What Medicine Can Teach Us About Low Probabilities: A Personal Experience
I’m recently home after 10 days in hospital. It was meant to be a simple procedure, home the next morning, but two low probability complications arose. I was largely out to it for a few days, but then I was fascinated to watch as my doctors monitored numerous measures, ordered tests, adjusted treatments, collaborated and, most generally, made numerous decisions as they worked to get my systems back to something like normal. I’m delighted to report that they did a great job and I’m now fine–and thanks for asking.
Is there anything relevant to write here? I have two thoughts, both concerning how medicine has learned to deal with low probabilities.
The differential diagnosis
A saying in medicine is that “If you hear galloping hooves, think horses, not zebras.” That makes sense–consider first the most likely explanation, the most likely diagnosis. But medical students are probably even more strongly admonished to always bring to mind the differential diagnosis: think horses, yes, but pause and also bring to mind zebras, and consider that as a possibility if there are no horses.
It’s extremely wise, given the strength of the confirmation bias, to bring to mind the one or two next-best possibilities (the differential diagnoses). Then the currently favoured diagnostic hypothesis can be considered on its merits, along with other possibilities. Our strong cognitive bias to leap to a conclusion, then preferentially see evidence to support that conclusion, is a real danger. Low probability events DO occur, sometimes even two of them, as I experienced.
I can’t be sure that the differential diagnosis rule helped me, but I strongly suspect it did. So I’m all for it–when interpreting data, or in any other situation, guard against leaping to a conclusion too soon–always bring to mind at least one other possible interpretation. Here’s to the differential diagnosis!
Atul Gawande: Dr Checklists
Atul Gawande writes on a vast range of medical topics for The New Yorker. His writing is a pleasure to read. His book Being Mortal on the end of life is by far the best I’ve read on the topic. (My brother’s 92 year old mother-in-law gave me a copy of that book for my 70th birthday–an inspired gift!) Possibly Gawande’s most famous writing is about checklists, starting with a New Yorker article in 2007. He later published The Checklist Manifesto, a wonderful book telling the checklist story.
Aviation long ago discovered that checklists are a simple but highly effective way to minimise errors when busy people have to carry out complex tasks. These tasks may be highly routinised, even boring, but even so the checklist ritual reduces errors. Checklists can be even more vital when emergency overload very occasionally arises.
Gawande’s achievement was to translate the checklist lessons to the operating theatre and the intensive care unit. Evidence for the effectiveness of requiring and using checklists has now come in from numerous hospital settings, in poor as well as the richest countries. Errors may be low probability events, but appropriate checklists can lower the probability further. Hooray!
I don’t know to what extent checklists were used when I was in theatre or intensive care. But my surgeon told me he had read Gawande, so I was greatly reassured. Thank you Atul!
BTW, his other two books are great reads also.