4 ms·

>> And the FDA has quite likely killed more people than it has saved, > My immediate reaction is to say you're full of shit. But I'll hold off judgment and see

by glenra 14y ago

>> And the FDA has quite likely killed more people than it has saved, > My immediate reaction is to say you're full of shit. But I'll hold off judgment and see if you can offer any shred of an argument that justifies this statement (not even asking for proof here). The result of the FDA's post-1962 efforts has been to dramatically slow the rate of introduction of new drugs with essentially no effect on the average quality level of the drugs that do get introduced. Some new drugs still get through, but less than half as many per unit time and the FDA approval process delays the use of those drugs and makes them vastly more expensive to develop. These extra delays and higher costs are estimated to have caused, at a minimum, hundreds of thousands of preventable deaths. Good drugs save lives. Good drug information saves lives. The FDA often finds itself in the business of postponing the use of good drugs and preventing the dissemination of valid drug information. It has a powerful institutional bias towards inaction, but failing to approve good drugs in a timely manner kills people with at least as much certainly as does letting bad ones through. The notion that the modern FDA on net hurts us more than it helps is a pretty common belief among those economists who have studied the matter. I'd start with this general overview of the issues: http://www.econlib.org/library/Enc/PharmaceuticalsEconomicsandRegulation.html http://www.econlib.org/library/Enc/PharmaceuticalsEconomicsa... And then browse this site - lots of good stuff there: http://www.fdareview.org/ http://www.fdareview.org/ Bryan Caplan summarizes the general state of opinion as: "The public thinks the FDA is great. Regular economists think it's pretty good. And economists who specialize in the FDA think it's pretty bad." (source: http://econlog.econlib.org/archives/2005/04/fda_public_econ.html http://econlog.econlib.org/archives/2005/04/fda_public_econ.... )

hackinthebochs 14y ago
Well, you certainly passed my low bar for rationale. I still can't agree though. Any argument based on statistics is going to be strongly biased against the FDA's current procedure. It is impossible to predict how many deaths were prevented from keeping dangerous drugs off the market, so it is impossible to make a meaningful comparison. Furthermore, a simple utilitarian argument about the net lives saved is not how most people would comprehend the situation. The situation where a drug kills someone will be weighed much more gravely than the situation where a life-saving drug was kept from you because it couldn't be proven effective. So a net-lives-saved calculation simply doesn't capture the intuition that most people have. Being biased towards conservatism is appropriate at the scale of approving drugs and advertisement towards an entire country. Most people do not have the knowledge (even considering the internet) to appropriately evaluate a drug. I would wager that even doctors are not equipped to make this sort of safety evaluation for their patients. This is exactly the sort of thing that scientists taking a conservative approach should be doing.
- glenra 14y ago
  Read the links I gave. I wasn't expecting you to be convinced just by what I wrote above. Before 1962, the FDA was tasked to verify safety but NOT to verify efficacy. Everything you just wrote only argues for that - for keeping drugs off the market that "kill someone" when they take it. It does NOT argue for keeping drugs off the market that we know are safe but can't yet conclusively prove work "better than a placebo" or "better than the standard treatment" in a study. After 1962, the FDA was tasked with determining efficacy. That is the part where the main problem lies. It turns out that proving that a drug probably doesn't kill people is MUCH easier than proving that a drug "works" at a level better than some metric. To prove a drug is safe, we have models and processes to calculate the LD50 and verify that people who take it at expected dose levels don't get sick. But to prove it's effective is hard. Hard enough that, in practice, insisting on it it kills hundreds of thousands of people. As one popular example, when the FDA approved beta blockers for use in preventing heart attacks, the FDA's own press release claimed that the newly approved drugs would save 10,000 lives per year, based on the results seen in Europe where the drugs had been in use for the prior decade. Economists look at that and notice that seems to suggest keeping that one class of drug OFF the market in the US after it had been approved in Europe killed on the order of 100,000 people all told. That is a very large death count. The worst death count we've ever seen due to drugs that "slipped through the cracks" prior to 1962 is a few orders of magnitude smaller than that. > The situation where a drug kills someone will be weighed much more gravely than the situation where a life-saving drug was kept from you because it couldn't be proven effective. So a net-lives-saved calculation simply doesn't capture the intuition that most people have. That is true, but in this case the "intuition that most people have" is WRONG. By any rational standard, killing a hundred people by preventing them from getting their medicine is MUCH WORSE than killing one person by letting him get his medicine...even if we know who the people affected are in the latter case and not the former. That is approximately the sort of tradeoff we're looking at. > It is impossible to predict how many deaths were prevented from keeping dangerous drugs off the market, so it is impossible to make a meaningful comparison. It is difficult, but probably not impossible. One might look at the death rates attributed to medicinal prescription drugs under various regulatory regimes. In the US, you can compare drugs approved pre-1962 versus post-1962 - if the FDA's extra testing in the modern era were really valuable, there should be a bend in that curve in the correct direction (hint: there isn't). One can also compare with other countries that changed their regulatory regimes to be stricter or looser. Sure, you can't get an exact number of deaths, but you can make a plausible estimate. And this calculation WAY favors the FDA approach, due to the "what is seen and what is not seen" factor - people who die from taking a drug show up in newspaper headlines blaming the drug whereas people who die from NOT taking a drug do not show up in newspaper headlines blaming the FDA. :-) BTW, one simple huge improvement would be to just say that any drug approved for use by some other country's regulatory apparatus could be used here too with a mere "not yet approved by the FDA" warning label.

