Hacker News Reader

> Even simpler is to prohibit payments for these experimental treatments, but let people who whould die anyway participate in them as long as it is free for them. In return the drug companies could use the outcomes as data points towards their aproval applications.

That would never happen for good reason. Drug companies would have every incentive to bias the results towards approval by giving the drugs to detain patients. I’m not exaggerating when I say this might be the worst idea I’ve read in this entire thread.

What do you mean with "by giving the drugs to detain patients"? Is that a typo? I don't see how companies holding on to terminally-ill patients makes the trial results look more favourable. Are you suggesting that companies might not release the bodies of the patients that died during the trial?

I meant to write “certain”.

> Drug companies would have every incentive to bias the results towards approval by giving the drugs to detain patients.

What do you mean “detain” patients? Remember we are talking about people who have a terminal illness with no known cure. How can drug companies “bias the results”? Obviously I’m not saying we take anything they say as gospel, but in short order these patients are expected to die. If a drug managed to cure them that would be quite hard to falsify don’t you think?

> I’m not exaggerating when I say this might be the worst idea I’ve read in this entire thread.

Well, thank you. I aim to please. Are you sure that you understood the idea well? If you are, could you please explain with more words what is the misbehaviour you are worried the drug companies would commit?

> What do you mean “detain” patients?

I meant to write "certain".

> How can drug companies “bias the results”?

Choosing appropriate patients will lead to longer survival rate unrelated to the drug in question. The whole point of double-blind studies is to avoid this.

> Well, thank you. I aim to please. Are you sure that you understood the idea well? If you are, could you please explain with more words what is the misbehaviour you are worried the drug companies would commit?

I did understood you. You, however, don't seem to understand statistics or economics. Your idea directly incentivizes drug companies to choose patients more likely to make their drugs look better. It is a truly horrible idea.

> Your idea directly incentivizes drug companies to choose patients more likely to make their drugs look better.

Yes. So? The whole patient group in question is destined to die fast. The best we currently can offer them is aleviate their discomfort and help them come to terms with their demise. If the drug company can identify a subgroup who they can possibly save and actually help those that is winning.

> You, however, don't seem to understand statistics or economics.

And you are abrasive and rude. These exclamations don’t add much of anything to the conversation. If there is something particular I am overlooking please adress that instead of speculating about the areas you think I am deficient in. Thank you.

You are aproaching it as if the question we are asking is if this is the best study design. But that is not the question we are asking. There is a patient who we believe is about to die soon. In comes a drug manufacturer and says to the patient: “you do not have to die. I think I have the cure for you”. The patient wants the drug the drug manufacturer is offering. The question is if we are to interject and stop the drug manufacturer giving the patient the drug, and if so under what circumstances.

I am saying: “yes, we should stop the drug manufacturer if the drug is not proven to be effective yet, and he is asking for money for it. There is too much of a chance that the drug manufacturer is a snake oil salesman in disguise and only wants to extract the wealth from the patient and the patient’s family. But if he is honest they can give the drug for free.”

In turn you are saying. “We should stop the drug manufacturer giving the unproven drug to the dying patient even if they are giving it for free, because it is possible that they cherry picked this patient. So though luck dear patient, you must die in the name of scientific rigour. I hope you understand.”

Do I understand you right?

> You are aproaching it as if the question we are asking is if this is the best study design. But that is not the question we are asking. There is a patient who we believe is about to die soon. In comes a drug manufacturer and says to the patient: “you do not have to die. I think I have the cure for you”. The patient wants the drug the drug manufacturer is offering. The question is if we are to interject and stop the drug manufacturer giving the patient the drug, and if so under what circumstances.

This is already largely possible in most states and called "right to try".

> Do I understand you right?

No you seem not to be. Giving away drugs is fine and largely already possible. Choosing patients and then pretending that the data gathered has any reasonable statistical validity when considering general approval of a drug is totally crazy. I think there is essentially zero chance the FDA would ever go for an idea with such overt conflicts of interest so I'm happy your idea will never turn into reality.

> This is already largely possible in most states and called "right to try".

The article’s author is convinced it is not covering them. Do you know otherwise? Perhaps you should let them know. You might save a life.

> Choosing patients and then pretending that the data gathered has any reasonable statistical validity when considering general approval of a drug is totally crazy.

Would like to mention that it is your adition that the drug manufacturer chooses the patients. It was not part of my original comment. And it is not core to the idea.

One can simply imagine a protocol where the manufacturer declares their criteria (what condition, what severity[1]) and how many doses they are able to provide and those “places” get filled up with volunteers. And this of course can be made even more buletproof.

1: and before you point it out, of course if the manufacturer is only willing to test their drug on not-so-bad-but-dying people then they get approval to be used on the same selection.

This plan seems to have excellent incentive alignment and basically mirrors the system in place, but without the need to be approved for clinical trials.

Experimental treatments being available to people with no options (there must be no applicable option with a proven probability of success or this becomes very questionable) could save drug companies millions-billions in the off chance that it proves to be incompatible with human testing or fails to produce the expected effects in humans.

I would take this up a notch or two for patients in a short-range situation that is certainly fatal, opening the door for them to try treatments that have just barely crossed en-vidrio and seem to have acceptable toxicity in mouse models. This would essentially be donating one’s body to science but in a much more useful form, with the side benefit of a slim chance of life extension.

This could benefit humanity by significantly reducing time to market for some drugs and cutting costs on eventual dead-ends.

I don’t think these people would be useful. Pharma needs controlled double blind studies to actually tell if the therapies work.

Of course, for actual efficacy, but you can tell if a treatment has a negative reaction or fails to effect the expected physiology in the expected way with a very small number of samples.

If everyone you give it to dies from reactions, or the expected effects seem to be happening on a physiological level, you don’t know much about efficacy but you do know a lot more about the eventual prospects for an actual clinical trial. It’s basically an extension of en-vidrio testing.

No, that does not set the right incentives. You have completely written off the harm of dangerous or deadly drugs.

The HepC drug Sovaldi brought in around $50B-100B in sales, but probably costs less than $50 to administer. Under your scheme it would make economic sense for a company to kill 1,000,000,000 people testing variants until they chance upon the right one because they incur no costs for administering deadly experiments.

The problem with your scheme is that human lives are valuable. A unsafe drug that causes harm or death is immensely destructive. If we want to be completely reductive, the actuarial value of a human life is around $10,000,000. A unsafe drug could kill the person resulting in a $10M loss to society. Therefore, the company should be required to escrow $10M per person until the treatment is done, paid out proportionally to any potential death or harm in excess of the gold standard treatment. Only then would the incentives be reasonably aligned.

This is obviously a very reductive treatment, but it helps illustrate the problems with unrestricted testing.

> You have completely written off the harm of dangerous or deadly drugs.

I was only talking about testing on people who we know are going to die fast because of their condition, and we do not have a way to save them in any other way. And naturally only with their consent. People who have a ticking clock and therefore willing to try anything.

This is the context the article in question talks about.

I think you forgot to price in the fact that some people live with what is effectively a time limit sometimes counted in months or weeks (like the person from the linked article).

> Even simpler is to prohibit payments for these experimental treatments, but let people who whould die anyway participate in them as long as it is free for them. In return the drug companies could use the outcomes as data points towards their aproval applications.

Hell I'd even approve my tax money to go to those companies which end up saving lives as reward.