« WoW, glider, and the difficulty of attestation | Main | I gotta get me one of those suits »
March 29, 2008
Making sense of the heparin incident
This NYT article talks about the problem of assuring the quality of food and medical products sourced outside the US (though all the incidents here actually are of products from China).When cold medicine containing a poison made in China killed nearly 120 Panamanians in 2006 and early 2007, Americans could take some comfort in the belief that a similar epidemic could never happen here, not with one of the best drug regulatory systems in the world.The article also comes with an extremely scary photo of one of the small workshops in China where the the pig intestines are processed. That doesn't appear to be the problem, though. Based on the article and the Wikipedia article on heparin, the contamination doesn't seem to have been a manufacturing problem but rather an intentional adulterant, as, it seems, were the toothpaste and pet food incidents. That's a different story entirely and it seems a lot less likely that just jacking up the inspection rate or having onsite inspectors is going to be very effective, for several reasons:Then last spring, hundreds if not thousands of pets died or were sickened in the United States by a Chinese pet food ingredient that contained lethal levels of melamine, an industrial product used to artificially boost protein levels. That was followed quickly by the discovery that Americans were brushing their teeth with Chinese toothpaste containing a poisonous chemical used in antifreeze.
Still, no Americans died from the chemical.
And then came heparin.
A hugely popular blood thinner used in surgery and dialysis, heparin turned out in some cases to contain a mystery substance that sophisticated lab tests earlier this month determined to be a chemically modified substance that mimics the real drug. The United States Food and Drug Administration has linked it to 19 deaths and hundreds of severe allergic reactions, though the agency is still investigating whether the contaminant was the actual cause.
...
Congressional Democrats are talking about authorizing more money so the F.D.A. can do more overseas inspections, particularly in China, where more and more drug ingredients are made. The agency is also completing a plan to permanently station employees in China for the first time.
- You only inspect a small fraction of the product and processes (e.g., you come by one day a month), so it's a lot better at detecting systemic carelessness than cases where the inspected party is deliberately trying to defraud you, which seems to be what is happening here.
- Because inspection regimes only catch a small percentage of the offenses, they need to be coupled with a pretty aggressive punishment regime. Without that, you just stop the particular incident, but what you want is to deter all incidents.
- One of the people quoted in the article, Dr. Roger L. Williams from USP, suggests that we could have a better testing procedure. Again, this is something that works best when you're looking for sloppy processes, since those follow a somewhat predictable pattern. When you're dealing with active attack, we should expect the attackers to simply adapt their methods.
Fundamentally, the entire food and drug system is based on trust-but-verify. If we're dealing with suppliers which can't be trusted at all, we need to either get a different attitude or deal with a different set of suppliers.
Posted by ekr at March 29, 2008 9:08 PM | Filed under: Pharma