To Err is Human
Reflections on Medical Training, Giant Cell Arteritis, David Hume, and the Silicon Buttressing of Carbon-based Walls
My medical education carried a baseline expectation of perfection. Excuses, no matter how justified – I was never notified; that’s not my patient; I was up all night; my dog died this morning – never got you very far. In fact, they only ever made things worse.
Criticism and fear of humiliation were strong motivators underlying long hours and spartan conditions. Kombucha on tap? Two different kinds of cold brew? Not quite. You’d be lucky if there was a container of Jello or Ensure pudding in the mini-fridge for a late-night snack.
You studied hard and did your homework partly because you knew you would be “pimped” on rounds, quizzed in front of the team until you got the wrong answer and let the attending move on to someone less dumb.
The funniest jokes were the ones about how nothing you ever did was right. What do you ask the attending surgeon before cutting the end of a suture he just finished tying? “Too long or too short?”
Things are different today, partly for the better, partly for the worse. That’s another blog. But one lasting consequence of this ethic of perfectionism is this: you may forget the cases you had, but you never forget the ones you missed – the overlooked diagnosis; the surgical complication; the abnormal EKG; the buried lab; the chest x-ray finding you could swear wasn’t there the first time you looked, but now, upon review, there it unmi-frickin-stakably is.
The other day, I sat down with a new patient to take her medical history. “I have an interesting diagnosis,” she told me. “Have you ever heard of Giant Cell Arteritis? It’s a rare disease that causes headaches.”
Indeed I had heard of it. I’d even seen it a few times. And it’s always a popular question on the medical boards. Formerly known as Temporal Arteritis, GCA is a serious rheumatological condition associated with inflammation of the arteries in the head. Untreated, it can cause blindness or stroke.
While it may be relatively uncommon, I wouldn’t call it rare. The prevalence of GCA is estimated at 278 per 100,000 people in the U.S. over the age of 50. It should be on the differential – the list of diagnostic possibilities – for all patients in that age group with new or unexplained headaches, and can be easily ruled out with a simple blood test for inflammation called an ESR.
“Well,” she said, “I saw nine doctors over the past six months until one of them finally figured it out.”
“You must be exaggerating,” I said. But she wasn’t – she was understating.
She pulled out a sheaf of papers with a handwritten chronological list of all the doctors she had seen and all the studies that were done. I looked it over. She had visited a total of ten different doctors, some of them more than once. She’d had CT and MRI scans, and been treated with various analgesics, migraine medications, and several rounds of antibiotics before someone thought to send an ESR – which came back through the roof.
The doctors on the list were mostly affiliated with a certain distinguished New York medical center, and some of them I even knew professionally. I had little doubt that they were all perfectly familiar with GCA, and no doubt that had it come up on a test, they would have gotten that question right.
Now, in general, I’m smarter than to think I’m smarter than my peers. I view compliments from patients that include disparaging comments about their prior doctors as a major red flag – I assume that eventually, I’ll be next on the list. And while there are obviously some glaring exceptions, most of the doctors I know tend to be intelligent, hardworking, and conscientious.
Would I have made the same mistake? The main lesson of the Milgram Experiment, is that while everyone likes to think of themself as the exception, in reality and by definition most of us are the rule.
So what happened to be the problem? I don’t know. I wasn’t in the room where it happened. But let me propose some distinctions that may shed some light.
First, the environment.
There’s a big difference between getting something right on the boards and getting it right in the clinic. You take the boards in a nice, air-conditioned environment, hushed as a library. Most likely you had a good night’s sleep the night before, and a healthy, nutritious breakfast that morning. All external distractions are put aside, and the worst of them, like cell phones, are locked away.
The clinic is very different. The environment may be hectic. Either you or the patient may be running late. Emergencies, urgencies, and trivialities intrude. Cellphones, like the beepers of old, are a necessary evil and constantly demand attention with buzzes and beeps, like an incessant toddler pulling at her father’s pants. I’m not saying that it’s impossible to possess a calm and zen-like focus amid the chaos of life – like a master chef in a busy kitchen – just that it’s easier while taking a test.
Second, the mentality.
