Trial pre-registration and reporting in eMental Health for anxiety and depression: the importance of being earnest

Context
The replicability crisis has sparked a wide interest in transparency. In clinical trials, prospective trial registration ensures that results can be reported and interpreted transparently; leaving little room to alter, insert or hide results. When a trial is not (prospectively) registered, there is no way to assess whether questionable research practices such as outcome switching or cherry-picking have taken place, or whether a study was ‘file-drawered’. As a field that is consolidating its evidence base, we undertook a review of eMental Health interventions for anxiety and depression to assess transparency in trial reporting in this field.

Methods
We systematically searched clinical trial registries (clinicaltrials.gov, ISRCTN, WHO-ICTR) records researching internet-based interventions for anxiety and/or depression in adults. From these records, we traced publication status, whether these records were registered prospectively or retrospectively, and whether publications concurred with records in terms of sample size and outcome measures.

Results
Many published studies showed discrepancies when compared to their respective trial records. Most studies failed to recruit a pre-specified number of participants, causing them to be underpowered according to their own power calculations. Outcome switching was common, and new, statistically significant primary outcome measures in favour of the intervention being studied were often inserted. Retrospective registration was common despite many journals explicitly requiring prospective pre-registration.

Conclusions
Trials in eMental Health for anxiety and depression – although mostly underpowered – are reported relatively transparently; but authors, reviewers and journal editors all have a responsibility to improve reporting and maximise study result reliability and credibility.