Protocol registration and selective outcome reporting in recent psychiatry trials: new antidepressants and cognitive behavioural therapies.

Abstract

The selective reporting of favorable outcomes has a serious influence on our evidence base. However, this problem has not yet been systematically investigated in the field of psychiatry. Our study aimed to evaluate registration and outcome reporting in randomized controlled trials (RCTs) of standard treatments for depression: cognitive behavioural therapy (CBT) or new-generation antidepressants (ADs). We searched for reports of RCTs examining the efficacy of CBT or AD for depression that were published between 2011 and 2013. We then compared their primary outcomes in the trial registries and those in publications. We identified 170 trials. Among them, 92 trials (54.1%) were registered, 43 trials (25.3%) were properly registered, and only 32 (18.8%) trials were both properly registered and reported (the primary outcomes as recorded in the registries were reported in publications). There was no statistically significant difference in the proportions of properly registered and reported trials for CBT or AD (relative risk: 0.51, 95% CI: 0.25-1.03). High impact factor journals, commercial funding, publication of protocol, and relatively large sample size were significant predictors of proper registration and reporting. The prevalence of proper registration and reporting is still very low in depression trials.