Risk assessments for acute violence to third parties: a review of the literature.

OBJECTIVES To provide an overview of risk assessments for acute violence to third parties by combining a clinical and research focus and to offer guidelines to physicians conducting clinical assessments. METHOD A computerized literature search of the MEDLINE and PSYCHINFO data bases from 1967 to 1996 was completed using the key words violence, aggression, dangerous behaviour, risk, risk assessment, risk factors, and practice guidelines. The search yielded 116 relevant references, 26 of which were original research articles on risk factor identification. A secondary search, based on the citations from the primary search, yielded an additional 8 general discussion articles. RESULTS Risk assessments may be conducted using different methods, although all methods should be systematic and comprehensive. Research shows that risk assessments do have validity for use in short-term prediction and that it is possible to develop clinical guidelines in this area. A combined clinical and research approach holds the most promise for improving the accuracy of probability estimates, and most published guides and tools rely on such a combination. CONCLUSIONS Risk assessments are an important and necessary part of the clinical examination. Because this field has sufficiently evolved, there is abundant literature to refer to when determining what constitutes an acceptable assessment for risk of violence to third parties and when it is appropriate to conduct such an examination.