Paramedics turn to expired drugs due to shortages

[Carnildo]

Aerospace is also a highly regulated industry, for much the same reasons. Yet production shut downs by the regulatory agencies due to safety issues is not a common occurance. Is the FDA (and co.) that much stricter, or is there something inherent in producing drugs that make quality control that much harder than say producing titanium fan blades (which have to have a tightly controlled alloy composition and micro-structure, as well as have very tight dimensional tolerances and an internal system of hollows for cooling air)?

In general, defective parts coming off an aerospace production line are due to random variations in the manufacturing process (in which case you just discard the bad parts), and when they aren't, it's generally due to a failure of a machine (which tends to be obvious). The only time you'd shut down an assembly line for more than a day or two is if you'd identified a fundamental flaw in the manufacturing process, which can be (and usually is) prevented during the design process. I can't think of anything equivalent to a bacterial infection of a sterile production line, where a random event requires you to shut down an assembly line for an extended period.

[Vaniver]

A company wants me, quite possibly a dieing person ready to cling to any possible straw that might keep me alive, to fork over lots of money based on the claim that their drug will make me better. How is requiring proof of efficacy then a "tremendously harmful" practice?

So, let's talk about heart attacks and aspirin. Taking aspirin during a heart attack significantly decreases the chances of death, and regularly taking low-dose aspirin can lower the risk of heart attack in the first place. It took the FDA several years to permit drug companies to publicly advertise what they knew to be true from scientific studies- with how many preventable deaths in the meantime? If they only required safety, well, aspirin's been in regular use for over a century, and so anyone could make any non-fraudulent claims they want about it, as soon as they're confident enough to make those statements.

You may recall the cease-and-desist letter sent to General Mills, in which the company was charged with making "unauthorized" claims. Not untrue claims, just claims that the FDA had not approved. That's what it looks like when you require submitting proof of efficacy to the FDA- you can't even say things that both you and the FDA know to be true unless you jump through hoops first. That doesn't strike me as an effective public health measure.

If someone is in desperate straits, and a new and unproven drug might help them out? Why would I want to take that hope from them? The only excuse I can think of is a randomized controlled trial to obtain the scientific knowledge to help others- doing it for the safety of the dying seems like extremely misplaced priorities.

I don't think I want to go back.

Any reason in particular? The pre-1962 FDA was already effective enough to prevent Thalidomide from deforming more than a handful of Americans. The trouble with any regulatory agency is that they face a tradeoff between false positives and false negatives- the damage done by not allowing good drugs is just as real as the damage done by allowing bad drugs, though it's a bit harder to notice. But the FDA is only punished for allowing bad drugs, and not rewarded for allowing good drugs. There were recent reforms to try and realign their incentives a little more in line with American health, but it's clear they're still way too cautious of an organization, and thousands of people die prematurely a year as a result.

[johnny_7713]

Aerospace is also a highly regulated industry, for much the same reasons. Yet production shut downs by the regulatory agencies due to safety issues is not a common occurance. Is the FDA (and co.) that much stricter, or is there something inherent in producing drugs that make quality control that much harder than say producing titanium fan blades (which have to have a tightly controlled alloy composition and micro-structure, as well as have very tight dimensional tolerances and an internal system of hollows for cooling air)?

In general, defective parts coming off an aerospace production line are due to random variations in the manufacturing process (in which case you just discard the bad parts), and when they aren't, it's generally due to a failure of a machine (which tends to be obvious). The only time you'd shut down an assembly line for more than a day or two is if you'd identified a fundamental flaw in the manufacturing process, which can be (and usually is) prevented during the design process. I can't think of anything equivalent to a bacterial infection of a sterile production line, where a random event requires you to shut down an assembly line for an extended period.

Ah, that makes sense. Thanks for the explanation.

[morriswalters]

Any reason in particular? The pre-1962 FDA was already effective enough to prevent Thalidomide from deforming more than a handful of Americans. The trouble with any regulatory agency is that they face a tradeoff between false positives and false negatives- the damage done by not allowing good drugs is just as real as the damage done by allowing bad drugs, though it's a bit harder to notice. But the FDA is only punished for allowing bad drugs, and not rewarded for allowing good drugs. There were recent reforms to try and realign their incentives a little more in line with American health, but it's clear they're still way too cautious of an organization, and thousands of people die prematurely a year as a result.

