The effect of treatment as usual on major depressive disorder: A meta-analysis

Background: Health-economic models are used to evaluate the long-term cost-effectiveness of an intervention and typically include treatment as usual (TAU) as comparator. Part of the data used for these models are acquired from the literature and thus valid information is needed on the effects of TAU on depression. The aim of the current meta-analysis was to examine positive and negative outcomes of major depression for patients receiving TAU. Methods: We conducted a systematic literature search in PubMed, EMBASE, Psyclnfo, and the Cochrane Central Register of Controlled Trials. Eligible studies were randomized controlled trials including a TAU group for depression. The quality of the included studies was assessed using the criteria described in the "Risk of bias assessment tool". Four separate meta-analyses were performed to estimate remission, response, reliable change and deterioration rates at short-term (<= 6 months from baseline). Results: Thirty-eight studies including 2099 patients in the TAU were identified. Nine studies (24%) met five or six quality criteria, 17 studies (44%) met three or four quality criteria and 12 studies (32%) met one or two quality criteria. After adjusting for publication bias, the first meta-analysis (n=33) showed that 33% of the patients remitted from depression. The second meta-analysis (n=13) demonstrated that 27% of the patients responded to treatment, meaning that their depressive symptom decreased at least 50% from baseline to follow-up measurement. The third meta-analysis (n=7) indicated that 31% of the patients showed a reliable change, meaning that their depressive symptoms improved more than expected by random variation alone. Finally, 12% of the patients deteriorated, meaning that their depressive symptoms became more severe. Limitations: Statistical heterogeneity was substantial in most analyses and was not fully explained by subgroup analyses. The quality of the included studies was moderate. This may result in overestimation of the true effects. Conclusions: The treatments labelled as TAU for depression were clinically and statistically heterogeneous. We demonstrated that a few patients benefited from TAU and a small number of patients suffered from worsened depressive symptoms at the short term. The results can be included in health-economic models that compare depression treatments to TAU.