Is health care quality threatened by manufacturer-funded drug research?

Health care researchers from the Drug Commission of the German Medical Association have concluded that published pharmaceutical research sponsored by drug makers overwhelmingly report positive findings. More so than studies not funded by manufacturers. That's probably no surprise to anyone. For a number of reasons there's a general publication bias towards studies that report positive findings. For example, journals would much rather publish studies showing what works rather than what doesn't work. This new study adds to research suggesting that the bias goes further than that, and could, potentially, threaten some of the basic fundamentals that support quality health care.

Evidenced-based medicine is held up by many as the ultimate goal towards which health care practitioners should strive. That's the idea that, wherever possible, you practice medicine that is backed by reliable published evidence. But what if the published evidence isn't reliable? That's the implication of the German findings, as well as previously published supporting literature. Not only are studies with positive results more likely to be published, but the research protocols, and the interpretations of the results also tend to be more favorable when the funder is also the manufacturer of the product being evaluated. For example, manufacturer-funded research is much more likely to compare new treatments to placebo. While this might seem like sound science, the "randomized, placebo-controlled trial" is, after all, the gold standard. But what placebo control measures is how an active treatment compares to no treatment. Many researchers recommend comparing new treatments to existing treatments, so that we see how much better a new product is compared to existing options, rather than how much better it is than nothing. The interpretation of the results is also a place where findings can be polished to seem more favorable. Other tricks include reporting relative risks rather than absolute risk (an improvement from 1% to 2% is a whopping 100% relative increase, but only a 1% absolute increase.)

Good scientific protocol is supposed to overcome health care research biases inherent in commercial funding in part by including conflict of interest statements in the published studies. But does that imply that a declaration of a potential conflict magically makes the conflict disappear? Or whether it means we should discount the findings all together? What, concretely, does a financial disclosure say about the quality and validity of research? Until we know, it may simply be "reader-beware" with industry funded research. We just don't fully understand the impact it has on research validity, but its pretty clear that it does, at least in aggregate analyses, have a meaningful negative impact on the quality of the science that's supposed to drive improvements in health care.

Photo credit: The author