Friday, July 24, 2009

Congress puts off action on health care reform...

Since Congress isn't going to pass a health care reform bill any time soon, this looks like a good time to take some vacation. I will resume blogging the week of August 3.

Tuesday, July 21, 2009

Is scientific fraud more common than we'd like to think?

In response to my blog on Frances Collins last week, a reader pointed out that the kind of data image manipulation that Collins belatedly discovered in his own lab is more common than many scientists would like to admit: as much as 25 percent of scientific papers submitted for publication contain such image manipulation.

For example, in manuscripts accepted by the American Journal of Respiratory Cell and Critical Care Medicine, the American Thoracic Society found that “approximately 23 percent of images had undergone some alteration “including ‘erasure,’ ‘filling in,’ ‘splicing and ‘cloning.’” In addition, the journals of Cell Biology and Blood have both reported that 20-28 percent of accepted manuscripts had signs of image manipulation, according to The Scientist.

Yet, according to these sources, the Office of Research Integrity at the NIH, which Obama has appointed Collins to direct, only investigates one in 100 cases of suspected data manipulation and thus far only 17 of those investigations have been resolved. Sounds like Collins has some cleaning up to do, should he be confirmed for the post.

On another note, it was good to see The Washington Post enlarging upon some of the reporting I and other bloggers have done on the health care industry's generous contributions to senators like Max Baucus and Chuck Grassley, who are playing a pivotal role in the current health care reform effort.

It was also nice to see The Boston Globe weigh on the same problem, since it was a Globe reporter who first revealed that the biotech Amgen had given $1 million to the planned Edward Kennedy Institute at the University of Massachusetts in Boston. Kennedy aides, of course, say there is no connection between Amgen's healthy donation and the fact that Kennedy's health committee recently voted out a bill that would give biotech and pharm companies a 12-year patent monopoly on new biologic drugs before the drugs become generic. But one can't help but wonder. What's obvious is who the losers will be if this bill passes: American consumers who have to pay top price for new drugs when much cheaper generics should and could be available.

Monday, July 13, 2009

Has Francis Collins, Obama's pick as NIH director, learned his lesson?

When the news broke last week that President Obama had selected Francis Collins to be the new director of the National Institutes of Health, I didn't immediately think of this noted geneticist in connection with a case study in ethics that a visiting scientist had discussed with my health and science journalism class this past spring. It wasn't until I read The New York Times article on Collins' appointment that I remembered the scientist's astute analysis and decided to share it with a wider audience.

The topic of my class this particular afternoon was ethics in science, and in the context of discussing several case studies, my guest speaker, a distinguished geneticist in his own right, brought up an incident that had happened in Collins' lab in the mid-90s when he was director of the National Center for Human Genome Research at the NIH. In August 1996, Collins was notified by the journal Oncogene that a reviewer who was reading an unpublished manuscript about leukemia genes submitted by Collins and a graduate student in his lab at the University of Michigan had discovered data that looked like a "cut and paste job," raising questions of intentional deception. Collins started checking into these allegations and discovered an extensive body of systematically fabricated data by this one graduate student of his. He confronted the suspected perpetrator, who confessed, and wrote a letter to about 100 colleagues whose own work might be affected by the falsified data. In the end, two published papers that contained fake data were retracted in toto and three others corrected. The scandal was covered by a number of major news outlets, including The New York Times.

In the NYT article, Lawrence Altman noted that "Dr. Collins rejected the idea that closer supervision would have prevented the problem because the trainee 'got quite a bit of attention from me' and with others in the laboratory." Interestingly enough, that was precisely the point around which my guest speaker wove his ethical case study. Collins' name was on all the papers that were found to contain falsified data, the guest speaker noted. He then asked my students, "Can a person with Collins' enormous management responsibilities (he was not only overseeing the $189 million human genome project at the time but still operating his lab at the University of Michigan) pay adequate attention to the daily operation of his lab, ensure the quality of the data produced, and guarantee that the papers published under his name are valid?" The conclusion we reached together was no. Collins was juggling too many responsibilities and should not have put his name on these scientific papers because he didn't really have the time to scrutinize the data. What Collins was guilty of, the visiting scientist said, was both a lack of judgment and a conflict of effort.

