My Daily Life Is a Game of Roulette

Medicine has never been about certainty, but odds.

Illustration of red cross made out of dice
Adam Maida / The Atlantic

Updated at 11:50 a.m. ET on October 21, 2021.

A pair of realities: This week, Colin Powell, the former secretary of state whose service under President George W. Bush is most prominently associated with the American invasions of Iraq and Afghanistan, died at 84 due to complications from COVID-19, despite having been fully vaccinated. And: There are conditions under which any generally healthy person would most likely have a seizure.

These facts, as strange as they seem in juxtaposition, are related—neither describes events at all out of the ordinary, and yet both arrive as a kind of shock, with the sense that something must be amiss. They share this: They’re matters of probability, and probability is a fact of life in medicine, and a chaos agent in discourse.

Powell’s passing sparked a fresh round of social-media venom from vaccine critics, with the general’s death serving as ammunition for their claims. The onslaught of told-you-so’s sent news networks scrambling to invite experts to explain why the death of a single vaccinated person did not prove the overall worthlessness of the COVID-19 vaccines.

In Powell’s case, there is plenty of evidence to point to: The man was 84 years old, having lived, in other words, a very long life and having arrived at the highest risk category for death from COVID-19. Over time, bodies simply break down—our defenses become frail, our organs fragile—and no amount of medical intervention, no matter how heroic or generally efficacious, can forestall the end we all eventually meet. Nor was Powell in sterling health by the time he became infected with COVID-19. According to Powell’s family, he was in treatment for multiple myeloma, a cancer that affects white blood cells. These cells play a crucial role in protecting the body from infection; without them, even the most effective vaccines cannot perform their role. (A vaccine, after all, is more or less a highly skilled agent training your immune system to fight invaders; it doesn’t matter if the vaccine expertly whips immune cells into shape if, when the time comes, they simply don’t fight.)

To which the vaccine skeptic will likely say: “Blah, blah, blah. That’s all just a long-winded way of saying vaccines don’t work. If you get your shots and you still die of COVID-19, that means the vaccine didn’t work. Case closed.”

That this implicitly misstates the promise of vaccines is perfectly understandable. This is where a healthy person’s hypothetical seizure—and my real ones—comes in.

If you are fortunate, you will live the majority of your life without ever having to think too much about medical decision making. (In fact, this is my hope for you, whoever you are: that you enjoy good health, and that if and when you do fall ill, your medical problems are straightforward, well researched, and easily treatable.) If you, like me, are a little less lucky, then you will become familiar with the way medical decision making actually works: probabilistically.

When I was 14, I was diagnosed with epilepsy. What this broad blanket term for dozens of syndromes means is that a person has unprovoked seizures, in my case both grand mal seizures—that is, the convulsive ones most commonly associated with the word—and myoclonic jerks, which are short, repeated, electric-shock-like contractions of specific muscle groups, typically in my hands, arms, and shoulders. This array of symptoms is categorized as Janz syndrome. For me, the disorder will be lifelong. It has no known cure.

When you are diagnosed with epilepsy, one of the things you learn is that a seizure is a burst of sudden, uncoordinated electrical activity in the brain that can be prompted by all kinds of disruptions—even in perfectly healthy, non-epileptic people. An ordinary person without any neurological abnormalities whatsoever may have a seizure if, for instance, they sustain a serious head injury, develop a significantly high fever, go into shock, suffer blood loss, or use particular prescription or illicit drugs. All of those factors lower what neurologists refer to as a person’s “seizure threshold,” the barrier between the average brain and a full-on electrical meltdown. For most people, the threshold is relatively high, and lowering it such that their likelihood of having a seizure comes to match, say, mine takes quite a disruption. That’s not the case for me. My seizure threshold is naturally quite low: It simply doesn’t take much to scramble the grid in my part of town. That doesn’t mean I’m constantly having seizures, or that anytime I run into a factor that could trigger a seizure, I will have a seizure; it just means my odds of having a seizure are higher than the average gambler’s.

Thinking in terms of odds helps patients with illnesses like mine consider treatments. Through one lens, no epilepsy medication works—at the end of the day, you’re still epileptic, and your life is still circumscribed by limitations that do not apply to the lives of others. But that implicitly misstates the purpose of treatment: Epilepsy medications aren’t intended to cure epilepsy, or to guarantee that an epileptic never again has another seizure. They’re meant to lower the odds that, on any given day, in any given situation, an epileptic person will have a seizure. Or to put it another way: They raise your seizure threshold to something closer to a non-epileptic person’s.

A windy day in an apple orchard with my kids is a roulette wheel, red and black. By nature’s accounting, my wheel is two-thirds black, and I can bet only on red. With anticonvulsant medication, I can reshuffle the pockets to 50–50; with a good adjunctive therapy, I can cut the black down to a quarter of the game or less, and mostly forget that I’m even placing a bet, though I always am.

And so it goes for medicine in general. In order to live, we need to forget that we’re placing bets, though we always are. So many of our treatments are so efficacious and well established that their benefits are never really called into question, and we take for granted that they simply work, when the reality is that they mostly work, most of the time, for most people.

The COVID-19 vaccines are no different. Elderly people living in states with higher vaccination rates during the Delta surge were more likely to survive than elderly people living in states with lower vaccination rates; unvaccinated Americans are 11 times more likely to die from infection with the Delta variant than vaccinated Americans; fully vaccinated Americans 65 and older are 94 percent less likely to be hospitalized than their unvaccinated counterparts, per a CDC assessment published in the spring. In any of those categories, there are undoubtedly outliers; there always are. But medicine isn’t about playing the outliers. It’s about playing the odds.


This story has been updated to clarify the mechanism by which vaccines produce immunity.

Elizabeth Bruenig is a staff writer at The Atlantic.