The bolded section is specious. I don't want the FDA rewarded for anything. Why should I want that? That's a incentive to approve, I don't want any incentives. I want the FDA to act as trusted broker of information. Pressure enough to approve exists now.

The approval rate for drugs will always be too slow. It's a complex problem and by it's nature it has to be slow. And drugs companies routinely show that they will manipulate the process both during and after for their benefit. They lie by omission, encourage off label uses, and use direct advertising to affect the judgement of the population least likely to have the tools to evaluate a particular drug. This.

As to Cheerios, so sorry about their luck. What do you suggest they should be allowed to do. If safety is the only concern why not call it a cure for cancer? It's about trust. Can the claims they make be trusted to mean what they say they mean?

[The Great Hippo]

The bolded section is specious. I don't want the FDA rewarded for anything. Why should I want that? That's a incentive to approve, I don't want any incentives. I want the FDA to act as trusted broker of information. Pressure enough to approve exists now.

Organizations work best when they have an incentive to work as they should.

As to Cheerios, so sorry about their luck. What do you suggest they should be allowed to do. If safety is the only concern why not call it a cure for cancer? It's about trust. Can the claims they make be trusted to mean what they say they mean?

Because 'Cheerios can cure cancer' would be a lie. 'Cheerios can reduce your cholesterol by 4% over six weeks' assumedly isn't.

Also, trust is merely a tool by which safety is better guaranteed. Safety is priority one; all this other stuff are just tools by which you increase safety. If the result of decision X is a higher death rate, decision X has decreased the safety of the public.

[EMTP]

If they only required safety, well, aspirin's been in regular use for over a century

Old does not mean safe. If you put a million people on the lowest possible dose of aspirin, between ten and twenty thousand or so will be diagnosed with gastrointestinal bleeding. It took a lot of population-based studies to find the trade-off between bleeding risk and MI/stroke prevention. We argued about the correct dose for a long time -- everything between full dose aspirin (325mg) to a baby aspirin (81mg). We found that really high-risk patients, like those with artificial heart valves or atrial fibrillation, need even strong anticoagulation causing even more bleeding (coumadin rather than aspirin).

These decisions aren't easy, and I'm not sure we want to leave them exclusively in the hands of corporation boards, marketing executives, and the public.

I'd be curious:

* When you think the science of aspirin for MI/CVA prevention was settled, as regards the safe dose, relative risk of complications, etc.
* How long after that drug companies were able to argue that.

As to terminal patients, I agree. They should be able to make an informed choice about risky experimental treatments. We also need to streamline the process for all medical trials; only 4% of doctors participate in clinical trials, and it's because the process of institutional review is so incredibly painful.

[morriswalters]

[Because 'Cheerios can cure cancer' would be a lie. 'Cheerios can reduce your cholesterol by 4% over six weeks' assumedly isn't.

Also, trust is merely a tool by which safety is better guaranteed. Safety is priority one; all this other stuff are just tools by which you increase safety. If the result of decision X is a higher death rate, decision X has decreased the safety of the public.

Granted. But if the claim was based on only the harm done by the cereal, and since, per se, there is none, and any claim would be valid. General Mills wants to tap dance on the line and push their claim as far as they can. The FDA pushes back to make sure the claims meets the facts. Which was achieved by a rewording of the claim. This is the process, but the process can't be perfect.

Trust in terms of the FDA is my overriding concern, since I don't have access to the data that makes a discussion about the relative safety to the public meaningful. I have to operate in an zone where, due to the complexity of the issues, I can't make any judgement of the claims. I have to trust the source. To do otherwise is to show the Dunning–Kruger effect. This is why I don't like incentives to act, but disincentives for acting incorrectly. I want to force caution.

Show me a reliable metric, which tells you in advance of any given decision, what the comparative death rates will be, given a particular decision? Hindsight is 20/20. The thalidomide incident speaks to this, it was only a matter of luck that the drug wasn't approved here. In any case it didn't kill. And the deaths due to a drug aren't just about base toxicity, but about statistical comparisons in cases where the drug is merely one factor among many. Safety is not just about death rates.

[The Great Hippo]

Show me a reliable metric, which tells you in advance of any given decision, what the comparative death rates will be, given a particular decision?

"How many deaths can be directly linked to a lack of drug X? Okay, now what are our high-end estimates for deaths as a result of allowing for drug X?"

