val lawless / shutterstock / arsh raziuddin / the atlantic

Regardless of who wins the Democratic primary, the 2020 presidential election will feature two septuagenarian candidates, either of whom, if elected, would be the oldest person ever sworn in to office.

It’s hardly surprising, then, that voters want more information about the candidates’ health, especially as the coronavirus rips across the country and presents an outsize threat to older people and those with preexisting conditions. Bernie Sanders recently suffered a heart attack. Joe Biden’s meandering statements have raised concerns about whether he’s lost a step. And Donald Trump, obese and averse to exercise, has long been the subject of psychiatric speculation.

The majority of Americans now think that presidential candidates should release their medical records. These demands for greater transparency are not unreasonable. The public deserves a basic understanding of candidates’ health. But as is often the case in medicine, the question is not whether, in some cases, a prescription could help. It’s whether, all things considered, the benefits outweigh the risks. In this case, a wholesale release of medical records would do more harm than good. Pundits and campaign operatives dissecting years of medical records would be only more fodder for political theater.

History is full of presidential ailments—from Abraham Lincoln’s debilitating depression to Franklin D. Roosevelt’s paralyzing polio. Dwight Eisenhower had a heart attack and a stroke while in office. John F. Kennedy was plagued by a number of conditions, including Addison’s disease, hypothyroidism, and severe back pain, none of which was publicly disclosed during his lifetime. Woodrow Wilson had a stroke so severe that his wife took over many of his routine duties. His successor, Warren Harding, had congestive heart failure and died of a heart attack two years into his presidency.

The length to which presidents have gone to hide their medical problems is captured by Grover Cleveland’s disappearance during the summer of 1893. As the United States slipped into what was then the worst economic crisis in its history, Cleveland decided to take a fishing trip.

Several months earlier, the president had noticed a small bump on the roof of the left side of his mouth—his “cigar chewing side.” After it had grown in size and roughness, his intrigue turned to concern. According to Matthew Algeo, the author of The President Is a Sick Man, the White House physician delivered an unambiguous assessment: “It’s a bad-looking tenant, and I would have it removed immediately.”

But Cleveland, who had just started his second term, worried that if he was officially diagnosed with cancer, the news would rattle the country’s teetering financial markets and further depress its flagging confidence. So he hatched a plan.

Under the guise of a fishing expedition, the president boarded a friend’s yacht and set sail for his summer home in Cape Cod. Once aboard, a team of six surgeons used ether and nitrous oxide to anesthetize the president, and in less than two hours they removed the tumor, five teeth, and part of his jaw—without making a single external incision, and leaving his signature mustache intact. Everyone involved was sworn to secrecy, and within weeks Cleveland was speaking publicly.

(It’s hard to imagine a president pulling off such a covert operation today. Despite his penchant for golf, Trump probably couldn’t get away for a week straight of Mar-a-Lago tee times during the coronavirus pandemic.)

All told, nearly half of U.S. presidents have suffered serious illnesses or injuries during their terms; eight have died in office.

Calls for presidential candidates to release their medical records presumably stem from a desire to understand whether an individual is likely to survive his or her term and meet the demands of what is perhaps the world’s most stressful job. But complete disclosure doesn’t tell us as much as we might think. Even with perfect information, doctors have a hard time prognosticating—especially, as they say, about the future.

Consider a landmark study in which doctors were asked to predict how long terminally ill patients could expect to live after being referred to hospice. Researchers found that 80 percent of estimates were inaccurate. On average, physicians were off by a factor of five. Those who knew the patients better were less accurate in their assessments.

And this was for patients who doctors knew were going to die. The uncertainty of predicting how long relatively healthy adults will live, including senior citizens running for president, is far greater.

Artificial intelligence doesn’t seem to fare much better. In a widely cited 2018 study, researchers from MIT, Harvard, and Stanford University used sophisticated machine-learning techniques to examine more than 6 million elderly Medicare patient records and predict who was going to die. They found that death, even among the sickest patients, was highly unpredictable. Among patients in the top 1 percentile of expected mortality, nearly half survived. And less than 10 percent of patients who ended up dying might reasonably have been expected to.

