A study published in September 2016 found major discrepancies in the reporting of clinical trials involving children. A team of researchers examined 20 trials and discovered that “all 20 trials selectively reported or failed to report main outcomes”.
In addition, the researchers found:
- 13 trials completed early without justification
- 11 downgraded or modified primary outcome or upgraded secondary outcomes
- 9 had discrepancies in declaring sponsorship
- 8 had discrepancies in the sample size
- 9 failed to respect inclusion or exclusion criteria
The authors of the study looked at 20 trials whose findings had been published in a widely read peer-reviewed journal and cross-checked that information with data from clinical trials registries.
Paola Rosati, who led the team of researchers, works in Rome’s largest children’s hospital, helping doctors to integrate new medical discoveries into their day-to-day work. In a blog, she wrote:
[P]aediatricians need to know about and critically appraise possible discrepancies in reporting when seeking reliable evidence from trials to guide their day-to-day clinical practice… We found major discrepancies… including changes in the original study hypotheses, trial designs, study conduction and reporting outcomes. These findings raise concerns about whether published scientific paediatric research can be accepted at face value.
Paola Rosati concluded:
As well as highlighting trial fallacy, our study should stimulate users to investigate how the clinical research protocols progressed from trial registration to publication… [and] stimulates developers of guidelines and systematic review researchers to use our scoring method, alerting journal editors to possible discrepancies.
In August 2016, a different team of researchers discovered that clinical trials involving over 69,000 children never had their results published in a peer reviewed journal, and that the findings of nearly a hundred of those trials remain completely unknown.
Recent data from other sources suggests that outcome misreporting is widespread in medical research. For example, the Oxford-based COMPare project, which also tracks discrepancies in the reporting of clinical trials, so far has checked 67 trials. Comparing trial registers with academic publications, project staff have discovered that across these 67 trials, 354 outcomes have remained unreported, while 357 new outcomes have been added retrospectively.