Synopsis: Clinical trials are much of the basis for the practice of medicine. Funding of trials has dramatically shifted so that today, trials paid for by pharmaceutical and device makers outnumber publically funded trials 6-to-1. Some studies indicate this makes bias in trials more likely, jeopardizing public health. Experts discuss.
Host: Reed Pence. Guests: Jocelyn Ullrich, Assistant Vice President of Science and Regulatory Advocacy, Pharmaceutical Research and Manufacturers of America (PhRMA); Dr. John Abramson, Lecturer in Healthcare Policy, Harvard Medical School; Dr. Joel Lexchin, emergency medicine physician and Professor of Health Policy, York University
Links for more information:
Industry funding of clinical trials
Reed Pence: Clinical trials are the basis for much of the practice of medicine. They test a new drug or procedure against current therapy, or against a placebo, so we learn which ones work, which ones don’t, and which ones might have dangerous side effects. But doing clinical trials is expensive. Extremely expensive.
Jocelyn Ullrich: We do know that pharma member companies spent an estimated $51.2 billon in 2014 to discover and develop new medicine.
Pence: That’s Jocelyn Ullrich, assistant vice president of science and regulatory advocacy for PhRMA, the Pharmaceutical Research and Manufacturers of America.
Ullrich: We have found that it takes about ten years on average, and about $2.6 billion of investments to bring a new medicine to market. And that $2.6 billion actually includes all the costs of failures. Less than 12% of the candidate medicine that enter phase one of clinical trials and human testing will actually make it as an approved medicine at the end of the day.
Pence: Big Pharma, as it’s often known, is funding a lot more clinical trials these days–the number went up by 43% between 2006 and 2014. At the same time, the number of trials paid for by the National Institutes of Health, the government’s principal funding agency, has dropped by 24%. The result is that the pharmaceutical and medical device industries now outspend the federal government on clinical trials by a margin of six to one. And it’s completely changed how clinical trials are done, according to Dr. John Abramson, a lecturer in healthcare policy at the Harvard Medical School.
Dr. John Abramson: It used to be that about 80% of clinical trials were performed in academic medical centers. Around 1991, 80% of trials were conducted by researchers in academic medical centers. Now that number is only 26%. So what’s happened is the pharmaceutical and device companies have taken over the funding of clinical research. They’ve also taken over the control of it and removed control from academic medical centers to themselves. So now rather than paying universities to conduct their research, they pay contract research organizations, for-profit research organizations, to conduct their research. And their sponsors get to dictate how their studies are conducted and the sponsors actually own the data.
Pence: It’s not that research universities don’t do clinical trials anymore. They do a lot of them, but more often now, they’re paid for by industry. And many experts say that the goal of industry-paid research is different compared to scientists funded by a more independent source such as the National Institutes of Health. Dr. Joel Lexchin is professor of health policy at York University in Toronto and an emergency medicine physician at one of the city’s large teaching hospitals.
Dr. Joel Lexchin: The pharmaceutical companies are in business to make profit, increase shareholder value. And in doing so, they’re designing their clinical trials to give the best possible results, which doesn’t necessarily translate into actually showing whether or not the drug is effective.
Abramson: The evidence is clear that industry-funded trials are significantly more likely to show that the sponsor’s product is both more effective and less harmful than non-industry sponsored trials.
Pence: Abramson says that doesn’t mean every study funded by industry is biased. But their generally more favorable view of drugs and devices is more than chance.
Abramson: There’s approximately a 27% greater likelihood that industry-funded studies will show efficacy and an 87% greater likelihood that industry-funded studies will show less harm than non-industry funded studies.
Lexchin: I’m not saying that all researchers are unethical. In fact, I would expect that most of them are highly ethical. But once you have a relationship with a drug company, that relationship starts to bias what you write and how you do your research, and often those biases are operating in a sub-conscious level and people don’t realize what they’re doing.
Pence: However, PhRMA’s Ullrich says that’s not necessarily true.
Ullrich: You know, I think conflict of interest is a very important topic of discussion. Research is an iterative process and there really needs to be collaboration between academic institutions and independent investigators and the industry in order to truly advance science. The assumption that because a researcher has received money from a company to conduct a clinical trial and therefore the results are no longer valid, I think, is maybe doing an injustice to the scientific process.
Pence: However, Lexchin says the scientific process is likely to be a little different in an industry-funded trial than in a government-funded one. He who pays gets to pick what you’re looking for and what a new drug is compared against.
Lexchin: They either compare their drug against placebo, which means that if it’s marginally better than nothing at all, it may get approved when in fact, that marginal benefit may, in clinical terms, may not mean anything. They also pick comparators, if they’re active drugs, they will pick comparators that either have more side effects than the product that they’re trying to test or that are less effective.
Pence: But Lexchin says doctors don’t want to see if a new treatment is better than no treatment at all. They want to know how a new treatment stacks up against the best that’s already available. And if a drug doesn’t do all that well, doctors want to know it. They don’t want to see bad news de-emphasized or even disappear.
