical realities of risk assessments and release recommendations. The current study found that several clinical variables (delusions, insight problems, homicidal ideation, psychosocial treatment noncompliance, and violence in the hospital) were associated with clinicians’ risk level determinations; however, the HCR-20 score was the single most impactful predictor of violence risk level. While HCR-20 and some clinical variables (delusions, hallucinations, insight problems, psychosocial treatment noncompliance, and violence in the hospital) were related to release recommendations provided by clinicians; this time, gender was the most significant predictor. Risk level determinations were significantly associated to release recommendations. When looking at changes in dynamic risk factors across repeated risk assessments, the change (or lack thereof) in violent behaviors in the hospital was the most prominent predictor of whether an acquittee was or was nor recommended for release during the studied period.application/pdfenRisk assessmentDischarge decisionsInsanity acquitteesNGRIThesis2020-08-17