3 ms·

The proper way to look at this is a dynamic analysis, not a static analysis. For the first person to try the drug, it's a big gamble, but as each additional "r

The proper way to look at this is a dynamic analysis, not a static analysis. For the first person to try the drug, it's a big gamble, but as each additional "right to try" participant takes the drug, everyone else gains knowledge as to its effectiveness. Is this as good as stage 3 trials at assuring safety? No. But you have to count all the people who could have been helped but weren't because of delays against a no right to try policy.

jcranmer 8y ago
> For the first person to try the drug, it's a big gamble, but as each additional "right to try" participant takes the drug, everyone else gains knowledge as to its effectiveness. The plural of anecdote is not data. We have come to understand what it takes to run a good study for effectiveness of a drug: double-blinded, randomized, high-power control studies, with further guards against p-hacking (did you fail a phase III but find a subgroup that was responsive? great, do another phase III on that subgroup; turns out it's mostly spurious). The data that comes from right-to-try is mostly not useful because it's not going to tell you how much of the benefit comes from things other than the drug (such as the placebo effect).
olliej 8y ago
Cool, except right to try is not blinded -- you're paying to have access to drugs, so you have to get them. Any data from a person who buys a drug without knowing if it works or not (or kills them faster) is useless for everyone else (it actively reduces the trial population) due to placebo effect and trial controls. This is ignoring 50% being (i would say optimistic) for stage 3 trials, but right to try kicks in before stage 3 trials so suddenly you're dealing with stage 2 trials: and iirc only like 10% of candidates make it from stage 2 to stage 3, so now you've got maybe a 5% chance of it helping, and a much greater chance of it making your situation worse. And you're still not providing useful data that might help other people. It also ignores that there are typically multiple drugs with the same or similar targets running trials at the same time: which do you take given currently there's no evidence of which (or any) is effective, no data on how dangerous any (or all) of them are.

jcranmer 8y ago

> For the first person to try the drug, it's a big gamble, but as each additional "right to try" participant takes the drug, everyone else gains knowledge as to its effectiveness. The plural of anecdote is not data. We have come to understand what it takes to run a good study for effectiveness of a drug: double-blinded, randomized, high-power control studies, with further guards against p-hacking (did you fail a phase III but find a subgroup that was responsive? great, do another phase III on that subgroup; turns out it's mostly spurious). The data that comes from right-to-try is mostly not useful because it's not going to tell you how much of the benefit comes from things other than the drug (such as the placebo effect).

olliej 8y ago

Cool, except right to try is not blinded -- you're paying to have access to drugs, so you have to get them. Any data from a person who buys a drug without knowing if it works or not (or kills them faster) is useless for everyone else (it actively reduces the trial population) due to placebo effect and trial controls. This is ignoring 50% being (i would say optimistic) for stage 3 trials, but right to try kicks in before stage 3 trials so suddenly you're dealing with stage 2 trials: and iirc only like 10% of candidates make it from stage 2 to stage 3, so now you've got maybe a 5% chance of it helping, and a much greater chance of it making your situation worse. And you're still not providing useful data that might help other people. It also ignores that there are typically multiple drugs with the same or similar targets running trials at the same time: which do you take given currently there's no evidence of which (or any) is effective, no data on how dangerous any (or all) of them are.