Hacker News Reader

Comment by fullstackchris | Hacker News Reader

Did anyone hear NPR's reporting on this? They said the early study was originally halted BECAUSE of it's positive effects at preventing hospitalization and death... very weird, I hope it is a typo or I'm understanding it wrong...

View on HN

I didn't hear the NPR story, but studies can be halted when it's shown that the treatment is so clearly effective that it becomes unethical to continue administering non-treatments/weak treatments to your comparison groups. This usually indicates that research can and should advance to the next stage immediately.

Here's something to read on this (note the three criteria for a TL;DR): https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5024796/

I haven’t read anything about this study, but this is common for highly effective treatments.

Let’s say you have a randomized clinical trial with a control arm (placebo) and test arm (drug). If the vast majority of the test cases do significantly better (like 90% of the test cases), at some point it becomes unethical to continue. Why? Because you already have the evidence that the treatment works, so having the control arm becomes meaningless. You already know that the drug works, so there is no longer a point to collect more data AND you know that the control arm isn’t going to help half the patients. So, you stop the trial early so that you can get started with the approval process so that the drug can get to all patients faster.

Pardon my naive question, but aren't there more objectives for the clinical trial? Particularly testing for negative side effects? How are those judged if the trial is discontinued?

You don't need people to actually have covid to find side effects. So you don't need sick people in your control arm.

To be cleared though, both arms or the trial are people who have COVID. The control arm is not getting the drug. If there are side effects that occur in some combination with having COVID, they will find those.

Safety continues to be evaluated during Phase 3 (and after approval), but trials don't proceed to Phase 3 without establishing some reasonable baseline for safety in the earlier phases.

Data on the most obvious side effects are collected in the first phase which is a very small group of healthy patients who take the drug under medical supervision in increasing doses until the side effects are so bad they can't keep going. They use this data to establish safety limits for the rest of the clinical trial and continue monitoring the situation until well after the drug is approved.

This was a phase 2/3 trial. So there was some side effect data collected, but enough people went through the trial to see any large/common side effects.

I generally think of the phases like this (this is just how I think of it):

Phase 1 - is it safe? If you don’t see severe toxicity then you can go to phase 2.

Phase 2 - does it work? You look for good outcomes (cures, better survival, etc). If there is significantly better outcomes relative to placebo, then go to phase 3.

Phase 3 - does it work for a large population? What is the appropriate dose? Dialing in the numbers before public release.

Phase 4 - after you’ve made the drug public, monitor the larger population for side effects or adverse events. You can’t hope to cover all patients in a phase 2/3 trial, so you keep monitoring the first waves of patients that get the drug.

That’s correct. If a positive effect in the treatment arm of the study is found to be strong enough, it is deemed unethical to continue giving patients the placebo as control. Hence, they terminate the study and proceed to the next phase of approval

That it's unethical to continue giving half the participants a placebo makes sense.

I am curious though: do they ever continue with the trial, with the patients on the actual drug? Might that be useful for monitoring rare-ish side effects, even if the drug is highly effective?

They almost always do this as an approval requirement.

Gotcha, thanks.

I believe they've only halted recruiting for the trial, not the trial itself. The Pfizer CEO this morning also said they're continuing with plans for two other trials starting soon, which is in people with non-co-morbidities, and to household contacts of confirmed cases.

Presumably they're kicking the placebo participants out? As others say, it is not ethical to give them ineffective treatment (placebo), given the effectiveness of the drug has been clearly demonstrated.

It's a semi common issue for highly effective novel treatments. In order to continue the study you need to keep some portion as a control group.

I can't imagine wanting to keep people in the control group when the experimental treatment is so dramatically effective.

Maybe they meant that further study at that level was not necessary and they were moving on to later stage trials? In order to speed up the release I mean.

If something proves to be too effective, too early in a trail, it's no longer ethical to give placebos to the control group.

When something obviously works, it's standard practice to halt the study and seek out emergency use authorization.

> There were six hospitalizations and no deaths among the 607 patients who received Paxlovid within five days of symptom onset, compared to 41 hospitalizations and 10 deaths in the placebo cohort.

I guess halted because meaning they are confident enough to go to phase 3 with larger amount of trial participants?

Maybe to reduce risk to the control group?

Reply on news.ycombinator.com