We read with interest the paper by Duke et al (Reference Duke, Pantelis and McPhillips2001) reporting on non-alcohol substance misuse among people with schizophrenia. We recently conducted a study addressing that question among patients with schizophrenia from the Parisian suburbs, using the Composite International Diagnostic Interview for DSM—III—R diagnoses of substance abuse or dependence (Reference Dervaux, Bayle and LaqueilleDervaux et al, 2001). We found that 41% of the patients had a lifetime history of substance abuse or dependence, including 30% on a drug other than alcohol (v. 16% in the Duke et al study). As in Duke et al's study, in our study the patients with substance abuse were younger than nonabusers. We also found no difference in respect of psychiatric symptoms using the Positive and Negative Syndrome Scale, mean age of first psychiatric treatment or mean number of previous hospitalisations. The patterns of consumption were slightly different for cannabis (27% of the patients in Paris v. 18.7% in London) and opiates (7% v. 5.3%, respectively). The main difference concerned the misuse of cocaine and crack, less common in France (1%) than in the UK (8.7%), and considerably less than in North American studies (ranging from 20 to 30%; Reference Cantor-Graae, Nordstrom and McNeilCantor-Graae et al, 2001). Although methodological issues could explain some of these differences (e.g. the use of different assessment instruments; mean age of the subjects was 34.7 years in our study v. 50.3 years in the English study), these patterns most probably reflect the different availability of drugs in these countries. Taken together, these observations do not argue in favour of the self-medication hypothesis but instead emphasise the role of alternative, and probably multiple, factors leading to substance misuse in schizophrenia.