He then asked an even more thought-provoking question: what does this incident say about the contention that science polices and corrects its own errors through peer review? If this and more recent cases of extensive scientific fraud (I'm thinking here of the Massachusetts anesthesiologist case I wrote about in my blog and the former Army scientist and Medtronics consultant currently in the news) are any indication, the answer is: maybe. Eventually, most cases of extensive fraud may be exposed but only after a great deal of harm has been done to patients and the credibility of the medical research process. Our current system of peer review is far from foolproof in ferreting out fraud, especially when we have star scientists, like Collins, who are too busy trying to burnish their reputations to pay much attention to what is going on beneath their noses.

I can only hope that Collins, if he becomes the next director of the NIH, has taken to heart the lessons from his own lab.

Wednesday, July 8, 2009

Conflicts galore: Statin study authors on the take from statin makers

Just yesterday, Senators Charles Grassley and Herbert Kohl announced that language requiring the public disclosure of payments to doctors from drug and medical device companies has been included in the draft bills for health care reform now being debated in the House and Senate. In case you haven't been convinced by the parade of doctors on the take (who in some cases were found to have skewed scientific data as a result of their conflicts of interest), here's one more reason why public disclosure of such conflicts is so important.

On June 30, the British Medical Journal published a meta-analysis purporting to show that cholesterol-lowering statins reduce the risk of heart attacks and improve survival rates in people without established heart disease. Up till now, evidence that statins like Lipitor, Crestor, Zocor and Pravachol prevent heart attacks in people without known heart risks remains ambiguous at best. And there is little evidence that statins reduce the risk of heart attacks in women, since most of the clinical trials have been in men. Now comes a meta-analysis that examines 10 previous clinical trials and concludes that lo and behold, statins do improve survival rates for men and women and other groups not previously shown to be helped by these drugs.

But before you break out the champagne, please consider these caveats. First off, as Gary Schwitzer pointed out in his health news blog, not everyone is convinced this meta-analysis shows what its authors claim it does. As one online commentator noted, if you look carefully at the data in the study, there is no evidence that the benefits of statins are significant for all 4 end points listed in women, people over the age of 65 and patients with diabetes. And as other commentators noted, the study also doesn't pay much attention to the serious side effects of statins or the fact that there are safer ways to reduce cardiovascular risk, such as exercise and dietary changes.

But what bothers me most about this meta-analysis are the substantial conflicts of interest among six of its 12 authors, most of whom are doctors. According to the fine print at the end of the BMJ study itself, six of the authors are consultants for one or more of the drug companies that make statins, which, by the way, are among the top five classes of drugs sold in the US today. For example, two of the authors are consultants for Merck, which makes Zocor; another researcher, a professor at Cornell University, consults for Bristol-Myers Squibb, which makes Pravachol, another statin; a third has received speaking fees from Pfizer, which makes Lipitor, and a fourth researcher has consulted for Astra-Zeneca, which makes Crestor, as well as Merck. One of the authors (who like six of his colleagues hails from the Netherlands) also sits on the board of two biotech-pharm companies (Aegerion and Arisaph), which are developing cholesterol-lowering drugs. Still another co-author (the one who consults for both Merck and Astra-Zeneca) has developed cardiovascular biomarkers that have been licensed to Siemens and Astra-Zeneca.

In sum, what we have here is a study rife with conflicts of interest. But what we don't know from the fine print is how much these fine fellows are making from their ties to the statin industry; is it $10,000 a year, $100,000 or a $1 million? As I've reported in previous blogs and Side Effects, even a little money can affect the outcome of clinical research and a lot of money can affect it a lot.

That's why we so badly need public disclosure of the financial ties between specific health professionals and health product companies. So consumers can see for themselves just how conflicted the researchers of a particular study are and decide for themselves how much weight they want to give the study's results. In the absence of such key information, I'd have to agree with one of the online commentators on the BMJ study: that the jury is still out when it comes to the benefit of statins in reducing the risk of heart attacks in folks without known heart problems.