It’s the holiday season, a time for family, festivities — and shoddy medical practices?!
Okay, so maybe that’s a bit alarmist. But a new, holiday-related study has us thinking about how doctors are humans with cognitive biases, quirks, and flaws like the rest of us — and how the holidays may be a time we should be a bit more conscious of this than usual.
The holiday the researchers focus on is not Christmas or Hanukkah. It’s Halloween. Specifically, the study looks at whether doctors are more likely to diagnose kids with Attention-Deficit/Hyperactivity Disorder (ADHD) when they see them on October 31st. But, still, there are lessons from this study that can apply to any holiday, or really, any day in general.
An Economist-Physician
Anupam Jena is one of the co-authors of this new study (with Christopher Worsham and Charles Bray). Jena is a professor of health care policy at Harvard Medical School. He’s both a physician and and economist, which means he spent an ungodly amount of time in graduate school. But it also makes him uniquely qualified to study the behavioral economics of our health care system.
Jena and his coauthors have long used the tools of economics — in particular, analyzing “natural experiments” and crunching “Big Data” — to document startling facts about the sometimes flawed decision-making of doctors. For instance, in one study, Jena found that physicians suffer from “left-digit bias.” This is the same bias that businesses exploit when they price something at $4.99 instead of $5.00.
“We showed that if you look at people who go to the hospital with a heart attack who are 79 years old and 50 weeks — so literally about to turn 80 — those patients are more likely to be given a cardiac bypass surgery than someone who is 80 years old and two weeks,” Jena says. “They’re similar in age, but one group is perceived by the doctor to be in their seventies and the other in their eighties.” Doctors, on average, are apparently more interested in giving aggressive treatment to those they perceive to be in the “younger” group.
The path to Jena’s most recent study — about the effect of Halloween on ADHD diagnoses — began with another study he did on ADHD. This first study on ADHD, which was published in the New England Journal of Medicine, showed the effect of arbitrary cut-off dates for enrollment in public schools on rates of diagnosis and treatment for ADHD.
In many states, this arbitrary cut-off date is September 1st. Kids who turn five shortly after this date will have to wait almost an entire year to enter Kindergarten and, when they do, they’ll be relatively old for their grade. And for kids with birthdays right before this date, it’s the opposite. (I know this personally. My birthday is August 30th, and I was almost always the youngest person in my class.)
Kids who are young for their grade are, almost by definition, more immature than their peers. And the economists hypothesized that medical professionals would be more likely to diagnose them with ADHD because younger kids are more likely than their peers to display symptoms of hyperactivity or inattentiveness. There’s apparently a fine line between being just a young, excited kid and meeting the official criteria for a medical ADHD diagnosis.
To an extent, all medical diagnoses are subjective, but diagnosis of ADHD seems especially so. Doctors can’t just give a kid a blood test or an X-ray or something to diagnose them with it. They rely on things like reports from teachers and parents on behavioral patterns, quick, “snapshot” observations in exam rooms, and comparisons of kids to their peers to judge whether they are acting abnormally. This subjectivity — the reliance on human judgement in the absence of crystal clear, objective data — opens the door to all sorts of cognitive biases and errors long documented by behavioral economists and psychologists.
In Jena’s first study on ADHD, he told us last week, “we showed that there was this pretty dramatic difference in the likelihood that those August-born kids were diagnosed and treated medically for ADHD.”
The Genesis of The “Halloween” Study
It was after conducting this first ADHD study, Jena says, that he began thinking of other natural experiments to show how other random or arbitrary circumstances can affect the judgment of medical professionals.
A couple years ago, Jena says, he saw how excited his son was about getting candy, and his mind turned to Halloween.
Jena, Worsham, and Bray’s hypothesis in their new study is pretty simple: Halloween is exciting for kids and their displays of that excitement in the doctor’s office on the holiday could increase the likelihood that doctors diagnose them with ADHD.
Diagnoses of ADHD have exploded in recent decades. Nearly 1 in 11 American kids, aged 3 to 17, is now diagnosed with the disorder.
“And there’s a lot of debate as to whether children are being overdiagnosed with ADHD or had we been historically underdiagnosing children with ADHD, and now we’re just catching up,” says Christoper Worsham, a Harvard Medical School physician who co-authored this study. If doctors are over-diagnosing kids with ADHD, then that could obviously be a problem. That diagnosis, for example, often comes with giving kids medication that can have side-effects. If doctors are under-diagnosing ADHD, that’s also a problem. Kids, for example, might need special accommodations at school.
