Only about 29 percent of completed medical trials conducted at major American academic centers lead to published results within two years, according to one of the most detailed analyses of the problem.
The findings, published on Wednesday in BMJ, suggest that universities and their funders still are falling well short on a major yardstick of open science and of responsibility to participating patients.
“The academic institutions are doing very little about this — nothing, in fact,” said a lead author, Harlan M. Krumholz, a professor of medicine at Yale University.
The general problem is recognized, said Deborah A. Zarin, director of ClinicalTrials.gov, the government’s official website for reporting plans for medical trials and their subsequent results. For more than a year, Dr. Zarin said, federal officials have been working to establish new rules and penalties for violations, and the process should be finished soon.
But the solution will probably require more than tougher rules, both Dr. Krumholz and Dr. Zarin said, and must involve a change in fundamental attitudes at medical centers. To that end, they said, the listing of individual universities in the BMJ analysis might generate public pressure for change.
For their analysis, Dr. Krumholz and his colleagues tallied all interventional trials registered on ClinicalTrials.gov with a primary completion date from October 2007 to September 2010, then saw how many either published the results in a journal or reported the results to ClinicalTrials.gov within 24 months of that completion date. Their report in BMJ covered 51 institutions — those with at least 40 completed trials registered on the government site during that time.
The average was 36 percent, with the institutions ranging from the University of Minnesota-Twin Cities, with a reporting rate of 55 percent, to the University of Nebraska Medical Center, at 16 percent. Just counting trials that led to published reports in journals within 24 months, the average among all 51 institutions was 29 percent. Previous studies have pointed out the problem, though this appears to be the most detailed institution-by-institution examination, Dr. Zarin said.
Nebraska’s associate vice chancellor for clinical research, Christopher J. Kratochvil, said that he recognized the problem and that his institution was working to improve its record. Still, some delays will be inevitable, Dr. Kratochvil said, because of the time needed to analyze data from a trial, draft an article, and engage in the back-and-forth of the peer-review process. And some major journals might reject an article if the data are placed first in the ClinicalTrials.gov public repository, he said.
Other considerations perhaps not captured in the BMJ analysis include the fact that data from a trial are often presented to the scientific community ahead of publication, such as at medical conferences, Dr. Kratochvil said in a written response.
“All of this being said, we as a scientific community need to continue to optimize the speed with which we disseminate clinical-trial results, and to do so consistently,” he said.
‘Effort and Sacrifice’
About a half-dozen co-authors of the BMJ study spent perhaps 150 hours apiece digging through records to identify the trials and figure out which had been reported, said one of them, Ruijun Chen, a resident physician in internal medicine at the University of California at San Francisco.
Dr. Chen, who already works about 80 hours a week in his residency, said he had participated because he understands “how much effort and sacrifice” both researchers and patients invest in a clinical trial. Some researchers may feel that reporting is not worth the bother if their study is not a 10,000-patient trial that ends up in The New England Journal of Medicine, Dr. Chen said.
They may not realize the value that even a small trial may have, Dr. Chen said. He said he had learned something that might help him with a patient someday, for instance, from a group discussion of a small trial examining a rare form of an eye cancer.
The new federal policies, when they take effect later this year, will carry penalties that include a potential loss of eligibility for federal grants, and fines of up to $10,000 a day for institutions that fail to publish results of clinical trials within set periods. The rules would cover about 60 percent of clinical trials, mostly because they involve drugs or devices, or have federal grant support.
It’s not clear that such threats alone will be sufficient, said Dr. Krumholz, who also serves on the Advisory Committee to the Director at the National Institutes of Health. The culture within academe has evolved to the point where many researchers regard reporting as optional, and that needs to change, he said.
Dr. Krumholz said he had asked for help on that from the Association of American Medical Colleges, whose members include all 145 accredited U.S. medical schools. “They weren’t interested,” he said. An AAMC spokeswoman, Jamila Vernon, said the association had nobody available to discuss the BMJ study. She instead provided a written statement attributed to Heather Pierce, the association’s senior director for science policy and regulatory counsel, saying the AAMC agrees on the need for improvement in reporting results of clinical trials, and is working on ideas that include the development of new data-reporting systems and outreach campaigns.
“We support publishing results of clinical trials,” Ms. Pierce said, “but we also believe it is important to publish information that is meaningful, accurate, and complete.”
Corrections (2/18/2016, 1:26 p.m.): This article and the accompanying table originally misnamed the University of Nebraska unit analyzed in the BMJ study. It is the University of Nebraska Medical Center, not the University of Nebraska at Lincoln. Both the article and the table have been updated to reflect this correction. The article has also been revised to clarify that the 29-percent figure applies to the rate of publication in journals, while the 36-percent rate includes both journal publications and posting of results on the ClinicalTrials.gov website.
Paul Basken covers university research and its intersection with government policy. He can be found on Twitter @pbasken, or reached by email at paul.basken@chronicle.com.
