Thread by @DiseaseEcology

Should we use human challenge trials to assess #COVID19 vaccines?
I'd argue Yes.
S thread
(I'm very much looking forward to symposium w/ @mlipsitch @nireyal below; I want to understand why challenge trials weren't set up 3-4 months ago & therefore in progress now.) @JeremyFarrar

https://twitter.com/sconc/status/1291802575516446720

Several vaccines (6!) are in phase 3 clinical trials for COVID19. But unless I've missed a huge announcement, none of them include human challenge trials. Why not?

nytimes.com/interactive/20…

There's been tons of discussion on it with some strongly in favor & others strongly against.
ama-assn.org/delivering-car…
statnews.com/2020/06/23/cha…

A key reason I think challenge trials should be done *in addition to* the usual phase 3 trials that are being done now (w/ 10,000s of people) is that the larger trials won't provide data on a key issue highlighted by @mlipsitch & others- does vaccine reduce infectiousness?

In large trials currently underway it's not feasible to test every person every few days. I'm pretty sure none of current trials include swabbing everyone 2-3x per week, for example (please correct w/ a link if I'm wrong).

Why would this be critically important information?

B/c infections in trial will primarily be detected by symptoms (people getting sick). Sick people will be tested w/in 72hrs after reporting symptoms. Thus this won't provide detailed data on viral loads BEFORE symptom onset. Why does it matter?
nih.gov/news-events/ne…

B/c data show that most transmission occurs from people 1-3d before symptom onset (

https://twitter.com/DiseaseEcology/status/1271281847754846211

). If vaccine reduces infectiousness by reducing early viral loads, we'd need to swab people before they got sick to find out. That's not possible w/ normal phase 3 trial.

But it would be possible w/ human challenge trial. You'd vaccinate volunteers (young healthy people w/ low probability of severe illness) & wait a few weeks for immune response. Then you'd experimentally infect them w/ known dose (or doses) of virus.

Unlike in normal phase 3 trial you could swab all of them daily (or even more often) starting IMMEDIATELY after infection to carefully characterize their viral loads. You could even measure how much virus they exhale when breathing, sneezing, etc.

Since study would be done on v small number of people (10s of people) you could carefully monitor all these aspects of infection. At same time you could team up w/ contact tracing teams & do same frequent detailed testing on recent age-matched household contacts of cases.

Household attack rates are 10-20% so you'd only need to follow a couple hundred household contacts to get dozens of infections & produce robust dataset on viral loads over time since infection.
thelancet.com/journals/lanin…

This would give viral load data for non-vaccinated. The two datasets could determine if vaccine reduced infectiousness. This will be key issue of vaccine efficacy.

Some vaccines prevent illness, some prevent infection, & some only reduce infectiousness. But even just reduced infectiousness could be hugely helpful. But w/out human challenge trial we're won't be able to collect data needed to assess this rigorously.I hope they'll start soon.

Note that human challenge trial can't ethically be done on older people who might die if vaccine isn't protective so we still need phase 3 trial. But challenge trial would provide qualitatively different kind of information that is badly needed so both should be done ASAP.