Sailing the Muddy Waters of Clinical Medicine
A Reflection on Heuristics and Cognitive Biases in Clinical Reasoning
A cold snap has arrived, and with it snow, to the delight of my ski-racer kids and our friends in Windham, NY, where we go on weekends.
One of those friends is Hilla Steinberg, an ophthalmologist with a practice on the Upper West Side. While on a hike in the woods, she told me about a case.
Guess what, she said. This week I saved a patient’s life!
It’s not every day you hear that from the local eye doctor, I said. What happened?
A woman had come to see her about a stye. I’ve had them before, the patient said, but this one is not getting better.
On exam, the woman’s right eye was puffy and droopy, which you can see with a stye, but something was off. Hilla looked closer. It turned out she also had a subtle miosis, where the pupil on that side was slightly smaller, and less reactive to light.
Alarm bells went off in Hilla’s head, and she is not the poker-faced type.
What? the woman asked, reading her expression. I’m telling you it’s a stye – I get them all the time – don’t you dare tell me I have something else – I don’t have time for that.
Horner syndrome? I ventured.
Exactly, Hilla said.
Horner syndrome is a classic neurological triad of unilateral ptosis (eye drooping), miosis (small pupils) and anhidrosis (absence of sweating). It is caused by a blockage in the circuitous 3-neuron pathway that starts in the brain and ends at the eye.
The neurons originate in the hypothalamus, travel down the spinal cord, exit at the level of the upper back, course over the top of the lungs, travel back up the neck, follow the carotid artery to the middle of the head, and finally branch out to the eye.
Lesions can occur anywhere along this pathway, and include such lovely conditions as brain tumors, strokes, lung cancer, demyelinating diseases like multiple sclerosis, and vascular conditions like aneurysm or dissection (tear) of the carotid artery.
Horner Syndrome can also be due to benign conditions like migraine, or be idiopathic, which is medical jargon for ‘beats me’.
In Hilla’s patient’s case, the culprit turned out to be an occult carotid artery dissection, which could have disabled or killed her had it not been picked up.
The diagnosis could easily have been missed at that stage – and, I’m sorry to say, would have been missed in most cases from what I have seen over the years.
And yet, if I presented the case this way to any doctor (or medical student for that matter) they’d have to be asleep at the wheel to miss it.
Why the discrepancy? Why is a case like this so obvious when you hear it, and yet so tricky when you see it?
Part of the answer lies in the packaging. A well organized, clearly presented clinical case is like a glass of water, whereas a patient in the room can be more like a murky lake.
Ok, granted, we all know that life can muddy the waters – but what comprises the mud, exactly?
In 1974, the psychologists Tversky and Kahneman came out with a landmark seven page paper called “Judgement under Uncertainty: Heuristics and Biases.” In it, they described the heuristics, or mental “short cuts,” that people use to make complicated judgements, as well as the cognitive biases that skew them.
As it turns out, heuristics and biases are a great way to think about clinical and diagnostic reasoning in medicine. Let’s look at a few that are relevant to this case.
Probably the most common and famous heuristic in all of medicine goes like this: when you hear hoofbeats, think horses, not zebras.
Common things are common. If someone comes into my office these days with a fever, cough, and shortness of breath, they probably have Covid, not Legionella. A headache is most likely from tension or a migraine, not a brain tumor. And a puffy, droopy eye is probably a stye, not Horner Syndrome.
Thinking in terms of horses and zebras is a useful heuristic that doctors employ every day. But it comes along with a cognitive bias: the Zebra Retreat.
In reality, zebras do exist, as do their hoofs; the longer you practice medicine, the more likely a zebra will eventually gallop into your office.
If every time you hear hoofbeats you think of horses, you will miss the rare zebra. Staying keenly alert to red flags that raise this possibility – as Hilla did by noticing a subtle miosis – is how you overcome the Zebra Retreat.
Another heuristic and cognitive bias in medicine is the Bandwagon Effect, where a diagnostic label sticks to a patient.
The most common risk factor for a given condition is a history of that condition. For example, if a patient with a history of gout comes in limping with a swollen toe, then gout is most likely the diagnosis.
Unless it’s not.
Just because Hilla’s patient is convinced that she has a stye, carries that diagnosis, and was treated for it successfully in the past, doesn’t automatically make it what she had that day.
Everyone loves a bandwagon, and jumping off of one is psychologically difficult. Which bring us to Hassle Bias and it’s evil twin, Regret Bias.
Make no mistake, diagnosing Horner syndrome is a monumental hassle.
First you have to deal with the patient, who is already annoyed that she didn’t go to City MD, get prescribed a tube of erythromycin ointment, and call it a day.
Then there’s the insurance company, which requires prior authorization for the MRI and will torture you with bureaucracy and phone trees, tying up your staff for hours.
Not to mention the patients who are piling up, waiting to be seen, pressuring the receptionist, who is already frazzled from covering a colleague, who went on vacation to Costa Rica, and got stuck there with Covid…
I’m getting cold sweats just thinking about it: actually, you know what? It’s probably just a stye after all….
Unless it’s not. And you miss it.
If that ever happens, Regret Bias will rear its ugly head, and turn you into that doctor – the one who sends every patient for an MRI, all because of the one time when she didn’t.
There are many other heuristics and cognitive biases that we could discuss. They are not, of course, limited to medicine, and perhaps you know them from your own life and work.
Using them or avoiding them, navigating through them and around them, is what makes the waters of real-life medicine muddier than the Evian in the textbooks – also more challenging, engaging and fun.
I asked Hilla which mental and psychological forces were at play that day, and why she thought she arrived at the right diagnosis.
Easy, she said. I’m always afraid of missing something like that.
And there you have it – the ideal ingredients of a good doctor: someone who can sail the muddy waters with intelligence and knowledge, experience and fortitude, not to mention empathy and compassion – all topped off with a healthy dose of fear.