by glenra 14y ago

hackinthebochs 14y ago

Well, you certainly passed my low bar for rationale. I still can't agree though. Any argument based on statistics is going to be strongly biased against the FDA's current procedure. It is impossible to predict how many deaths were prevented from keeping dangerous drugs off the market, so it is impossible to make a meaningful comparison. Furthermore, a simple utilitarian argument about the net lives saved is not how most people would comprehend the situation. The situation where a drug kills someone will be weighed much more gravely than the situation where a life-saving drug was kept from you because it couldn't be proven effective. So a net-lives-saved calculation simply doesn't capture the intuition that most people have. Being biased towards conservatism is appropriate at the scale of approving drugs and advertisement towards an entire country. Most people do not have the knowledge (even considering the internet) to appropriately evaluate a drug. I would wager that even doctors are not equipped to make this sort of safety evaluation for their patients. This is exactly the sort of thing that scientists taking a conservative approach should be doing.

glenra 14y ago
Read the links I gave. I wasn't expecting you to be convinced just by what I wrote above. Before 1962, the FDA was tasked to verify safety but NOT to verify efficacy. Everything you just wrote only argues for that - for keeping drugs off the market that "kill someone" when they take it. It does NOT argue for keeping drugs off the market that we know are safe but can't yet conclusively prove work "better than a placebo" or "better than the standard treatment" in a study. After 1962, the FDA was tasked with determining efficacy. That is the part where the main problem lies. It turns out that proving that a drug probably doesn't kill people is MUCH easier than proving that a drug "works" at a level better than some metric. To prove a drug is safe, we have models and processes to calculate the LD50 and verify that people who take it at expected dose levels don't get sick. But to prove it's effective is hard. Hard enough that, in practice, insisting on it it kills hundreds of thousands of people. As one popular example, when the FDA approved beta blockers for use in preventing heart attacks, the FDA's own press release claimed that the newly approved drugs would save 10,000 lives per year, based on the results seen in Europe where the drugs had been in use for the prior decade. Economists look at that and notice that seems to suggest keeping that one class of drug OFF the market in the US after it had been approved in Europe killed on the order of 100,000 people all told. That is a very large death count. The worst death count we've ever seen due to drugs that "slipped through the cracks" prior to 1962 is a few orders of magnitude smaller than that. > The situation where a drug kills someone will be weighed much more gravely than the situation where a life-saving drug was kept from you because it couldn't be proven effective. So a net-lives-saved calculation simply doesn't capture the intuition that most people have. That is true, but in this case the "intuition that most people have" is WRONG. By any rational standard, killing a hundred people by preventing them from getting their medicine is MUCH WORSE than killing one person by letting him get his medicine...even if we know who the people affected are in the latter case and not the former. That is approximately the sort of tradeoff we're looking at. > It is impossible to predict how many deaths were prevented from keeping dangerous drugs off the market, so it is impossible to make a meaningful comparison. It is difficult, but probably not impossible. One might look at the death rates attributed to medicinal prescription drugs under various regulatory regimes. In the US, you can compare drugs approved pre-1962 versus post-1962 - if the FDA's extra testing in the modern era were really valuable, there should be a bend in that curve in the correct direction (hint: there isn't). One can also compare with other countries that changed their regulatory regimes to be stricter or looser. Sure, you can't get an exact number of deaths, but you can make a plausible estimate. And this calculation WAY favors the FDA approach, due to the "what is seen and what is not seen" factor - people who die from taking a drug show up in newspaper headlines blaming the drug whereas people who die from NOT taking a drug do not show up in newspaper headlines blaming the FDA. :-) BTW, one simple huge improvement would be to just say that any drug approved for use by some other country's regulatory apparatus could be used here too with a mere "not yet approved by the FDA" warning label.