For the problem of vaccine trials ending prematurely, is it possible to setup a voluntary post-approval trial? Online registration and randomizing people to the control and treatment group seems easy enough, although you’d need to somehow get pharmacists on board if you want to do blinding (presumably pharmacists won’t go along with it if they get a sealed letter asking them to inject the person in front of them with a saline solution). Unfortunately you probably need blinding, since people who know that they don’t have a vaccine presumably act different than people who know they have a vaccine :\
Another option is a trial where everyone gets a vaccine. Each patient is randomized to get one of multiple vaccines. That should provide better evidence about the relative desirability of each vaccine. I’m pretty sure that we ought to want to focus on which vaccines are best, not on whether vaccines meet some pre-defined standard of effectiveness.
I’m unclear why blinding is critical. It should have almost no effect on safety evidence. It seems likely to have a moderate effect on the quality of the efficacy evidence. But the harm from delaying vaccines for a few weeks seems a good deal larger than the harm from not having ideal information about which vaccine is most effective.
For the problem of vaccine trials ending prematurely, is it possible to setup a voluntary post-approval trial? Online registration and randomizing people to the control and treatment group seems easy enough, although you’d need to somehow get pharmacists on board if you want to do blinding (presumably pharmacists won’t go along with it if they get a sealed letter asking them to inject the person in front of them with a saline solution). Unfortunately you probably need blinding, since people who know that they don’t have a vaccine presumably act different than people who know they have a vaccine :\
Another option is a trial where everyone gets a vaccine. Each patient is randomized to get one of multiple vaccines. That should provide better evidence about the relative desirability of each vaccine. I’m pretty sure that we ought to want to focus on which vaccines are best, not on whether vaccines meet some pre-defined standard of effectiveness.
Alex Tabarrok has some related discussion in How to Vaccinate and Continue Clinical Trials.
I’m unclear why blinding is critical. It should have almost no effect on safety evidence. It seems likely to have a moderate effect on the quality of the efficacy evidence. But the harm from delaying vaccines for a few weeks seems a good deal larger than the harm from not having ideal information about which vaccine is most effective.