The actuarial approach follows ‘objective’ procedures to classify risk, looking at probability and predictability. For example, clinical algorithms are ‘an explicit description of appropriate steps to be taken in the care of a patient with a particular problem’, which should account for a full clinical history, with subsequent recommendations for diagnosis and/or treatment based on the data obtained. Algorithms include ‘branching logic’, which allows recommendations to be ‘individualised according to the patient’s age, gender and speciﬁ c clinical ﬁ ndings’ (Komaroff 1982).
Of course, assessing the likelihood of dangerous behaviour in the context of mental health problems is not the same as assessing chest pain or simply basing a prediction of future behaviour on what has gone before. It involves a degree of clinical assessment and this should be taken into account when considering the reliability of using an actuarial approach, which is of very limited value on its own (Bouch and Marshall 2005).