Public health agencies versus science?

Oh man did I love Yglesias post today (it’s a public one, so you can read it all yourself), “Our public health agencies should follow the science.”  Some highlights:

Today my kid’s elementary school is hosting a pop-up vaccination clinic for 5 to 11-year-olds to get their second dose of the Pfizer vaccine, exactly three weeks after a pop-up clinic for first shots. It’s organized that way because of the CDC-recommended dosing schedule which, in turn, is in line with the Emergency Use Authorization the FDA granted the kiddie vaccine, which followed an EUA and eventual standard licensure for adults.

And that, in turn, was based on the dosing regime that Pfizer opted to use during its Phase III clinical trial to demonstrate the safety and efficacy of the vaccine when used in a large population that was randomly assigned either the vaccine or the placebo.

So everyone is following procedure and doing things by the book.

But unfortunately, the best available evidence suggests that this three-week spacing is not optimal and that the vaccine would be more potent if the doses were spaced further apart. There is not a gold standard Phase III clinical trial to demonstrate that, in part because Pfizer has no particular financial incentive to organize such a trial. But even though those kinds of randomization studies are the ideal form of evidence, they’re not the only kind of evidence that exists. And based on other evidence, it seems like we are sticking with a non-optimal dosing scheduling for arbitrary reasons of path-dependency... [emphases mine]

By the same token, the pharmaceutical companies also needed to come up with a timing strategy on which to run the clinical trial.

One consideration was that they wanted to come up with a schedule that would show high efficacy in the trial and get the medicine approved. But another consideration was that they wanted to run the trial quickly and get approval. So if some guy in a lab coat came to you and said “our best guess is that with a three-week gap we’ll get 90% efficacy but with a 12-week gap it’ll go up to 95%,” I think the right answer might be to say “fuck it, let’s do the three-week gap and get approval faster — that will both save more lives and make us more money.”

But once the vaccine was approved, many countries found themselves dealing with a shortfall of vaccine supply. A few countries responded to that by adopting a “first doses first” strategy of trying to give everyone one shot and then looping back around with second shots once all the first doses had been given. The theory was that this would maximize population-level protection because, as you can read in this extensive document from Canada’s National Advisory Council on Immunization, the first dose does offer meaningful protection.

To be clear, the motivation for the various different dose-timing strategies that different countries adopted was about stretching supply, not maximizing efficacy. But because different places tried different things, we ended up with what the clinical trial didn’t do — experimentation that lets us draw conclusions about the efficacy of different schedules. And a team of researchers out of Oxford found that an eight-week gap is optimal.

Since the United States continues to be newly vaccinating some people and certainly is officially encouraging the unvaccinated to get vaccinated, we ought to update our guidance to be in line with evidence about best practices — an idea that, frankly, we’d be well-advised to take in other areas too…

The government should be recommending Fluvoxamine

It never got the hype of ivermectin or hydroxychloroquine, but back in the early “throw shit against the wall and see what sticks” phase of the pandemic, another old generic drug that some clinicians thought might help was fluvoxamine…

but when subjected to a rigorous RCT, fluvoxamine turns out to actually work and generates a 30% reduction in hospitalizations.

That’s the same efficacy as Merck reported for its much-hyped new antiviral candidate molnupiravir. But fluvoxamine is cheaper and more widely available. What’s more, some scientists are worried about molnupiravir because its mechanism of action is that it induces fatal mutations in the virus. The concern is that if people don’t complete the full course of treatment, this could end up inducing non-fatal mutations which might worsen the pandemic. Largely for this reason, the FDA advisory committee vote on molnupiravir was closely divided with 13 in favor of approval and 10 against.

I’d be inclined to side with the 13 on the grounds that the downsides here are pretty hypothetical, if not for the fact that an equally effective alternative treatment — fluvoxamine — is already available. That consideration, to me, tilts things in the direction of the 10. But I think it’s a tough call. What’s not a tough call is that we should be recommending fluvoxamine as a safe and somewhat effective treatment. Unfortunately, the federal government’s official treatment guidelines on this have not been updated since April so they don’t include the results of the big RCT and don’t recommend the treatment.

And, this paragraph, is perfectly in keeping with my thoughts from this morning’s post:

I don’t really know what to say about this. I am a generalist political pundit working with an editor, an intern, and a part-time copy editor, and I am apparently more up to speed on the scientific evidence about Covid-19 treatments than the official NIH document. I found out about it because I follow Kelsey Piper on Twitter. If you’re lucky, your doctor does, too.

To be fair, I expect plenty of the people at the FDA and CDC are up to speed and do know this.  But the institutional culture is just so insanely “follow the established protocol” that they are unable to do what needs to be done (some political leadership needed here?).

In a recent conversation, a friend said to me something along the lines of “but what about the next pandemic?”  You know me, I’m an optimist and I’ve liked to thing we’ve learned so damn much that will be really valuable when the next pandemic comes along.  But, on the pessimistic side, what really disturbs me is how much we’ve learned that we are somehow not putting into practice.  That’s what really scares me about the next pandemic is we’ll know so much better what to do and just not do it.

The specificity of expertise

The fact that I know that viruses mutate for contagiousness and not for more/less virulence does not make me an expert on viruses, but it does mean that in this fairly specific domain, I clearly have more “expertise” than many public health professionals and MD’s who have frequently claimed that viruses typically mutate to become less contagious.  Dylan Morris‘ guest post in Zeynep’s substack back in May was great on the nature of viral evolution.  So, by reading that and figuring out the smartest people to follow on twitter on viral evolution were Zeynep (who’s no viral expert, but an expert in figuring out what “experts” to listen to), Morris, Michael Mina, and Carl Bergstrom and reading what they had to say made me more of an “expert” than those “viruses evolve to become less deadly” people who literally have degrees in medicine.

But, you know what, it’s really not that hard.  What we fail to appreciate, I think, is just how amazingly narrow expertise is.  I am a professor of Political Science, but if you read a couple good pieces in the Monkey Cage and Vox about some theories that explain an international conflict, you will assuredly be more of an “expert” in the matter than me.  Or, honestly, I literally teach about Congress in American government, but, again, read a few good articles that summarize social science on how Congressional committees work and you will know more than me.  I teach a class on Political Parties, but, again, spend not-all-that-long reading some good summaries of academic research on the role of amateur party activists and you will know more than me.  Now, if its a matter of partisanship or attitude measurement, don’t try me, but the reality is that expertise is really narrow and that most subjects are not so hard that an intelligent layperson can’t make sense of them when reading what good-communicating experts have to say.

As it happens, immunology is a ridiculously complex subject.  I’m reading a great book on it (Immune by Philip Detmer– also, you should totally familiarize yourself with Kurtzgesagt).  And to really understand the whole discipline it would surely take me years.  But that said, even then, there’s all sorts of bite-sized chunks that I can get my head around pretty well.  And if there’s somebody who’s a literal epidemiologist, they definitely might not be up on the latest research on something with Helper T Cells and Covid (or viral evolution!) just like I may not be up on the latest on party activists or Congressional committees.

So, we really need to be more humble about the limits of expertise and the very specific nature of expertise.  And, honestly, take lessons from Zeynep on how to think about issues you come to without built-in expertise.