Healthy Skepticism Library item: 8760

Warning: This library includes all items relevant to health product marketing that we are aware of regardless of quality. Often we do not agree with all or part of the contents.

Publication type: news

Underwood A.
Too Good to Be True?
Newsweek 2007 Feb 26
http://www.msnbc.msn.com/id/17344198/site/newsweek/

Full text:

A new analysis finds that clinical trials of breast-cancer treatments by drugmakers show positive results far more often than non-industry trials. Why the difference?

When clinical trials for new drugs are published, they have a tendency to show positive results-some critics say, too positive. Questions of bias have been raised repeatedly, particularly in trials with drug-company funding. A new study, appearing Monday in the online version of Cancer, the journal of the American Cancer Society, analyzes 140 trials of breast-cancer drugs. The bottom line? In 2003, a whopping 84 percent of trials with pharmaceutical-company involvement showed positive results, compared to just 54 percent for trials without industry backing. The study was led by Dr. Jeffrey Peppercorn, assistant professor of medicine at the University of North Carolina Lineberger Comprehensive Cancer Center, along with three researchers at the Dana-Farber Cancer Institute in Boston. NEWSWEEK’s Anne Underwood spoke with Dr. Peppercorn.

Excerpts:

NEWSWEEK: How did this study come about?
Dr. Jeffrey Peppercorn: As clinical researchers and breast-cancer specialists, we recognize that an increasing number of studies are developed through collaborations between academic medical centers and drug companies. In general, this has been a very productive collaboration. However, the amount of funding from pharmaceutical companies now exceeds that from the National Institutes of Health. In fact, pharmaceutical-industry investment in research exceeds the entire operating budget of the NIH. That makes it important to understand the influence that industry involvement may have on the nature and direction of breast-cancer research. We need to know if there are important clinical issues that aren’t being addressed.

How did you conduct this study?
We selected 10 leading journals for breast-cancer research and attempted to identify every breast-cancer clinical trial for a medical therapy that appeared in 1993, 1998 and 2003. We identified a total of 140 studies.

What did you find?
There were essentially two major findings. First, in 2003, as compared to 1993, there were more studies with drug-company involvement-either drug-company funding or actually having a drug-company scientist as one of the authors. In 2003, 58 percent of studies reported pharmaceutical involvement, versus 44 percent in 1993. But that may be because of more stringent disclosure requirements in recent years. Second, studies backed by pharmaceutical companies were significantly more likely to report positive results. In 2003, the likelihood of positive results was 84 percent for studies with pharmaceutical involvement, versus 54 percent in studies without clear industry connections.

Our analysis was small, consisting only of 140 trials, but similar associations have been documented before, in stroke trials, psychiatry trials, cardiovascular trials and several other areas of clinical research. The one previous study in oncology, looking at multiple myeloma, also found that pharmaceutical studies reported positive results in 74 percent of trials compared to 47 percent of non-industry-sponsored trials.

How do you explain the difference?
There could be multiple explanations. It could represent bias in the reporting or in the interpretation of results. Or it could be that these studies are in some way superior. Certainly for-profit companies are under a lot of financial pressure. Maybe that causes them to make smarter or safer choices about what drugs to bring to trial. We need larger and more detailed studies to figure out why there is an association between pharmaceutical involvement and positive results, but we should at least consider the possibility that this difference is valid.

That’s a generous interpretation.
To use an analogy, if a baseball player keeps hitting more home runs, it may be because he’s using steroids, or it may be that he’s a better baseball player.

I’d be more likely to bet on the steroids.
I’m as worried about bias as the next person. But these collaborations have led to some real advances in cancer treatment, such as the targeted drugs Herceptin for HER-2-positive breast cancer and Gleevec for chronic myelogenous leukemia. Some of the connection between industry and positive results may be because industry focuses on drug development and they do it well, in which case it would be a shame to bash them for it. My patients definitely need more and better therapies, not fewer.

I’ve spoken to cardiologists who say that some of the trials for stents, for example, cherry-picked patients-enrolling only those with simple lesions and no complicating factors like diabetes. The studies therefore didn’t address the needs of “real-world” patients. And the studies also were fairly short in duration. They picked short-term endpoints like arterial narrowing rather than the more important issues of whether stents actually decrease heart attacks and death.
In oncology trials, you don’t see as many problems with inappropriate endpoints. Most studies have what we call hard endpoints, meaning they measure overall survival or the length of time before the disease progresses or the chances of shrinking the tumor. Intermediate endpoints are less common. In cancer research, it may be harder to hide the fact that a drug either works or doesn’t work.

And you don’t get cherry-picking of patients?
Selection bias in clinical trials is well documented, but that’s an issue in both pharmaceutical- and non-pharmaceutical-funded studies. The patients selected for trials are often healthier than others with the same disease, because it’s safer to test new drugs in people without other medical problems. It’s also easier to interpret the results if there aren’t complicating factors that could influence the outcome.

On the other hand, this can lead to a false sense that a drug is effective. Selection bias was a problem in the early, uncontrolled trials of high-dose chemotherapy with bone marrow transplant for breast cancer, which proved to be no more effective than standard chemotherapy once it was studied in larger, randomized trials. It is very important to make sure you see the same effect in later randomized trials and in more complicated cases outside of trials.

Does this argue for greater monitoring of patients after drugs are approved?
Absolutely. We need improved reporting and monitoring of what happens after new drugs are in use in the broader population.

You note that a higher percentage of drug-company studies are one-arm studies rather than randomized trials. In other words, they had no control group.
True, but one-arm studies are also an essential part of drug development. In 2005, we presented data at the American Society of Clinical Oncology meeting looking at the subset of 2003 trials that were single-arm studies for metastatic breast cancer. If anything, patients were older and sicker in the pharmaceutical-backed studies. Ours was a very small exploratory subset analysis and needs to be confirmed by other researchers in larger studies.

The other issue I’ve heard is that when studies turn up negative results, they tend to be buried. They don’t appear in journals.
That’s true. But there is now a move to establish clinical-trial registries online. You will be able to see what trials are being done, then check if they’re stopped early or not published. That will help answer part of the question.

You are less cynical than I expected.
Prior research has documented the connection between positive outcomes and drug-company funding. But it’s a leap to conclude that it’s due to bias. That still has to be tested and proven. Equally important, we need to identify issues of patient care that are beyond the scope of drug development and which pharmaceutical funding may not support. We need to make sure these issues are being addressed by other funding sources. It is possible, perhaps even likely, that the pharmaceutical industry will do some things very well, like bring new drugs to market, but it will be up to academic researchers to make sure we understand how best to use them.