I bought Paul Kalanithi’s memoir, “When Breath Becomes Air,” the day it came out and set it down on my kitchen table unopened. Several of my best friends from medical school did likewise. We had read Kalanithi’s recent articles and knew the story of the Stanford neurosurgery chief resident diagnosed with metastatic lung cancer — a diagnosis that 22 months later would prove to be terminal.

But instead of jumping in, for several weeks we shared email exchanges about what we could expect to find, afraid to confront the actual words on the page. We were all at the end of our residency training (several of us, including me, in surgery), and reading the book, we knew, would make real a fear that no scientific articles and patient stories could: that terrible and unexpected things happen to doctors, too.

I had never really thought about the way physicians die, even though I was seeing so much death around me. But in hospitals, death can feel routine. We encounter it in intensive care units, on the floors and in the trauma bays; we see it happen to elderly patients with dementia, to newborns and to healthy teenagers. And yet I cannot think of any colleague who does not consistently draw a hard line of separation between what happens to “them” — the patients — and what can happen to “us.” There is an otherness to the bizarre accidents and the exceedingly rare cancers: In hospitals filled with unlikely catastrophes, the statistics are always on our side.

We trust in the double standard: one recommendation for our patients, a different one for ourselves. We routinely counsel patients on their end-of-life care, and we inquire about living wills before surgery. But in asking around, I learned that none of my fellow residents have living wills, designated proxies or advance directives. I don’t have them either, nor have I had a real discussion about what I would want done if the worst-case scenario occurred.

“None of my fellow residents have living wills, designated proxies or advance directives,” writes Daniel Marchalik. (Courtesy of Daniel Marchalik)

I did end up having a conversation with my best friend, an oncology fellow. “You would know what I would want,” I said to him rather nonchalantly, expecting this to be the end of the conversation. “Of course,” he said, “so do you. We’ve always been on the same page.” We easily agreed that neither of us would want prolonged breathing or feeding tubes, or holes made in our windpipe, or ending up stuck in a coma. Unless, of course, that coma was reversible.

“What percent chance of reversibility?” he asked. I didn’t have an answer — “I guess 5?” My best friend, it turned out, would draw the line at 2 percent, but not if there’s significant brain damage involved. How much brain damage is significant? As we considered our end-of-life preferences, it became clear that neither of us knew what the other would want because we had never actually thought through our own stances.

Physicians, like most people, do not want to discuss the implications of their own mortality. We forgo difficult conversations, assuming that our wishes would somehow be innately known by our friends and families. Haven’t we always been told that all doctors want the same thing? Indeed, a 2014 survey confirmed that an overwhelming majority of physicians — almost 90 percent — would choose no resuscitation. Most doctors also report wanting to die at home rather than in a hospital.

Perhaps it is these general assumptions that make physicians not feel the need to explicitly discuss and outline their end-of-life preferences. In a survey of almost 1,000 physicians whose mean age was 68, almost 90 percent thought that their family members were aware of their wishes for end-of-life care. Almost half of those surveyed did not think their doctor was aware of their end-of-life choices, with 59 percent of those participants having no intention of discussing these wishes with their doctor in the next year.

But we know that conversations about proxies and advance directives should happen long before they need to be utilized.

A 2016 study found that physicians were as likely to be hospitalized in the last six months of life as were non-physicians. On average, they also spent more days in intensive care units at the end of life and were as likely as others to die in a hospital.

So why are doctors dying in hospitals and in intensive care units instead of at home, when we know that their wishes tend to align with avoiding extreme measures at the end of life? My best guess is that physicians and non-physicians alike are skilled at believing that bad things happen only to others. Repeatedly witnessing tragedy does nothing to temper this human tendency.

At the same time, questions about the end of life are never easy even if we try to pretend that they are, and dealing in absolutes and generalities is seldom helpful. We see patients spending years on life support or clinging to hopeless cancer treatments and agree that we would never want that road for ourselves. Most of the time, however, the issues are markedly less clear. Often, the calculus shifts as life itself changes, as the needs of significant others and children become factored into the equation.

For that reason, living wills or designated proxies cannot exist in isolation. Instead, these questions must begin with a lot of self-reflection and difficult conversations with our loved ones. But none of that can happen until we come to terms with the unsettling reality of our own mortality. Not reflecting on or discussing it, like not picking up a book, does not change our reality: that illness and death are often unpredictable, and that this is as true for us doctors as it is for our patients.