TURIN, Italy-A massive failure globally to publish phase III clinical study findings in radiotherapy was revealed at the 2016 European Society for Radiotherapy and Oncology (ESTRO) Annual Meeting from analysis of 802 trials due for completion by January 2013. More than 75 percent of studies did not publish data as planned (Abstract PV-0087).
This was in spite of the fact that since 2007 U.S. law has effectively mandated release of such data for all trials around the world within 12 months of any trial's planned completion.
"These findings came as a surprise for many reasons, not least of which was that many had been funded by the U.S. National Institutes of Health. It is particularly worrying that the law is being ignored on such a wide scale," said lead author Jaime Perez-Alija MSc, Medical Physicist from the Hospital Plato in Barcelona, Spain.
The investigators concluded a large number of patients with cancer have been exposed to the risks of trial participation without being able to give any benefit to future generations of patients by sharing data and publishing results.
Reasons for Not Publishing
A subset of the studies begun after 2007 was analyzed separately to compare practice before and after the U.S. legislation came into force. There was almost no difference. In the overall study sample, a total of 655 trials (81.7 percent) did not deposit a summary result within 12 months. And 422 out of 552 (76.4 percent) of clinical trials starting after the 2007 law came into force also failed to publish a deposition of their results.
Perez-Alija added that half of the studies had not reported any findings at all. "So this evidence seems to be lost forever," he noted.
One possible reason for non-publication, he told Oncology Times, was that some trials may have been granted deadline extensions that were not made public. "Therefore, our first problem is that we do not know with any certainty whether a trial is truly overdue," he said, adding that the U.S. registry stipulates that all dates must be updated if an extension has been allowed. "But it seems likely that this is not happening in many cases," he said.
He suggested there could be publication bias since there were several reasons why it was easier to publish a positive finding than a negative one.
"If you got a positive finding, you'd run to a big, important journal, maybe [get] a better career or get a grant," he noted. But he suggested negative findings could have the opposite effect. "Maybe you feel insecure about yourself [thinking]: Maybe I screwed up? Maybe I didn't do the things right? Or: Nobody cares about my trial!
And he explained that it seemed to be impractical to police the 2007 law that has been ignored despite the obligation it carries to make results available to everybody. "It's not like a peer-reviewed journal. Everyone-a patient [enrolled in the trial]-could go there and see what happened in the trial," said Perez-Alija. Yet, this has been ignored with impunity, possibly, he suspected, because enforcement could discourage the launch of clinical investigations in the future.
But, he was disturbed by a number of issues. "Most patients enrolled in these trials have taken the risk without the benefits." And he added there were also economic matters causing health systems inefficiencies. "Doctors may be giving a treatment they think is well-suited for patients. But there [may be] other [unpublished] trials that say it's not so well-suited," he said. He also was concerned about the failure to carry out good science by having access only to distorted evidence of simply not knowing about other potentially contradictory data.
Need for Evidence-Based Guidelines
Former ESTRO President, Michael Baumann MD, Chair of the Department of Radiation Oncology at the University of Dresden in Germany, who did not participate in the study, told Oncology Times he shared Perez-Alija's ethical and scientific concerns, though he could not yet tell what effect failure to publish could have had on current practice.
"At the moment, we only have the fact that there's under-reporting of studies. And that by itself is of concern," he said, adding that he, too, was worried about the risk of abusing informed consent given by patients who volunteer altruistically to help doctors improve the outcome of radiotherapy for future generations of patients. "Studies need to be reported. That is a must," he said.
When he was asked how to rectify the situation, he insisted that all studies need to be registered, with scope for updates if they are running longer than expected. "But when they have been finished, we should expect a report of the outcome," he said. And he added that doctors participating should encourage their principal investigators to report data.
He said under-reporting occurred more often among trials that were not industry supported.
"Obviously, studies which are well-funded by industry are reporting at a higher percentage than studies which are not," he said. And he added that not all studies were of interest to industry and that there also needed to be independent studies. "So maybe we have to face the fact that there's not enough funding for studies which are not industry-driven. And that would be of concern for policy makers," he said.
Baumann pointed to the paramount importance of evidence-based guidelines. "But evidence-based guidelines might be affected by under-reporting. So we have to make sure that studies which can be reported are reported," he said.
Peter M. Goodwin is a contributing writer.