Discrepancy between trial goals, results may mask treatment risks

(Reuters Health) - Scientists often fail to publicly register plans for clinical trials or to publish the results, and the outcomes they do share may mask instances when new treatments are unsafe or ineffective, a small study suggests.

One goal of asking scientists to register clinical trial plans in public research databases is to highlight the main objective of experiments and make it easy to see whether a tested treatment achieved this goal when results are published, the study authors note in JAMA.

But more than one-third of the 113 clinical trials researchers examined for the study were never registered. Only 64 of the trials, or 57 percent, were published.

One in five trials didn’t define a primary outcome, or clear protocols for determining if the tested treatment had achieved its main goal.

“Doctors rely on published evidence to guide patient care decisions, while researchers use the published literature to guide which promising areas of inquiry to pursue,” said lead study author Dr. An-Wen Chan of the University of Toronto.

“When research is hidden from view, then this evidence is unavailable to guide patient care or future research,” Chan said by email.

For the study, Chan and colleagues examined 113 clinical trials approved in Finland in 2007.

Overall, 61 percent of the trials appeared in public registries within one month of the trial start date.

Among the unregistered trials, 55 percent of the results that were published had discrepancies, such as a primary outcome that changed after the trial started or a lowered bar for determining whether the treatment met its main objective in the experiment.

For registered trials, only 6 percent of the published results contained these kinds of discrepancies between the pre-trial stated goals and the objectives described in the final publication.

The study was small and limited to research done around the Helsinki region in Finland, making it possible that the findings don’t reflect how scientists work in other places, the authors note.

Even so, the findings add to growing evidence highlighting a lack of transparency in clinical trial results, said Jonathan Kimmelman, a biomedical ethics researcher at McGill University in Montreal who wasn’t involved in the study.

“Unfortunately, it is easier to access positive results than negative ones,” Kimmelman said by email.

Unpublished trials tend to be the ones that found treatments didn’t work or weren’t safe, said Roberta Scherer, a researcher at Johns Hopkins Bloomberg School of Public Health in Baltimore who wasn’t involved in the study. Published results, meanwhile, tend to highlight successful experiments.

This can put patients at risk, Scherer said by email.

“The worse-case scenario is that patients are receiving an ineffective treatment or possibly even a harmful treatment,” Scherer added.

Without complete transparency, it’s also hard for patients to make informed decisions, said Dr. Keith Ferdinand of Tulane University School of Medicine in New Orleans.

“Public reporting is essential to advance science and protect the public,” Ferdinand, who wasn’t involved in the study, said by email.

Patients may not understand that publication doesn’t guarantee they’re getting complete information about a treatment, said Dr. Joseph Ross, a public health researcher at Yale School of Medicine in New Haven, Connecticut.

“It often takes years for sufficient evidence to accrue to support an intervention’s efficacy and safety, and it is difficult to differentiate good and complete information from bad and incomplete information, even when published,” Ross, who wasn’t involved in the study, said by email.

For patients who want to do their own homework about their treatment options, systematic reviews that examine results from a variety of clinical trials are the best way to go, said Dr. Barbara Mintzes, a pharmacy researcher at the University of Sydney in Australia who wasn’t involved in the study.

The Cochrane Collaboration, for example, publishes reviews of drug treatments with summaries of key findings, good and bad, that may help guide patient decisions, Mintzes said by email.

“These systematic reviews combine all available scientific evidence on a subject, including a thorough attempt to find unpublished as well as published studies,” Mintzes said.

SOURCE: JAMA, online September 11, 2017.