Sorry U2, but people find what they are looking for. Common things are common. Or, to use the medical cliche, when you hear hoof beats, think horses, not zebras. On the medical boards, the correct answer is likely to be something obscure, like GCA. But in the exam room, the correct diagnosis is overwhelmingly more likely to be something mundane, like migraine or sinusitis. In fact, to pursue rare conditions as the cause of common symptoms is a classic rookie error.
Behind every cognitive bias is a habit of mind. And while in general, you would rather be treated by an expert than by a rookie, the scenario of the uncommon diagnosis is sometimes paradoxically better served by the mentality of a rookie – perhaps in some ways analogous to the Buddhist concept of beginner mind? – than of an expert.
Third, the specialty.
Looking over my patient’s list of doctors, it immediately struck me that they were all specialists, mostly ENTs. Now, I have nothing against specialists. Some of my best friends are specialists! But remember – they chose to specialize for a reason. They would prefer to restrict their practice to their chosen organ system. They expect and deserve to see patients for whom the referring doctor has already done the basic work of ruling out conditions outside of their specialty.
We like to think of privilege as something that leads to advantages, but there are situations where it leads to handicaps. The common, big-city custom of seeing specialists first – a cardiologist for chest pain, an orthopedist for back pain, or an ENT for a headache that may be sinusitis – is one great example.
Symptoms can have different causes across multiple organ systems. A good primary care doctor will see the big picture; narrow down the diagnostic possibilities; identify the acuity; treat the most likely cause; and refer you to the proper specialist for further workup if necessary. Seeing the specialist first is a classic case of putting the cart before the horse.
Now, maybe I’m biased (you think?), but my best piece of medical advice to all patients is simple: find a good primary care doctor and see them first. And by the way, that does not include specialists who do primary care because they can’t fill their practice with specialty patients. A primary care doctor is someone who was specifically trained in – who specializes in – primary care.
Last month, our family drove down to Virginia to drop my son, Ariel, off at college. A few weeks later, he texted me, “I wanna read this philosophy reading w u. 5 pgs.” He emailed it over: David Hume, A Treatise of Human Nature.
Now, this is exactly the kind of thing I want him to be studying in college, and I was flattered that he asked me for help. But, man! Five pages can be short and five pages can be long! It’s been a while (i.e. college) since I’ve read any classical philosophy, and I forgot how difficult it can be to disentangle even one page of dense, discursive, sentences, especially when you have a fuzzy baseline understanding of the key terms.
I read it through twice and was left baffled, but with a sneaky suspicion that underneath the dense layer of text lay a few basic concepts that, if I knew what they were, would help me decipher the whole. I didn’t know who to ask about this particular essay so I did what – as I understand it – all the kids are doing these days: I asked ChatGPT for help.
It was amazing. Three short paragraphs later I had the basic idea, and after a few targeted follow-up questions, I returned to the original text. This time I read it through with more clarity – enough to navigate my own way, identify some key lines, discuss it with Ariel, and help him come up with a couple of good questions to ask in class.
It was like having an assistant who happened to be a philosophy grad student – not to do the work instead of me, but to provide enough guidance to enable me to do it myself. It made me better and it made me think: Wouldn’t it be great to have that in the exam room?
Speculation is cheap. Whatever the real reasons my patient’s diagnosis was missed, the root cause is human fallibility. Yes, the expectation in medicine is perfection. The problem is, that not only are doctors imperfect at best, but neither are we our best selves in every situation. Whether for personal or structural reasons, ultimately that, too, goes along with being human.
So when it comes to AI, I’m optimistic – more eager to be improved than afraid of being replaced. Partly because I’m lazy enough to view a future where machines work instead of humans with a benevolent eye – as long as it comes with a universal income. But mostly because, deep down, I believe that human beings are fundamentally irreplaceable.
For a patient, nothing can replace a human touch. We all want attention, empathy, and human understanding, especially when we are sick, and especially from our doctors. No one wants to be seen by someone who even acts like a machine, let alone be treated by an actual machine.
But for a doctor to have a virtual assistant listening in on the visit? One that never gets tired, never gets distracted, never feels the pressure, never has cognitive bias… and can even write the clinical note? That doesn’t sound too bad at all.
More to the point, if I had been one of those ten doctors and an assistant had whispered in my ear before I moved on to the next patient, “psst, Doc, don’t forget to send an ESR to rule out GCA”, I imagine I would have been very appreciative, regardless of whether they were carbon or silicon-based.