Lexchin: If the studies are negative, they may not be published at all and if they are published, quite often when they are published, they show something completely different than the data that is given to the FDA. In other words, if the study given to the FDA shows that the drug is not effective, doesn’t work, when that study is published in a medical journal, somehow miraculously the data will show that the drug actually does work.
Pence: Obviously, data can be interpreted a lot of different ways, and Abramson says published studies are more likely to make sunny conclusions when the data is owned by a pharmaceutical maker. He says the drug company may even get to write the study, though a major research university’s name is on it.
Abramson: When you look at the contracts between the academic medical centers and the drug companies. If the study was published in The New England Journal of Medicine, when you look at those contracts which are going to give the most power and control to academic researchers, 50% of those contracts actually allow the drug company sponsors to write-up the article and they give the right to the researchers to review the manuscripts that the drug companies write. And the contracts say that manufacturers may accept revisions suggested by the researchers who actually did the research. 50% of these contracts give the right to the manufacturer to ghost-write these articles, the results of the research, to ghost-write them for the researchers who actually did the work.
Pence: However, Ullrich says misinterpretation of data isn’t in a drug company’s interest. The industry has a clinical trials code, and she says one reason drug companies turn to academic medical centers is the layer of safeguards built in to their research process, adding to those in place through the regulatory system.
Ullrich: The FDA has established, and Congress through legislative action has established, a very robust framework for how medicines are developed and review for approval. There are many, many safeguards along the way to ensure that research is being conducted in the most scientifically rigorous way possible. And in fact, when companies invest in research, they use the academic medical institution to conduct the research at their hospital site. Everything is overseen by independent institutional review boards. The FDA self-ensures that sites are being monitored by independent parties. They audit sites. So it’s very, very difficult to falsify data in a clinical trial and when that is done, investigators are found out and there’s actually a list with the FDA for investigators who have tried to falsify data.
Pence: Now you might think that peer review required by a medical journal might also weed out studies that come to exaggerated conclusions. But Lexchin, often a peer reviewer himself, says the data may not be all there. What you see may be only the data that’s needed to support the intended conclusion. And even that goal may have changed. Lexchin says the claimed purpose of studies is sometimes altered in the face of unfavorable evidence.
Lexchin: This is called in some places “data dredging,” so you go looking for positive results in the study whether or not those were actually the things you were supposed to be looking for and you write up the study as if the original intent was to discover what this data dredging unearthed and that’s very difficult for peer review to pick up. They other thing that peer review misses quite consistently is what’s called ghost writing. This is where the drug company who has run the trial hires a medical writing company to write up the results in such a way as to put the drug in a positive light.
Pence: What most people don’t consider is that medical research has two audiences, at least when a pharmaceutical company is seeking approval of a new drug. Doctors who read medical journals are one audience. They see an interpretation of the results by the study’s authors. The only outside entity that gets to see all the data is the other audience, the FDA, and they never release it. So we never know what the data really show behind the agency’s thumbs up or thumbs down.
Lexchin: All the FDA requires is two trials that show that the drug works so there could be ten trials and then only two show definite evidence that the drug works. That’s enough for the FDA but that may not be enough for practicing doctors. If you knew that out of ten trials that two show that the drug worked and eight didn’t, you may not way to prescribe that. But if those eight are not published or if they’re published with conclusions that have been turned 180 degrees, then doctors are more likely to use that product.
Pence: The federal database clinicaltrials.gov is also supposed to promote research transparency. Government rules say that all clinical trial results are supposed to be posted there. But Abramson says that the law is routinely broken and that major academic medical centers are even worse than drug companies about that. So what can be done that will actually work to make all data more available and make study results more trustworthy?
Abramson: There is a possibility that progress can be made if professional organizations and medical journals and physicians themselves step up to their professional responsibility to make sure that the American people derive maximum benefit from the clinical science that’s done, whether it’s commercially sponsored or publically sponsored. It doesn’t matter. The American people deserve to have access to the data from most clinical trials. And if medical journals whose fundamental responsibility in society is to provide information to physicians and the public and physicians themselves and the professional organizations accept their social role and their social responsibility, they could make a huge difference and turn this around.
Pence: Abramson says the British Medical Journal, for example, now won’t publish a study unless its authors make all data available. But other journals will have to join that stand for it to have any teeth. Lexchin says that sanctions against cheaters need more teeth, as well. He suggests that those found guilty could be barred from doing business with the federal government, banned from contracts with Medicare and the Veterans Affairs health system. And he says CEO’s should be made personally responsible if their firm is caught cheating, with the prospect of jail time. A funding firewall is another suggestion, funneling drug company research dollars through an impartial intermediary. Abramson isn’t optimistic government will be able to get tough or more inventive, but public health demands more oversight than we have now. If clinical trials are the way we decide how to practice medicine, there’s something in that for everyone.
You can find out more about all our guests through links on our website, radiohealthjournal.net. You can also find archives of our segments there, as well as on iTunes and Stitcher.
I’m Reed Pence.