Evaluating a kid who may or may not have ADHD could be more tricky on Halloween. The doctor’s office might be festooned with fun decorations. The staff may be dressed up. The kid may be dressed up themself — and may be looking forward to trick-or-treating and getting candy. “There’s a lot of excitement being introduced to the diagnostic environment on this day,” Worsham says. Kids, he says, often have trouble channeling excitement, “and it turns into being restless and talking a lot and running around and moving around and looking at things. All of these are part of the diagnostic criteria for ADHD.”
What’s sort of brilliant about this study is it’s pretty much impossible to systematically peer into what’s in doctors’ heads and figure out what’s going on when they diagnose people with conditions. So the researchers look at this random, unrelated environmental factor to having ADHD — the fact the doctor is seeing these kids on Halloween — to get a sense of just how subjective and prone to error these decisions are.
Jena, Worsham, and Bray assemble a really impressive dataset. They use data from private health insurance claims to analyze more than 100 million doctor visits over more than five years. The economists compare rates of ADHD diagnoses on Halloween to the ten weekdays before and after.
The researchers find a significant Halloween effect. They find “a 14% increase in the rate of diagnosis of ADHD among children seen on Halloween compared to the surrounding weekdays despite these children having similar characteristics and estimated risk of ADHD diagnosis.”
Worsham says he suspects, in particular, that there’s one heuristic — in other words, a mental shortcut or simple rule of thumb — that may be leading doctors astray when they make an ADHD diagnosis: the representativeness heuristic. This is a mental shortcut where we judge someone or something based on a sort of representative ideal in our minds. In this context, the doctor may be thinking, “This is how a first grader should act,” Worsham says. “We have an idea in our head of what that is, and we are evaluating every other first grader against that idea. If we fail to recognize, I’m using this heuristic on Halloween — or I’m using this heuristic on a kid who is young for his class or I’m using this heuristic on someone who has a language barrier — it’s going to increase the chances of misdiagnosis.”
The researchers’ biggest hurdle with their Halloween-based methodology is proving that there’s no “selection bias” in their findings. That is, maybe the parents of kids who really do have ADHD are more likely to select Halloween as the day they go to see the doctor. Maybe these kids are systematically different. Like, maybe parents are so desperate to get their kids help that they’re willing to go to the doctor on a holiday.
But the researchers do a bunch of statistical work showing that the kids seen on Halloween are statistically similar to the kids who go on days surrounding Halloween. And the researchers also look at diagnoses of other conditions — including autism, asperger’s syndrome, and eating and personality disorders — and they don’t find a similar uptick on Halloween. Moreover, they also look to see if there is a similar uptick of ADHD diagnoses on Valentine’s Day, which also involves candy. They don’t find it.
What Does This Mean For Seeing The Doctor Around Christmas?
Of course, like on Halloween, kids also get excited about Santa and presents during the holidays. But many doctors offices are closed on Christmas, so the economists didn’t look at whether there is a similar Christmas effect for ADHD diagnoses.
We don’t have hard data on the question of whether the holiday season is a suboptimal time to seek medical care. However, there are reasons to believe that maybe it is a season we should be wary of when it comes to getting a diagnosis or treatment for something. For one, Jena says, staffing may be different and lower around Christmas. Practitioners could also be phoning it in or may be more distracted than normal— like by text messages or family drama or a yearning to leave and go be with loved ones.
“Anything that affects your judgment or the way you think about a problem could affect the ultimate diagnosis and outcome of the patient,” Jena says. “And we see this happen in a lot of other areas where doctors might get distracted. For example, we showed that when surgeons operate on their birthdays, they have higher mortality [rates] for their patients. And why would that be? We think it’s because someone’s birthday is a sort of distracting event.”
Worsham, who is a pulmonologist (a doctor who specializes in the respiratory system), says that doctors may rely on a different heuristic — again, a simple rule of thumb — during the holidays. “If I work in the hospital on Christmas, one heuristic will certainly be something like, ‘This guy came in with shortness of breath on Christmas morning. It must be serious… it must be bad if they’re showing up on Christmas.”
All this said, if you need care, both researchers say it’s a good idea to get it. “My main concern actually is underuse of care,” Jena says. “I worry about the person who delays going to the hospital.”
But, considering their growing mountain of studies about cognitive biases and flaws in clinical decision-making, Jena says he thinks the medical profession should work harder to nudge practitioners to slow down their thinking and more rationally consider the data in front of them so they can fight potential flaws in their decision-making. For us, as patients, he says “there’s never a harm to talking to your doctor to get them to explain their thinking.” Maybe you can even cite some of Jena and Worsham’s peer-reviewed studies next time you’re in the doctor’s chair.
If you’re interested in this field of research, check out Anupam Jena and Christopher Worsham’s new book, Random Acts of Medicine: The Hidden Forces That Sway Doctors, Impact Patients, and Shape Our Health