Dr Daly argues that the link between child sexual abuse and adult psychosis may be the result of confounding by psychotic symptoms in childhood or adolescence. He adduces evidence for this from his secondary analysis of data from the 1970 British Birth Cohort sample. ^{Reference Daly1} Of the female sample, 1.6% indicated that they had been forced to have sex by the age of 16, and this was associated with an elevated risk of visual and auditory hallucinations at age 29 (OR = 8.5). However, after controlling for the experience of such quasi-psychotic symptoms before the age of 16, the odds ratio fell to a non-significant 2.4. Daly interprets this as indicating that this relationship exists because children with quasi-psychotic symptoms are more at risk of abuse and also at greater risk of developing psychosis as adults.

Nevertheless, Dr Daly’s conclusion must equally be tentative. First, the British Birth Cohort sample apparently does not provide temporal discrimination between the occurrence of sexual abuse and the development of quasi-psychotic symptoms. Second, given that this is so, the diminution of the odds ratio after controlling for quasi-psychotic symptoms in adolescence could indicate mediation. In other words, the sexual abuse leads to adolescent symptoms which are then associated with adult symptoms. I find this explanation more plausible than the suggestion that psychotic symptoms themselves have a major effect in increasing vulnerability to abuse. There is some evidence that psychotic symptoms in adolescence are associated with prior abuse. ^{Reference Kelleher, Harley, Lynch, Arsenault, Fitzpatrick and Cannon2}

It would be good to resolve this argument with appropriate data from a cohort study. However, this might not be possible: there are considerable ethical difficulties in contemporaneous enquiry about sexual abuse in child and adolescent epidemiological samples. Current research has provided some indication that the psychological consequences of abuse show similarities to psychological antecedent and maintaining factors in psychosis, ^{Reference Fowler, Freeman, Smith, Kuipers, Bebbington and Bashforth3,Reference Smith, Fowler, Freeman, Bebbington, Bashforth and Garety4} and this does add plausibility to the aetiological role of sexual abuse. The particular association of early trauma with psychotic disorders (schizophreniform or bipolar) characterised by hallucinations is also difficult to explain in terms of confounding. ^{Reference Hardy, Fowler, Freeman, Smith, Steel and Evans5,Reference Hammersley, Dias, Todd, Bowen-Jones, Reilly and Bentall6}

The final worry about Dr Daly’s argument is that it may detract attention from therapeutic engagement with the consequences of sexual abuse and other trauma in people with psychosis.