Hassiotis et al ^{Reference Hassiotis, Gazizova, Akinlonu, Bebbington, Meltzer and Strydom1} describe an excess of probable psychosis in prisoners with intellectual disabilities (11.3% v.5.7%, P<0.01). We tried to replicate this finding in a large database of 21 857 pre-trial reports of Dutch defendants. ^{Reference Vinkers, Barendregt and de Beurs2,Reference Vinkers, de Beurs, Barendregt, Rinne and Hoek3} A diagnosis of intellectual disability (IQ<70) was made in 609 defendants (2.8%). However, these individuals had fewer psychotic disorders than defendants without intellectual disability (5.9% v. 12.7%, P<0.001). Furthermore, fewer defendants with intellectual disabilities reported misuse of hard drugs (13.4% v. 24.6%, P<0.001) and alcohol (16.6% v. 23.1%, P = 0.002) and their rate of cannabis misuse was similar to that of defendants with a normal IQ (12.9% v. 14.2%, P = 0.51). This again contradicts the findings of Hassiotis et al, who found more cannabis misuse and similar misuse of hard drugs and alcohol in individuals with intellectual disability.

What could explain these opposite findings? The diagnosis of probable psychosis in the Hassiotis et al study was, in 80% of the cases, based on a lay interview, and intellectual disability was defined as a low score on the Quick Test. Diagnosis in Dutch pre-trial reports is based on: (a) multiple examinations of the defendant by a psychiatrist and/or psychologist; (b) the defendant’s judicial and psychiatric history, including previous examinations; (c) information from relatives; and (d) IQ tests in 88% of defendants with intellectual disabilities. As Hassiotis et al themselves suggest, their method may have led to an overestimation of the prevalence of intellectual disability (4%). Indeed, a systematic review in 2008 showed that the prevalence of intellectual disability in prisoners ranged from 0.0 to 2.8%. ^{Reference Fazel, Xenitidis and Powell4} Moreover, low scores on the Quick Test are significantly related to the prevalence of psychosis. ^{Reference Marjoram, Gardner, Burns, Miller, Lawrie and Johnstone5} Confounding of the relationship between probable psychosis and intellectual disability is therefore probable. The conclusion reached by Hassiotis et al is premature and more studies on this topic are needed.