The Wall Street Journal reported (subsequently re-published by the Pittsburgh Post-Gazette) about FoxHollow Technologies' questionable strategy for studying the performance of its single product, a catheter system (SilverHawk) used to remove atherosclerotic placque from limb arteries.
The Journal reported that the only research that FoxHollow has sponsored to evaluate their product was a patient registry, i.e., an ongoing case-series. The article reported that FoxHollow "doesn't want to conduct [randomized controlled] trials to test SilverHawk's effectiveness against other procedures because that would be expensive and unnecessary, given the positive registry data."
A quick PubMed search indeed revealed only a few articles that described small case-series of patients on whom the device was used, (for example, see this link). I could find no controlled trials of this device.
Case-series, however, are subject to selection bias, since the patients physicians select to undergo treatment with the intervention of interest may systematically differ from those they choose not to treat in ways that could affect the outcomes measured. Controlled trials that randomly assign patients to treatment by the intervention of interest or a comparison intervention make it likely that the two groups will be similar in all aspects that might affect the outcomes of interest other than the treatment assignment.
Making the situation more murky is that many of the physicians who are involved in the registry have conflicts of interest. Most have received consulting fees from FoxHollow, some of which were provided in the form of stock or stock options, and many own substantial amounts of FoxHollow stock, some of which the company provided them at discounted prices. The article gave the example of Dr. R. Stefan Kiesz, identified as a cardiologist who supplied data to the registry, who owns options to buy 10,000 - 20,000 FoxHollow shares, and has sold options for another 8,000, all of which were granted as payment "for advice he gave to help develop the device."
Thus, FoxHollow's reliance on positive data from its registry ignores the possibility that the registry data was positive due to selection bias, and perhaps also the bias of investigators with a financial stake in the company and its one product.
This is yet another unfortunate variation on theme of conflicts of interest affecting medical research, with both physicians and company leaders bearing responsibility.
Are you a clumper or a splitter? - The Economist reports about an experiment in the National Health Service: *On February 25th it emerged that local authorities, clinicians and George Osborn...
2 hours ago