A pair of recent articles regarding the performance of the Food and Drug Administration recently came to our attention. The first was a New York Times editorial titled Not What We Call Due Diligence, berating the agency for the way conflicts of interest were identified and dealt with in the approval of medical devices. The second, also from the New York Times, was an article out of Washington City titled Report Criticizes F.D.A. on Device Testing, relating to devices getting approved too easily in testing.
In the first piece, the Times rightly excoriates the FDA for not adequately collecting and auditing information about the financial relationships between the manufacturers and the physicians that contract to do their testing, based on a report from the DSHS inspector general. One bizarre item was that the inspector general suggested that the FDA gather this information before trials began, which the agency opposed. The Times properly considers this to be backwards, far better to disqualify a contractor based on conflicts before the trial rather than attempt to evaluate a report afterwards when a conflict is revealed — which is probably impossible.
The second article covers a Government Accountability Office report dealing with testing of medical equipment. Apparently, a manufacturer of a pace maker, for example, having been through rigorous testing for one model can bring in the next one and get expedited clearance with very little testing based on the similarity. The story quotes Dr Peter B. Bach of Memorial Sloan-Kettering Cancer Center as saying: “So on the one hand, the manufacturer wheels in their new Ferrari to the FDA and says, ‘Look, it’s a car just like the Model T,’ Then they go out in the marketplace and say to doctors, ‘Why would you drive anything but a Ferrari?’ This drives up the cost of care without any necessary actual improvement in outcomes.”
While both articles raise valid questions about how the FDA does it’s job, both miss the point that the FDA is doing a job that shouldn’t be done by government at all. It makes a certain amount of sense for a government agency to run a new drug or device through a series of trials to make sure it is not generally dangerous, but to put a government stamp of approval on a product, endorsing its efficacy in a certain arena, does not. That information can only rarely be derived from limited tests. The most famous case has to be thalidomide, a powerful drug with good results in treating a range of conditions. Unfortunately, it can cause severe birth defects when taken by pregnant women. When widespread use demonstrated that the drug in its original form could not safely be used during pregnancy, a great hue and cry arose and the drug was pulled off the market around the world in 1961.
In 1964 an Israeli physician found that thalidomide was helpful in treating certain complications of leprosy, so the lightning-fast FDA approved its use in that area a third of a century later (1998). In 2006 the FDA approved the drug for use in treatment of multiple myeloma, and the drug is successfully being used in treatments for AIDS, various cancers, certain skin diseases, and macular degeneration. But for thirty years, the benefits of this powerful drug were unavailable to patients.
What we can learn from this is that clinical trials are not a reliable way to identify problems in specific cases, but that intelligent analysis of results in the general population are. We need to have the FDA take a quick look to make sure that a new drug or product isn’t widely dangerous, then get the hell out of the way. We further need to dramatically accelerate the adoption of computerized medical records so that investigators (and by this we mean scientists, not bureaucrats) can more quickly learn the conditions under which a drug should not be used and the relative benefits of different treatments for different situations. The current system tends to suggest that whichever product the FDA most recently approved for a specific condition must be the most efficacious, despite the fact that even the FDA makes no such claims.
If the FDA has a role, the best use of their energies will be to ensure a transparent marketplace in information, so intelligent patients and dedicated health-care professionals will know what analyses were written by doctors who were beholden to the drug companies, and thus make informed decisions. Federal gatekeepers are not the solution, they’re a large part of the problem. If I Were King, the FDA would be much smaller, and dedicated to a beneficial mission that it could perform instead of a dubious mission that it cannot.