Ever felt the thrill of making crucial diagnosis? The pain of missing a diagnosis? Yeah, us too
Clinical reasoning is hard, so let’s talk about it. Our 1st post in a new series starts today. Let’s talk about Biases.
#MedTwitter
>Heuristic: a cognitive shortcut based on prior experience & pattern recognition
>Cognitive Bias: when a heuristic leads to misdiagnosis or inappropriate treatment
We need heuristics b/c reasoning every decision from 1st principles would take forever!
There is a fine line b/t a heuristic and a bias
For example: 80yoM w/ CAD p/w SOB, ankle edema, and alveolar filling on CXR. You start abx PNA
👍If he improves, PNA was a good call
👎If he gets worse because he has ♥️ failure, PNA was “premature closure”
Heuristics are based on knowledge and prior experience. The more you practice and learn, the better you become at using them. Biases happen when heuristics go wrong.
Check out this comparative list of heuristics and biases.
Chart by @CaseyMcQuadeMD
Most biases fit into 1 of these 2 categorie:
1⃣You forgot the base rate of a disease - base rate neglect, representativeness restraint, availability bias
2⃣You want to be right (and don't we all?) - confirmation bias, premature closure
40yoF w/ obesity p/w RUQ pain, anorexia
🌡️VS normal
🩺RUQ pain. Murphy’s sign negative. No icterus
🧪WBC 14, AST 300, ALT 400, HAV/HBV/HCV neg
🥩RUQ US: dilated CBD, no stones
>ANA = 1:320
Q1: Which diagnosis is most likely?
A passed gallstone is most likely based on the prevalence of gallstones, CBD dilation, & elevated LFTs (commonly caused by gallstones/cholangitis)
Your MS3 saw autoimmune hepatitis last month and thinks AIH is most likely here.
Q2: What biases are likely at work?
All are possibly correct (I can’t read your student’s mind...):
1⃣Gallstones are common, AIH is rare
2⃣They just saw AIH recently (availability)
3⃣They focus on the ANA (nonspecific) but disregard the dilated CBD
4⃣“No gallstones” is not “classic” of gallstone dz
So how do you combat bias?
‼️Turns out it’s very hard:
Things that may decrease bias:
-Doing an ”effortful pause” to check yourself for bias
-Using diagnostic frameworks (e.g @CPSolvers)
-Using ☑️lists
🧠Remembering to do these is hard even with practice!
Thanks for reading!
We’re going to keep this series going in the future. If you’re interested in reading more, check out these:
“Thinking, Fast and Slow” by Daniel Kahneman
“Sources of Power” by Gary Klein
pubmed.ncbi.nlm.nih.gov/32311734/
qualitysafety.bmj.com/content/26/2/87