The Ethics of Business-Run Clinical Trials

Assume for a moment that you are overweight and have high blood pressure. Before writing a prescription, your doctor asks if you would like to take part in a clinical trial for a new drug that holds out the promise of lowering blood pressure faster and more effectively than existing medications.

"Does it work," you'd probably ask. And your doctor would answer something like this: "I don't know. That's why we're doing the clinical trial. But don't worry, there is no placebo arm in this trial because we couldn't ethically deny you the standard of care since there are already so many blood pressure control medicines out there that we know work."

"So I will get either an existing, effective treatment or the existing treatment plus this new drug?"

"Yes."

"So what's the point? How will you know if it works?"

"Well, what we hope to learn is whether it works faster or better."

"Oh, I see. Sure. I'll take part."

Thousands of people sign up for clinical trials on these terms every day. It's especially true in cancer clinical trials, where desperately ill patients are anxious to get their hands on the next drug in the pipeline without giving up the not-very-good but better-than-nothing standard of care.

But let's change the scenario slightly. As a patient, what whould you do if you had very strong suspicion that the new drug was going to turn out better? Indeed, what would you do if you knew, in a statistical sense, that it was going to turn out better? Would you willingly go into a randomized clinical trial where you had a 50-50 chance of getting the old standard of care?

"If I'm 90 percent sure a new treatment will be better, would I be randomized to a 50-50-chance? Of course not," said Benjamin Djulbegovic, a professor of medicine at the University of South Florida, who has devoted his career to applying the mathematics uncertainty principle to clinical trial ethics. "Ethicists and scientists have said you shouldn't do trials where you have a good idea what the outcome will be. You'd be exposing people to inferior treatment, which is unethical."

Yet, based on an article that appeared in today's Journal of the American Medical Association, people do it every day when they enter industry-sponsored clinical trials.

The study compared the past five years' results from industry-sponsored clinical trials on various cardiovascular medicines and procedures to the results of similar trials sponsored by non-profit organizations like the government or foundations. The authors, Paul Ridker and Jose Torres of Brigham and Women's Hospital in Boston, limited their research to the three leading medical journals -- the New England Journal of Medicine, The Lancet and JAMA -- so they would only consider "high quality" studies.

The results were not surprising, since any number of previous studies have reached similar conclusions. Fully 67.2 percent of the industry-sponsored trials turned out to favor the new product or procedure over the standard of care compared to just 49 percent of the studies funded by non-profit entities.

The article contains a good discussion about the various biases that may have contributed to these results. But I want to focus on just one issue physicians should begin thinking about if they take its findings to heart: It is ethical to put a patient in an industry-funded trial?

The math, according to Djulbegovic, is unassailable. If a clinical trial is well-constructed, its architects won't know what the outcome will be. In statistics, this is known as the uncertainty principle and suggests that over time, and with enough examples, the distribution of answers to a binary question (is it better or is it worse) should hover around 50 percent on each side. And as we can see from the latest Ridker-Torres study, that's precisely what happened when they looked at 101 randomized clinical trials conducted by non-industry-funded scientists. These scientists were more likely to be asking questions and picking subjects for randomization in an unbiased manner, thus ensuring that the ultimate outcome was, at the outset, truly unknowable.

But that isn't what happened with 137 industry-funded trials in the study. With a two-thirds success rate, the odds were more like the House in Las Vegas than a coin-toss. "Incentives surrounding for-profit organizations have the potential to influence clinical trial outcomes," the authors wrote.

"There's all kinds of problems," said Djulbegovic. "If you select a comparator you know is inferior, you are violating the uncertainty principle. Then there's profit-driven me-too drugs and poorly designed trials. Plus they don't publish unfavorable results."

That latter factor actually works to drag the industry-funded trial ratio back toward the expected average. But scientists -- and patients and regulators -- never benefit from that knowledge because it doesn't enter the public domain.

The cure for this disease is to have independent scientists test new drugs and devices with all outcomes, whether positive or negative, made publicly available. If medical journals aren't interested in negative trials, then the results can at least be posted on websites and on the government's clinical trials database.

This shouldn't be an argument that hinges on what one thinks about "big business" or "the drug industry." Given the growing body of evidence about industry-funded clinical trial outcomes, it is an ethical imperative.

Posted by gooznews at May 17, 2006 06:34 AM