That lack of accessibility not only wastes precious research dollars, but is also potentially harmful to patient care, argues a group of researchers, who are calling for changes to ensure study results are fully reported in medical journals.
"Overall, half of health research, nothing from it is made public, not even the basic results or the basic journal paper," said Dr. An-Wen Chan, a scientist at the Women's College Research Institute in Toronto.
And when research is published, a study's goals, methods and findings are usually summarized, resulting in significantly incomplete information, said Chan, lead author of one in a series of articles on research waste published in this week's Lancet.
"There are hundreds to thousands of pages of research information that describe how the study was done, as well as reporting all the analyses. And that is compressed into a journal paper that's less than 10 pages," he said.
"So clearly there's some loss of information. And when you select what information to report, as with any journal, the most interesting information is selectively reported."
Chan said the strongest predictor of a medical study being submitted for publication is finding positive results for a drug, medical device or other intervention being tested.
When results of research are negative — in other words, the drug or device didn't prove to be beneficial — researchers often don't bother to submit their study to a journal, he said.
But negative results are highly valuable in themselves, Chan stressed.
"To know that something doesn't work and we should not be treating our patients with a particular intervention is important to know. It wastes money, it wastes time and they're potentially harmful if they don't work."
Chan and his co-authors from the U.S., Europe and Australia cite several examples of incomplete study information leading to wasted health-care spending or harm to patients.
For instance, governments worldwide spent billions of dollars on Tamiflu (oseltamivir) to treat influenza, especially during the 2009 H1N1 pandemic.
But the decision to stockpile the drug was based on inadequate information, including clinical trials in which almost two-thirds of patient data were not reported. As it turned out, Tamiflu did not necessarily reduce hospital admissions or complications from flu and its harmful side-effects were unclear.
"It's really been based on questionable evidence of whether it works or not, questionable in the sense that regulators and independent researchers are unable to fully evaluate whether it works or not because not all the data are available," Chan said.
In the case of the blockbuster painkiller Vioxx, exclusion of data by its maker Merck from studies submitted to journals and to government regulators resulted in thousands of patients who took the drug having heart attacks or strokes.
A subsequent analysis by independent researchers that included full data from all patient trials of the drug — obtained through a lawsuit — revealed the risk of heart attack and stroke, leading to Vioxx being pulled from the market by Merck in 2004.
Those scientists were able to show the drug's inherent dangers only because they obtained access to complete information, Chan said.
"The tens of thousands of excess heart attacks and deaths that have been estimated could have been avoided."
In the Lancet article, Chan and co-authors make the following recommendations:
—Research institutions and funding agencies should provide incentives that reward scientists who provide full dissemination of study materials.
—Researchers, funders, sponsors, government regulators, research ethics committees and journals should develop and adopt standards for the content of full study reports and for data-sharing practices.
—These groups should endorse and enforce study registration policies, wide availability of full study information, and sharing of all patient-level health research.
Dr. Andreas Laupacis, executive director of the Li Ka Shing Knowledge Institute at St. Michael's Hospital in Toronto, said the article highlights an important issue that the public should be concerned about.
Patients often agree to take part in research trials because they're told it could benefit society or future patients, said Laupacis, who was not one of the authors.
"If we don't then publish results, the benefit's not nearly what it should be and my sense is that (people) might be less interested in entering trials," he said, adding that it also breaks an "ethical contract" made between researchers and patients.
"I think it's something that researchers will need to solve themselves, hopefully with a bit of a push from the public."