5 ms·

so... personally? were I to shell out unusual money for experimental treatments, I'd be way more comfortable with a pay-for-performance model. Like if my doct

by lsc 8y ago

so... personally? were I to shell out unusual money for experimental treatments, I'd be way more comfortable with a pay-for-performance model. Like if my doctors think I've got a year left, and you think you can cure me for $120K, sure, that seems like a totally reasonable deal if you can actually cure me (and maybe even good for everyone, regardless of if I live or die, if you then use what you learned from me to help others.) But I'm a computer nerd, not a doctor. I'm not really in a position to tell the difference if you are giving me sugar pills. Seems to me like we'd align incentives if I'll give you 1/10th that for every year I live past treatment, maybe for the next 20 years (say you have a 50% success rate; I'm not objecting to you taking my money if you can cure me, I just don't want to pay you if you don't) I mean, maybe even I put ten years of that in escrow or what have you up front... I mean, that's the thing, it is experimental and experimental medicine seems totally reasonable if I'm probably gonna die with non-experimental treatment... but it seems like incentives are misaligned if a for-profit entity gets to offer up "experimental" treatment that I can't tell from sugar pills and they get paid regardless of efficacy. (I think paying for conventional medicine regardless of efficacy is probably fine, 'cause I think there are standards treatments have to meet before they become 'conventional medicine' - and obviously, 'experimental' treatments don't have to meet that high bar.)

lsc 8y ago
to be clear, I think I'd be even more comfortable with some sort of socially accepted procedure to sign up for really experimental trials done by some non-profit seeing entity. Some sort of procedure where you get several conventional doctors to sign off on, you know, that conventional medicine isn't going to cure you, then sign up for something that looks good, but hasn't been fully tested yet. I'm just saying, going to a for-profit entity and saying "give me something experimental" and then paying the same on success or failure seems like a pretty bad idea to me.
mikeash 8y ago
There’s a chance you’ll survive even without treatment, and giving you sugar pills is way cheaper, so I’m not entirely sure that this scheme gives the right incentives to potentially unscrupulous doctors.
- winstonewert 8y ago
  Let's say we have 100 people and they'll each pay you $100K if they are cured. If you give them sugar pills, let's say 1 recovers naturally and pays you a total of $100k. On the other hand, if you give them an experimental drug that cures say 10 of them, then you'll get $1M dollars. So you'd be willing to spend around $900k on actually curing them. That works out to $9k per person. So the incentives seem correct to me. The incentive is to actually cure you as long as the cost/benefit ratio works out. However, in order for a treatment that costs $100k up front with a 10% cure rate to work out this way, you'd have to pay $1M if the cure worked. I suspect that most people paying for this cure wouldn't or couldn't make that kind of payment.
  - mikeash 8y ago
    Asking people to potentially pay $100,000 for a $9,000 treatment seems unrealistic. You’re right that the incentives can align when the numbers are right, but I suspect realistic numbers leave you with bad incentives.
    - lsc 8y ago
      eh, I think that for better or worse, the 'prices will fall to match the marginal cost' ship sailed a long time ago; Yes, this is going to make Adam Smith mad, no, I'm not going to fight that fight anymore. I'd pay $200,000 for another 20 years of life without complaint even if the marginal cost for the treatment was super low. I mean, I think the real problem in misaligned incentives here is that there would be an incentive for the for-profit experimental treatment providers to keep things that should become mainstream "experimental" so that they could continue to charge for things that worked. But even so I still think my suggestion is better, at least for those paying for treatment, than the case in the story where people without medical expertise end up deciding to spend lots of money up front for treatments that explicitly are not generally accepted as effective by the medical community.
      - mikeash 8y ago
        It might work if there was a monopoly on treatment. If your choice was sit there and wait to die, or try an experimental treatment where you’ll pay $100,000 if you’re cured, sure, sounds great. But there’s usually more than one doctor or hospital that can do a given treatment. And some of them will figure they’ll do better if they charge $10,000 up front instead. Would you still take the $100,000 option? The only way I see it working is if the odds are very good. If there’s a 90% cure rate, paying a small premium for a cure in exchange for paying nothing if it fails makes decent sense. Any retailer with a generous return policy works this way, for example. This also helps the incentives to align: the loss from not handing out treatment will be much higher. I think that this means the “I’ll pay you if I live” model is a great model for normal medical treatment. You get diagnosed with cancer that has a 50% cure rate? Set it up so you pay 2x if you’re cured and zero if you die, and you’ll help ensure you get the best treatment. You get diagnosed with a disease that will bear your name and your only hope is some crazy experimental treatment that has a 1% chance of saving your life? Going that route will just leave you stuck between death and bankruptcy.