If figure 1 is ridiculously higher than figure 2, that implies that we should allow for drug X. The problem being described here (and I don't know if it's true, but let's be precise about what's being said) is that the FDA is obsessed with ensuring that drugs are always safe and effective, even when the metrics they use could potentially be counter to their highest goal--protecting the public health.

It's probably unfair of me to speculate toward reasons why, but assuming this was true, I would wager a guess that it has to do with how we parse fault. If the FDA doesn't allow for drug X, and you die as a result of a disease that drug X treats, the FDA can claim they were just looking after your safety (the FDA doesn't treat disease, they just regulate drugs. Your disease wasn't their department). But if the FDA does allow for drug X, and drug X kills you, the FDA can't really make that claim anymore (even if drug X ultimately saved more lives than it took), because making sure drug X doesn't kill you is their department.

I'm probably over-generalizing, but keep in mind--I'm not arguing for changes to be made to the FDA. I don't know anything about the FDA, just that it makes decisions that I find really weird now and then. I'm just explaining what the position seems to be.

[morriswalters]

Then I believe that we are substantially in agreement. The FDA authorizing statute is constantly amended. There is even an amendment related to the use of unapproved drugs or therapies.

[Carnildo]

Show me a reliable metric, which tells you in advance of any given decision, what the comparative death rates will be, given a particular decision?

"How many deaths can be directly linked to a lack of drug X? Okay, now what are our high-end estimates for deaths as a result of allowing for drug X?"

If figure 1 is ridiculously higher than figure 2, that implies that we should allow for drug X. The problem being described here (and I don't know if it's true, but let's be precise about what's being said) is that the FDA is obsessed with ensuring that drugs are always safe and effective, even when the metrics they use could potentially be counter to their highest goal--protecting the public health.

The problem is that neither figure is easy to calculate. How do you determine "deaths linked to a lack of drug X", when drug X doesn't cure terminal condition Y, but rather delays death by a few years? How do you make a reasonable high-end estimate for deaths caused by the drug when those dying are teenagers whose mothers were pregnant when they took the drug? What about when the deaths are an indirect effect of the drug (eg. a drug that causes paralysis as a side effect, which in turn makes the person more likely to die prematurely)? How do you deal with a drug that causes near-miraculous cures in 90% of the treated population, and kills the other 10%?

[Vaniver]

The approval rate for drugs will always be too slow.

I'm confused by this statement. Do you mean that it's impossible for drugs to be approved too quickly? Or that there will always be people who want the approval rate to be faster than it currently is? Or something else?

Old does not mean safe.

Agreed! But my current impression is that the FDA does little to no safety-related testing of things that have been consumed for long periods, like salt, sugar, or wheat (all of which have credible health risks). Grandfathering things in like that is not optimizing public health, and I'm troubled by health-related regulatory agencies whose incentives are not lined up with public health. (For example, my impression is that several sugar replacements are healthier than sugar, but are required to have warnings because they're not zero risk, which is a misuse of the FDA's ability to warn.)

And we haven't even started talking about alcohol or tobacco!

These decisions aren't easy, and I'm not sure we want to leave them exclusively in the hands of corporation boards, marketing executives, and the public.

Exclusively? Of course not. I imagine most people would like to hire a medical expert to make that sort of decision for them. I also agree with the implication that private interest announcements and public interest announcements are often difficult to separate, especially when medical commerce is involved.

Most importantly, I agree that the decisions are not easy. We should expect most technologies to be tradeoffs- altering your blood makes it both less likely to have heart attacks and more likely to have intestinal bleeding- but most people naturally want to think about technologies as good or bad. Either nuclear power is cheap and clean- and not risky- or it's risky, and also expensive and dirty. From the outside, it seems possible that it could be cheap, clean, and risky, or expensive, dirty, and not risky, but very few people hold that impression.

The FDA seems like it plays into that fallacy- either drugs are safe and effective, or unsafe and ineffective. Drugs that are unsafe but effective, or safe but ineffective don't seem to have a place. (Why would you want to approve either of those? The first kind may be better than the next best option for some patients, and the second kind of drug may be effective for something, but what or for who is not known yet.)

And so free flow of information seems like it could be a significant solution, especially if that came with a clear separation of risks and rewards. Requiring transparency seems more flexible than requiring veracity (beyond basic anti-fraud requirements).

I'd be curious:

I'm not an expert in that field, and so can't defend any of my examples very far.

[morriswalters]

Or that there will always be people who want the approval rate to be faster than it currently is?

Close. If your the one who needs the drug than any delay will seem to long.