Author's reply

Kripalani et al express their concerns about biases towards the treatment arm of our study and the characteristics of our study group of patients who self-harm. With respect to biases towards the treatment arm, it should be noted that at the start of treatment no significant differences in anxiety, depression and suicidal cognitions were evident. Thus, the gradually growing difference in depression and suicidal cognitions from the first follow-up at 3 months and in anxiety at the 9-month follow-up in our opinion reflects a treatment effect. Just because the effects on secondary measures were stronger than on the target variable, we concluded that, as hypothesised, CBT primarily targeted maintaining factors of self-harm and that the specific self-harm effect was a secondary effect. Moreover, our study results remain silent on whether the treatment effects observed are attributable to specific ingredients of CBT or to the total package of CBT in addition to TAU. We agree with Kripalani et al, however, that the fact that assessments were not carried out masked to treatment group might have influenced outcome. With respect to characteristics of the study group, participants in our study manifested both self-poisoning (91%) and self-injury (9%) irrespective of the apparent purpose of the act, and therefore can be considered a representative sample of patients who self-harm. Of the contacted participants, only 7.3% were excluded because of schizophrenia or alcohol and drug misuse. Our final sample consisted of females (94%) with a long history of self-harm (77% reported 10 or more previous episodes of self-poisoning and/or self-injury) and severe psychological and psychiatric problems (on average four psychiatric diagnoses (mood and anxiety disorders in particular)). It is possible that CBT as an add-on to TAU is more likely to be effective for people with such chronic and severe self-harm. The fact that rate of withdrawal from CBT amounted to 17% underscores the feasibility of an intervention tailored to the needs of this particular group.

In conclusion, CBT appears to be an effective adjunct to TAU in chronic self-harm and further research on moderators and mediators of change seems warranted.