At our local journal club at the Gordon Hospital, Westminster, we recently read the excellent paper by Crawford et al on patient experience of negative effects of psychological treatments. ^{Reference Crawford, Thana, Farquharson, Palmer, Hancock and Bassett1} All present were first struck by the novelty of the concept of considering the side-effect profiles of psychological therapies – and then, a split second later, astonished by our own astonishment. As psychiatrists thinking about aetiology and treatment, we are fed and watered on the biopsychosocial model. We are also accustomed to sharing the potential benefits and problems associated with treatments we offer, but seemingly only in matters of medication. We are grateful to Crawford et al for bringing this ‘blind spot’ to our attention and hope their paper will help raise awareness of the simple yet fundamental observation that psychosocial interventions may also have downsides.

As the authors have acknowledged in their ‘Limitations’ section, their study is not without problems. First, we – like the authors – noted the low (19%) inclusion rate of participants relative to the original sample identified. There may well be significant differences between the characteristics of the 19% who did take part and the 81% who did not, creating considerable potential for bias. Second, with a view to excluding potential confounding, we would have liked to know a good deal more about the clinical details of the participants – their diagnoses and, in particular, what other treatments they may have been receiving.

In addition to these methodological observations, we were left with a sense that the practical applicability of the study's findings is significantly limited by the lack of what the authors term ‘qualitative data about negative effects’. When trying to imagine ourselves drawing on the paper as part of evidence-based practice, we strongly suspected that patients would not find it helpful to be told that there is a 5.23% chance they will have ‘lasting bad effects from the treatment’. We would be keen to know more about what the authors' ‘ongoing analysis of in-depth interviews’ has revealed in this regard.

Finally – more at the level of intrigue than critique – we were interested by two findings which appear to point in rather different directions. The first is the strikingly low rate (5.23%) of reported side-effects of therapy, with roughly 87% of respondents reporting no negative effects. Taking into account the earlier point about giving as much consideration to potential side-effects of psychological (and social!) interventions as biological ones, and considering that the efficacy of psychological therapy is, at least for some conditions, broadly similar to that of medication, the side-effect rates identified seem almost too good to be true. We wonder if this may reflect a corollary in patients of our own hitherto lack of awareness of the potential downsides of psychological treatment. On the other hand, our eyes were caught by Table 3 of the paper, which seems to indicate that receiving a large number of sessions of psychological treatment is associated with an increased rate of side-effects. Of course, it may be that the higher number of sessions is due to increased severity and complexity of cases, in which we would expect negative experiences (perhaps interpreted as side-effects) to be more frequent. However, we cannot rule out the possibility of the phenomenon of ‘cognitive–behavioural toxicity’, which should clearly be a focus for further consideration and research.