Treatment as usual (TAU) as a control condition in trials of cognitive behavioural-based psychotherapy for self-harm: impact of content and quality on outcomes in a systematic review

Background Randomized controlled trials (RCTs) are the mainstay of evaluations of the efficacy of psychosocial interventions. In a recent Cochrane systematic review we analysed the efficacy of cognitive behavioural-based psychotherapies compared to treatment as usual (TAU) in adults who self-harm. In this study we examine the content and reporting quality of TAU in these trials and their relationship to outcomes. Methods Five electronic databases (CCDANCTR-Studies and References, CENTRAL, MEDLINE, EMBASE, and PsycINFO) were searched for RCTs, indexed between 1 January 1998 and 30 April, 2015, of cognitive-behavioural interventions compared to TAU for adults following a recent (within six months) episode of self-harm. Comparisons were made between outcomes for trials which included different categories of TAU, which were grouped as: multidisciplinary treatment, psychotherapy only, pharmacotherapy only, treatment by primary care physician, minimal contact, or unclear. Results 18 trials involving 2,433 participants were included. The content and reporting quality of TAU varied considerably between trials. The apparent effectiveness of cognitive behavioural psychotherapy varied according to TAU reporting quality and content. Specifically, effects in favour of cognitive-behavioural psychotherapy were strongest in trials in which TAU content was not clearly described (Odds Ratio: 0.29, 95% Confidence Interval 0.15 to 0.62; three trials) compared to those in which TAU comprised multidisciplinary treatment (Odds Ratio: 0.79, 95% CI 0.63 to 0.97; 12 trials). Limitations The included trials had high risk of bias with respect to participant and clinical personnel blinding, and unclear risk of bias for selective outcome reporting. Conclusions TAU content and quality represents an important source of heterogeneity between trials of psychotherapeutic interventions for prevention of self-harm. Before clinical trials begin, researchers should plan to carefully describe both aspects