More clinical trials in kids? Nearly half are unfinished or unpublished

pediatric trials clinical trials
Of 559 interventional trials in children, 19 percent were stopped early and 30 percent of completed trials remained unpublished several years later, finds a new study. (Vmenkov/Wikimedia Commons)

Recent laws like the Best Pharmaceuticals for Children Act and the Pediatric Research Equity Act are encouraging clinical trials in children. Yet, as with adult trials, these trials commonly stall out or, if completed, remain unpublished several years later, finds a study published online today in Pediatrics.

“Our findings may speak to how commonplace discontinuation and non-publication are in medical research in general,” says Natalie Pica, MD, PhD, a senior resident at Boston Children’s Hospital and the study’s coauthor. “We need to make sure that when children participate in clinical trials, their efforts are contributing to broader scientific knowledge.”

Pica and senior investigator Florence Bourgeois, MD, MPH, of Boston Children’s, tracked 559 randomized, controlled pediatric trials registered on ClinicalTrials.gov from 2008 to 2010 and whose final status (completed or discontinued) was confirmed by the end of 2012. All tested interventions such as drugs, devices, procedures and behavioral interventions.

pediatric trials clinical trialsAn estimated 8,000 children were enrolled in discontinued trials and more than 69,000 in trials that were never published after an average of 58 months.

They then searched the medical literature for matching peer-reviewed publications through September 1, 2015. If they couldn’t find a publication, they reached out to study investigators and sponsors via email.

Overall, 19 percent of trials were discontinued early — two thirds of them after enrolling participants. Of the completed trials, 30 percent had not been published after an average of 58 months.

Pica and Bourgeois see this as a tremendous waste. They calculate that more than 8,000 children were enrolled in discontinued trials and more than 69,000 children in trials that were never published, at least as of their cutoff date.

“We need to make sure that the proper resources are in place to ensure that information gleaned from these studies reaches the scientific community,” says Bourgeois.

Industry vs. academia

Similar to studies in adults, industry-sponsored trials were more likely to be completed than trials sponsored by academic institutions. But, once completed, industry-sponsored trials were less likely to be published.

In a multivariate analysis, the odds of non-publication were more than doubled for completed industry-sponsored trials two years post-completion and more than tripled three years post-completion when compared with academia-sponsored trials.

Roadblocks to completion

Why are so many trials going unfinished or unpublished? Through data provided on ClinicalTrials.gov and follow-up emails with study investigators, the researchers delved into the 104 discontinued trials. In 12.5 percent of cases, the trials were halted for good reason: the drug or device raised safety concerns or was so effective or ineffective that to continue the trial would be unethical. But other reasons included:

  • Problems with enrolling enough patients (36.5 percent)
  • Technical or logistical issues compromising trial completion (12.5 percent)
  • A company/business decision (8.7 percent)
  • The principal investigator left the organization (7.7 percent)
  • Regulatory issues with IRBs, the FDA or other bodies (7.7 percent)
  • Funding issues (4.8 percent)

For the remaining 9.6 percent, the reason for trial discontinuation was either unreported or unclear.

Publish or perish?

The study didn’t formally query or tabulate reasons for trial non-publication, a topic ripe for exploration. Non-publication was common even for academia, which accounted for 30 percent of unpublished trials two years after completion and 23 percent of unpublished trials three years out.

Pica
Pica

While section 801 the FDA Amendments Act of 2007 requires investigators to report study results, they can fulfill this mandate by simply reporting on ClinicalTrials.gov. While 42 of 136 unpublished trials did so, Pica and Bourgeois see posting to ClinicalTrials.gov as a half-measure, not nearly as helpful as a full peer-reviewed scientific publication.

“[The] data were not always interpretable because they were often without statistical analysis or without clarification of the hypothesis tested, making it difficult to draw substantive conclusions,” they write.

Bourgeois
Bourgeois

Non-publication has also been found by others. A study from the University of Minnesota, published in 2012, searched ClinicalTrials.gov for closed studies involving children, registered from 2000 through 2010. Of 758 completed studies, just 29 percent were published, an average of two years after completion. Of 160 randomly selected NIH–funded studies, 53 percent were published. The study also found that completed interventional studies were published more often than observational studies.

A Yale-led study published earlier this year tracked 4357 interventional adult trials registered on ClinicalTrials.gov across 51 academic medical centers. All trials had a primary completion date between October 2007 and September 2010. As of July 2014, 56.5 percent had been published, and 29 percent were published within two years of completion.

Data recovery

Can this situation be salvaged? One proposed initiative is RIAT (Restoring Invisible and Abandoned Trials). Endorsed by the BMJ, PLoS ONE and BioMed Central, RIAT calls on researchers with unpublished trials to either publish within a year or provide public access to their data. Independent investigators can then step forward to become “restorative authors.”

A website called Alltrials.net is collecting examples of researchers who rescued the results of old trials. RIAT studies also reanalyze primary data from published studies, as a 2015 BMJ paper famously did for a 2001 study of antidepressants in adolescents.

“It’s hard to reanalyze others’ data,” says Pica, “but this may be a useful mechanism to make sure that findings from completed trials are disseminated in the medical literature.”

For more on the study, check out NPR and Forbes. The study was supported by the National Institute of Child Health and Human Development (1R21HD072382) of the National Institutes of Health and the Fred Lovejoy House-Staff Research and Education Fund at Boston Children’s Hospital.