This study examined the ability of a risk assessment algorithm, the Risk of Harm to Others Clinical Assessment Protocol (RHO CAP) to predict inpatient aggression within a mental health and addictions treatment facility in Ontario, Canada. Anonymized patient records were retrospectively reviewed from April 1, 2004 to July 31, 2009 (N = 6,425). Survival analysis using Cox's regression was used to predict time to inpatient aggression using the RHO CAP. Approximately 10% of inpatients were at moderate risk of harm to others, and 2% were considered high risk. The pattern of survival curves revealed that within the first month of admission, approximately 10% of inpatients at high risk of harm to others displayed physical aggression. Patients at high risk were also two times more likely to display physical aggression. Clinical teams can use the RHO CAP to implement preventive safety measures, reduce the incidence of inpatient aggression and improve quality of care.