by lsc 8y ago

lsc 8y ago

to be clear, I think I'd be even more comfortable with some sort of socially accepted procedure to sign up for really experimental trials done by some non-profit seeing entity. Some sort of procedure where you get several conventional doctors to sign off on, you know, that conventional medicine isn't going to cure you, then sign up for something that looks good, but hasn't been fully tested yet. I'm just saying, going to a for-profit entity and saying "give me something experimental" and then paying the same on success or failure seems like a pretty bad idea to me.

mikeash 8y ago

There’s a chance you’ll survive even without treatment, and giving you sugar pills is way cheaper, so I’m not entirely sure that this scheme gives the right incentives to potentially unscrupulous doctors.

winstonewert 8y ago
Let's say we have 100 people and they'll each pay you $100K if they are cured. If you give them sugar pills, let's say 1 recovers naturally and pays you a total of $100k. On the other hand, if you give them an experimental drug that cures say 10 of them, then you'll get $1M dollars. So you'd be willing to spend around $900k on actually curing them. That works out to $9k per person. So the incentives seem correct to me. The incentive is to actually cure you as long as the cost/benefit ratio works out. However, in order for a treatment that costs $100k up front with a 10% cure rate to work out this way, you'd have to pay $1M if the cure worked. I suspect that most people paying for this cure wouldn't or couldn't make that kind of payment.
- mikeash 8y ago
  Asking people to potentially pay $100,000 for a $9,000 treatment seems unrealistic. You’re right that the incentives can align when the numbers are right, but I suspect realistic numbers leave you with bad incentives.
  - lsc 8y ago
    eh, I think that for better or worse, the 'prices will fall to match the marginal cost' ship sailed a long time ago; Yes, this is going to make Adam Smith mad, no, I'm not going to fight that fight anymore. I'd pay $200,000 for another 20 years of life without complaint even if the marginal cost for the treatment was super low. I mean, I think the real problem in misaligned incentives here is that there would be an incentive for the for-profit experimental treatment providers to keep things that should become mainstream "experimental" so that they could continue to charge for things that worked. But even so I still think my suggestion is better, at least for those paying for treatment, than the case in the story where people without medical expertise end up deciding to spend lots of money up front for treatments that explicitly are not generally accepted as effective by the medical community.
    - mikeash 8y ago
      It might work if there was a monopoly on treatment. If your choice was sit there and wait to die, or try an experimental treatment where you’ll pay $100,000 if you’re cured, sure, sounds great. But there’s usually more than one doctor or hospital that can do a given treatment. And some of them will figure they’ll do better if they charge $10,000 up front instead. Would you still take the $100,000 option? The only way I see it working is if the odds are very good. If there’s a 90% cure rate, paying a small premium for a cure in exchange for paying nothing if it fails makes decent sense. Any retailer with a generous return policy works this way, for example. This also helps the incentives to align: the loss from not handing out treatment will be much higher. I think that this means the “I’ll pay you if I live” model is a great model for normal medical treatment. You get diagnosed with cancer that has a 50% cure rate? Set it up so you pay 2x if you’re cured and zero if you die, and you’ll help ensure you get the best treatment. You get diagnosed with a disease that will bear your name and your only hope is some crazy experimental treatment that has a 1% chance of saving your life? Going that route will just leave you stuck between death and bankruptcy.