Selective outcome reporting and sponsorship in randomized controlled trials in IVF and ICSI.

STUDY QUESTION: Are randomized controlled trials (RCTs) on IVF and ICSI subject to selective outcome reporting and is this related to sponsorship?

SUMMARY ANSWER: There are inconsistencies, independent from sponsorship, in the reporting of primary outcome measures in the majority of IVF and ICSI trials, indicating selective outcome reporting.

WHAT IS KNOWN ALREADY: RCTs are subject to bias at various levels. Of these biases, selective outcome reporting is particularly relevant to IVF and ICSI trials since there is a wide variety of outcome measures to choose from. An established cause of reporting bias is sponsorship. It is, at present, unknown whether RCTs in IVF/ICSI are subject to selective outcome reporting and whether this is related with sponsorship.

STUDY DESIGN, SIZE, DURATION: We systematically searched RCTs on IVF and ICSI published between January 2009 and March 2016 in MEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials and the publisher subset of PubMed. We analysed 415 RCTs.

PARTICIPANTS/MATERIALS, SETTING, METHODS: Per included RCT, we extracted data on impact factor of the journal, sample size, power calculation, and trial registry and thereafter data on primary outcome measure, the direction of trial results and sponsorship.

MAIN RESULTS AND THE ROLE OF CHANCE: Of the 415 identified RCTs, 235 were excluded for our primary analysis, because the sponsorship was not reported. Of the 180 RCTs included in our analysis, 7 trials did not report on any primary outcome measure and 107 of the remaining 173 trials (62%) reported on surrogate primary outcome measures. Of the 114 registered trials, 21 trials (18%) provided primary outcomes in their manuscript that were different from those in the trial registry. This indicates selective outcome reporting. We found no association between selective outcome reporting and sponsorship. We ran additional analyses to include the trials that had not reported sponsorship and found no outcomes that differed from our primary analysis.

LIMITATIONS, REASONS FOR CAUTION: Since the majority of the trials did not report on sponsorship, there is a risk on sampling bias.

WIDER IMPLICATIONS OF THE FINDINGS: IVF and ICSI trials are subject, to a large extent, to selective outcome reporting. Readers should be aware of this to avoid implementation of false or misleading results in clinical practice.

STUDY FUNDING/COMPETING INTERESTS: No funding received and there are no conflicts of interest.

TRIAL REGISTRATION NUMBER: N/A.