• Only 29% published in journal in two years, 13% on registry
  • Dissemination varies widely among 51 academic medical centers

One-third of clinical trials conducted at 51 major U.S. universities and academic hospitals were never published in a peer-reviewed journal or in a government registry online, according to a new study in the BMJ, formerly the British Medical Journal.

The researchers looked at 4,347 trials that were completed between October 2007 and September 2010. Of those, only 29 percent had results published within two years of finishing data collection, and 13 percent were posted on the government database ClinicalTrials.gov within the same period, the study found. Overall, about 67 percent of the studies disclosed their results by July 2014.

Researchers examined the publications rates at academic medical centers, including Stanford University, Dana-Farber Cancer Institute and Cornell University. Results varied widely among the institutions, from 11 percent of trial results at the University of Nebraska being published in medical journals to 40 percent at Yale University within 24 months after tests had finished.

“You’d like to think that academic institutions are role models for science,” said Yale Medical School Professor Harlan Krumholz, a co-author of the study. “The truth is we’re not doing very well.”

The study found that University of Florida and Ohio State University joined Yale with some of the highest rates of publication in peer-reviewed journals in the same time period while Boston University, Cornell and Oregon Health and Science University, or OHSU, had among the lowest rates.

Negative Results

When asked about the results, David Ellison, a professor of medicine at OHSU, said the university doesn’t currently track publication data, though he expects that will change. 

“I think everybody in the country will be focused on this over the next couple of years,” he said in a telephone interview. “I think this is an important issue and we need to do better.”

Ellison said researchers may not publish their results if the data isn’t positive or interesting, noting that it’s harder for studies with negative results to be accepted by medical journals. He said that regardless of the outcome of tests, it was still important to publish. 

“Pharmaceutical companies, the government and foundations invest a lot of money in clinical trials,” he said. “If the results aren’t made public, that’s wasted money.”

Despite having among the lowest rates of publication in medical journals, the study found that OHSU was in the middle of the pack for overall release of results, with about 68 percent shared publicly.

Cancer Studies

About a third of the trials studied focused on cancer, the most for any disease, while mental health and cardiovascular diseases were the second- and third-most tested. The National Institutes of Health sponsored 9.8 percent of the trials studied, and industry backed about 12 percent.

The International Committee of Medical Journal Editors requires that all studies published in medical journals are registered in a public database before the trial starts. It encourages, but doesn’t require, that results be published in a public registry as well.

“Disclosure of results is critically important to advance science,” Chris Kratochvil, associate vice chancellor for clinical research at the University of Nebraska Medical Center, said in a statement. “Whether those results are positive or negative, and are important as a part of our agreements with funders as well as the participants in clinical trials.”

Ethical Issue

Pharmaceutical companies have come under scrutiny in the past for not publishing their results, though Krumholz’s team found the problem is also widespread throughout academia. Beyond creating a data void that makes future research more difficult, unpublished results also may represent an ethical issue. Patients in clinical trials undergo experimental treatments with the understanding that doing so will advance medical knowledge, which can’t happen if the results aren’t published.

“It’s an obligation, an ethical obligation, to report results,” Krumholz said. “My hope is that this report card is shocking enough that it moves people to say that we need to fix this problem.”

Before it's here, it's on the Bloomberg Terminal. LEARN MORE