For-profit vs. non-profit: Sponsor type affects study outcomes
Experts analyze findings and talk about changes
It may not surprise most people in the clincial trials industry, but a new study says that when cardiovascular trials are funded by for-profit sponsors they are more likely to report positive findings.1
Experts say the reasons related to bias are partly systemic.
"I think it's a cultural problem right now," says Barbara J. LoDico, BS, CIP, executive director of the office of human subjects protection at the University of Medicine and Dentistry of New Jersey in Newark. LoDico will be the new chair of the regulatory and ethics forum for the Association of Clinical Research Professionals (ACRP).
"Whether it's in a private office or private research center or an academic center, contract research's goal is to reach a data set point for the sponsor," LoDico says. "And with the lack of investigator input into the protocol, it's a different paradigm."
The new study looked at reported outcomes of 349 randomized cardiovascular clinical trials, which were published between 2000 and 2005 in the Journal of the American Medical Association (JAMA), the New England Journal of Medicine, and The Lancet.1
Of the 349 trials, 31% were financed by not-for-profit organizations; 44% were financed by for-profit pharmaceutical or device manufacturers, and 19% were financed jointly by for-profit and not-for-profit organizations. The remaining 6% noted no funding source.1
The published results were rated as positive or negative and then analyzed according to the funding source of the trial. Investigators found that 49% of the not-for-profit studies reported evidence significantly favoring newer treatments, while 67.2% of the for-profit trials showed evidence significantly favoring newer treatments. Of the trials funded by both for-profit and not-for-profits, the percentage was midway between the other two groups at 56.5%.1
The study's authors conclude that incentives associated with for-profit sponsors have the potential to influence clinical trial outcomes.1 But they also point to other important factors, such as the reality that truly novel therapies that have initial negative results are unlikely to be funded later by for-profit entities and that for-profit sponsors often are interested in evaluating proven therapies in understudied patient populations.1
The more optimistic results reported by for-profit entities probably has more to do with study design than with interpretative bias, says LaDale George, JD, a partner of Foley & Lardner in Chicago and lead of the translational research group at Foley.
"The funder has a great impact on the design of the study, setting out end points and inclusion and exclusion criteria," George says. "And it's more narrowly tailored to try to identify a positive outcome, but that's by design."
Unless studies were designed in a way that did not limit the criteria of patients and would provide a broader risk and efficacy profile for the study product, then there always will be this issue, George says.
"The design by the for-profit sponsor and the identification of the end point is market-oriented," he says.
"It's not just a design to show the product is more efficacious than the other in general, it's designed to show it's more efficacious in a specific manner, which is going to a particular market opportunity for that product."
Any regulatory changes that take away this market incentive for research would back fire, experts predict.
"Who's going to fund this research if the pharmaceutical industry, in whose interest the research is being conducted, doesn't fund it?" says Ellen Hyman-Browne, JD, MPH, CIP, director of research compliance at The Children's Hospital of Philadelphia. Hyman-Browne, George, and LoDico are members of the editorial advisory board of Clinical Trials Administrator.
"We don't have a choice in how the research is funded," Hyman-Browne says. "But there needs to be a partnership between the pharmaceutical industry and research institutions to work towards maintaining the trust of the public, and I think there are ways we can improve on that."
Research is expensive, and no academic medical center or educational institution could ever fund on their own the type of research that needs to be done, LoDico says.
"They are really dependent on a combination of federal and private sponsorship," LoDico says.
The controversy that continues over the arthritis painkiller rofecoxib (Vioxx) highlights the problem of bias in study design and analysis.
Since Merck pulled rofecoxib off the market in 2004, reports have shown that clinical trial evidence showed some cardiovascular risks among trial participants.
According to a recent article in The New York Times, these problems were even evident from data in people who took the drug for as little as four months, despite Merck's assertion that people who took the drug for less than 18 months were not at risk.2
The FDA had approved the use of Vioxx for no more than 18 days of continual use, but physicians were prescribing it for 18 months or more, George explains.
"All COX-2 selective inhibitors ran into problems once they reached the marketplace," George notes. "The problem is the lack of association between FDA-specific use approval and physician practice of medicine exception."
Initially, no one was checking the data on the back end of physician use as thoroughly as they should have, he says.
The published studies about rofecoxib did not address long-term use, despite the clear market advantage to Merck of having a drug that people would use for more than 2.5 weeks at a time.
In this case, the omission of important data became a safety issue. And it's because of the Vioxx controversy and similar recent occurrences that research transparency is gaining momentum.
"Part of the problem is that in this country we don't have a governmental process that approves drugs," George says. "We have a governmental process that examines data that is collected in the private sector and uses that data to base its approvals on, but there is not an independent government testing of the product."
Whether the government should be involved in the direct testing process is a big philosophical debate, George adds. "I don't know if there should be an independent government body that runs clinical trials — we'd grind to a halt if that was the case," George says.
The climate in today's clinical trials industry is much more favorable than it has been in the past for honest reporting of research data, Hyman-Browne notes.
"This is not something we can expect the government to take over — not that that would be completely unbiased either," Hyman-Browne says. "So I think we have to work with the situation we have."
More transparency needed
The goal is to add more transparency to research, and most academic medical institutions contribute to this objective through their efforts to preserve the right to publish, Hyman-Browne notes.
Not-for-profit academic medical institutions work to monitor conflicts of interest and data integrity in their efforts to maintain research education service goals, she says.
"Publishing is very important in an academic medical center when clinical trial agreements are negotiated," Hyman-Browne says. "Every center I'm aware of is very careful to protect the integrity of the trial and the research and the public service nature of what they are doing, so they're not working for hire for the pharmaceutical industry."
The new emphasis on a clinical trials registry, required in some cases by the FDA and by medical journal editors, also has helped to improve transparency in human subjects research.
"Many institutions now have guidelines for clinical trial agreements that ensure the trial will be registered," Hyman-Browne says. "And some research institutions also have policies of registering all trials conducted at their sites on their own web sites."
Public disclosure of clinical trials is important, LoDico says. Research journals increasingly are publishing studies that show negative results, a trend that is needed to provide a balance, she adds.
The National Cancer Institute has a Clinical Trials Cooperative Group Program that provides one example of a model that could be a solution for eliminating bias in clinical trials. Under the cooperative group model, protocols are developed based on input from all of the investigators who will be conducting the studies, as well as the pharmaceutical sponsor or device manufacturer, LoDico explains.
"So cooperative groups, like AIDS clinical trial research, come up with the best possible protocol to answer the scientific question, as opposed to the purely industry sponsored research where the only question is, 'Will this drug in this trial be good enough to allow the FDA to approve a marketing application?'" LoDico says.
In the traditional model, the design of the study and the analysis are done by the sponsor and local sites have no input, LoDico says. While the cooperative group model may be ideal, it is very time-consuming and expensive, she notes.
"I think the research industry recognizes that the cooperative model is actually the best, but how pure does it need to be?" LoDico says. "I don't think you'll get an agreement on this by anyone."
Reference
- Ridker PM, Torres J. Reported outcomes in major cardiovascular clinical trials funded by for-profit and not-for-profit organizations: 2000-2005. JAMA 2006;295(19):2270-2274.
- Pollack A, Abelson R. Why the data diverge on the dangers of Vioxx. The New York Times. May 22, 2006;C1.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content