Answers are grounded in this paper and cite their sources.
How good are we? It turns out that clinicians make reasonably accurate judgments about dangerousness for the purpose of involuntary commitment: persons committed to hospitals for dangerousness are significantly more likely to behave violently during the first days of their hospitalization than are those who are committed for other reasons. 3 Our ability to predict violent behavior over the long term, however, has not yet been demonstrated. We can identify populations more likely to be violent, but we cannot predict the specific risk of violence for a given individual.4,5 In practice, we can assess the approximate likelihood of violence, but not its nature or its magnitude. We plan morerestrictive treatments when we judge the risk of violence to be unacceptable. However, no specific threshold has been defined, either clinically or legally. The standard of care is to complete an assessment and to exercise reasonable clinical judgment. 6 We identify risk factors that correlate empirically with dangerousness; we identify historical factors that characterize the context of an individual patient’s aggressive behavior; and finally, we use those data to formulate a clinical judgment in terms of relative risk. Thus, we aim to integrate our general knowledge of empirically determined risk factors with specific knowledge of an individual’s history. With that combination, we establish a plan that addresses whatever factors can be realistically addressed yet acknowledges the factors that cannot be changed. For example, acute psychosis can be treated, social supports can be enhanced, and a program can be recommended to a patient who abuses substances. In contrast, a past history of violent assault, although it is a factor to be considered, cannot be treated or modified in any way. Identifying the factors that can be addressed in treatment defines the extent and proper limits of what the treatment team can do.
No takes yet. Share an insight, caveat, or question.
Michael Craig Miller (2000) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: