Empirically derived structured violence risk assessment instruments are increasingly used by nurses in forensic mental health settings, typically demonstrating stronger predictive validity than unaided clinical risk assessments, and associated with reduced aggression and reduced restrictive practices including seclusion. However, these instruments are less often used in non-forensic mental health settings despite frequent aggression in these settings. This study represents the first test of the Dynamic Appraisal of Situational Aggression (DASA-IV), a structured instrument used to appraise risk for imminent aggression in a non-forensic mental health hospital. Predictive validity of DASA-IV, and unaided clinical and structured clinical judgements made after DASA-IV assessments were compared. Participants included 105 nurses at two mental health inpatient units in rural Victoria, Australia. During the study, 482 DASA-IV assessments and structured clinical judgements were compared with 997 unaided clinical risk judgements. DASA-IV total scores predicted aggression significantly better than unaided clinical risk ratings over the subsequent 24 hours and for the next shift. Nurses' structured clinical judgement ratings were more accurate than unaided clinical appraisals but less accurate than actuarial (DASA-IV derived) scores. The DASA-IV presents as a valid measure for appraising risk of imminent aggression in mainstream mental health inpatient settings.