And so, for all medicine’s risk calculators and prediction models, it’s not clear that we’re better at anticipating how likely candidates are to survive than simply observing them on the campaign trail. If a candidate can endure months of flights, rallies, debates, interviews, town halls, diners, and selfies, that says more about their fitness to serve than their cholesterol level 15 years ago.

Bernie Sanders had a point, then, when he responded to questions about his fitness with: “Hey, follow me around the campaign trail, three, four, five events today. See how you’re doing compared to me.”

If medical prognostication is an imperfect science when practiced by doctors, it’s dangerous speculation when spun by pundits and politicians. Full release of medical records would offer ample ammunition to partisan operatives concocting tales about how treatable medical conditions make candidates unfit to lead, thereby simultaneously stigmatizing patients with those same conditions.

Hypothyroidism? Low energy. Depression? Unstable. Low testosterone? How will he stand up to Vladimir Putin?

Medical records are far more likely to be used as a political weapon than as a dispassionate tool to evaluate candidates’ fitness for office. Consider Hillary Clinton’s famed pneumonia of 2016. At a memorial event for 9/11 victims, Clinton became light-headed and had to leave, later revealing that she’d been diagnosed with pneumonia. That same weekend, reports emerged of fraud at Trump’s charitable foundation.

The following day, television news outlets devoted 13 times more coverage to Clinton’s pneumonia than to far more substantive allegations about Trump’s foundation. Celebrity doctors and political pundits speculated wildly about her stamina. Politico ran a story titled “Hydrated Hillary: 9 Times Clinton Quenched Her Thirst”—literally just nine photographs of Clinton drinking beverages.

It is true that the public deserves a sense of a candidate’s physical and mental fitness, and that candidates have a moral responsibility to disclose serious conditions that could impair their ability to serve. It’s just not clear that indiscriminate release of medical records gets us there.

If genuine concern persists about candidates’ health, we could consider creating an independent panel of doctors to perform a physical examination and offer impartial assessments of candidates’ ability to serve. Such a panel has been proposed by Jimmy Carter, among others, and could vet not only candidates’ health, but also when and whether the Twenty-Fifth Amendment should be invoked for an incapacitated president.

Indeed, a more important concern about whether a president can effectively govern is declining mental acuity over time—subtle cognitive changes due to nascent dementia or other neurological conditions. Unlike cholesterol levels and colonoscopies—which are objective, diagnostic, and widely used—no single test determines whether a person has the mental or emotional deficits we might worry about with presidential candidates.

While screening exams exist, such as the one President Trump took in 2018, diagnoses of cognitive decline are made based on comprehensive assessments of in-depth interviews, laboratory data, and imaging tests. But a candidate who runs an effective presidential campaign is unlikely to arouse sufficient clinical suspicion for a doctor to actually pursue and document any significant cognitive impairment.

Moreover, not all patients—least of all the powerful, high-performing patients running for president—routinely receive this type of cognitive evaluation. Doctors calibrate neurocognitive testing and treatment to the individual patient in front of them—and even with the same patient, different doctors may consider different diagnoses and record different data. Doctors’ notes, then, are more like journal entries than tax returns—and thus don’t allow for objective, meaningful comparisons of mental fitness across candidates.

And so, annual nonpartisan health evaluations, ones that include cognitive assessments, would be far more helpful than old medical records or letters from longtime doctors. This is partly because, as Carter notes, unlike candidates’ personal physicians, an independent panel wouldn’t have to “balance patient confidentiality and personal interest vis-à-vis the nation’s interest.” But it’s also because most neurocognitive tests are more valuable as repeated measurements used to track trends in cognitive and behavioral functioning over time—as could be performed by annual evaluations—than as snapshots of mental acuity available through old medical records.

Ultimately, however, no medical stamp of approval exists for presidential fitness. The best test of whether a candidate is fit to serve is their performance during the campaign (that is, if we ever get back to a normal campaign). Just as impeachment is a political process—not a legal one—determining whether someone is fit for the highest office requires the public’s judgment, not the doctor’s note.

Aside from a few rumors, Grover Cleveland’s maritime operation remained largely secret for nearly a quarter century, until the mission’s last surviving surgeon finally revealed what had happened. Cleveland, a cancer survivor who weighed 260 pounds, recovered swiftly and completed his second term without incident. He went on to enjoy more than a decade of post-presidency life, and died, according to one report, “a perfectly natural death.”

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.