Paxlovid works. But don't ask anyone if it helped them.

Part II of the "reverse anecdotes" series.

A person who is not my mother-in-law. Getty Images, Daniel Grill.

This is the second installment of a miniseries I’m calling “reverse anecdotes.” That is, stories about medicine whose meanings could have done a 180° if just one detail had changed.

Last week, I wrote about how my daughter did not have Long Covid, but how it might have looked that way under a different circumstance.

This week, we’re talking about Paxlovid, the Covid-19 antiviral pill that had blockbuster effects in high-risk patients in a pre-vaccine era clinical trial, and still appears effective in the post-vaccine era for some high-risk individuals, albeit at far lower rates. (Here’s the NIH's take on who should get it, as of now.)

Now on to the best Paxlovid reverse anecdote that I ever could have dreamt up. Except I did not conjure it. It happened! It’s the story of how my mother-in-law took Paxlovid—and very well might have concluded that it wiped out her Covid-19 symptoms. But that’s not what happened, and I’m sure of it…

---

The time: July, 2022.

The era: Omicron subvariant.

The setting: My house in Cambridge, Mass.

The situation: There’s a Covid-19 outbreak in my home.

The details: My father-in-law (well, one of them. I have two), picked up Covid-19 in the airport on the way to visit us. My mother-in-law (well, one of them. I have two), tested positive the next day. We officially had an Covid outbreak in my home. Fun.

Both Fred and Kathleen had symptoms, but nothing worse than a typically nasty flu-like virus. They both felt fairly miserable, but I wasn’t remotely worried that either of them would be headed for the hospital any time soon. (Hooray for vaccines.)

That said, both were eligible for Paxlovid. Fred got a script called in by his doctor and started to take it right away. Kathleen was also prescribed Paxlovid, but there was a snafu.

Here’s the timeline:

Tuesday morning: Kathleen gets symptoms, and tests positive for Covid.

Tuesday midday: Kathleen is prescribed Paxlovid.

Tuesday evening: Something goes wrong at the pharmacy. Her Paxlovid is not ready. She takes ibuprofen and goes to sleep, still feeling pretty lousy.

Wednesday morning: Kathleen wakes up. She reports that she feels a million times better. She has more color in her face, looks peppier.

Wednesday later morning: the pharmacy situation is resolved. Her Paxlovid is ready for pickup.

Wednesday PM: Kathleen takes her first dose of Paxlovid. She reports that the taste is terrible, but she otherwise tolerates the drug well and there are no additional problems.

Over the course of the next few days, her lingering symptoms improve. She eventually tests negative and life goes on.

You see where I’m going with this. Let’s imagine what an observer might have concluded if she’d gotten her Paxlovid, as intended, on Tuesday evening.

Here’s a revised timeline:

Notice the ***? That’s the part where Kathleen, or Fred, or really anyone, might have said, “Wow, you went to bed feeling awful last night. But after just one dose of Paxlovid, you really bounced back overnight!”

Of course, we know that in reality she felt better without having taken the medication. And that’s the key here. When things happen, it is very difficult for people to know what the reason for that is. In this case, “tincture of time” could have easily been mistaken for "cause and effect” after taking a high-profile medication.

The fancy name for this situation is the post hoc ergo propter hoc logical fallacy. (I’m pretty sure I first learned of this formal concept via The West Wing.) More or less, it translates from the Latin as: “After this, therefore because of it.” But, as we know, that’s not always the case. (See: hilarious off-color example here.)

It bears mentioning that nobody is immune to this. Even physicians and scientists can easily be misled by what they see with their own eyes.

And it also is important to remember that people can be fooled in either direction. For example, what if someone took Paxlovid, but their symptoms didn’t improve for a week or longer? They might say something like, “Paxlovid did nothing for me.” But how do they know that? It’s possible that if they had not taken Paxlovid, their case would have progressed to severe illness, requiring a visit to the ER, or maybe even a hospitalization. Someone who takes a drug and has no severe outcome from their illness can’t possibly know whether the drug had anything to do with it or not. Only large amounts of data from well-controlled studies can tell us whether our treatments work, and what the effect sizes are. It might sound strange, but there are many therapeutics that are considered highly effective because they help just one in ten people who take them avoid a serious outcome. The other nine get no benefit. But on aggregate, “the drug helps people,” and we recommend treatment based on the overall picture.

So, if one person said, “Paxlovid worked great for me,” while another said “Paxlovid didn’t do jack squat for me,” my response to these two people would be the same: You have no idea, either way! Perhaps the first person would’ve gotten better no matter what. And perhaps the second person would’ve been much worse off without the medication. Again, this is why we like high quality clinical research.

My message to you: When you’re making decisions related to whether or not you want a medication, a vaccine, or even a test of some kind, don’t rely on anecdotes from even your smartest and well-meaning friends and relatives. In fact, don’t accept one-off anecdotes from physicians and scientists. (I’ll allow some wiggle room for a whole bunch of anecdotes that reflect months or years of clinical observations, but even that can be pretty dicey, as I’ve written before.) If someone is telling you what to do based solely on their personal experience with that thing, they are not thinking scientifically in that moment.

I hope you remember the concept of “reverse anecdotes” in the future. And I hope it helps you understand how wonderfully odd our world is, and why it’s always good to distinguish between when we are casually comparing notes and really talking science.

Do you have delicious examples of “reverse anecdotes”? I want to hear them! Please share them, or your questions, in the Comments section!