2019, book 15: The laws of medicine: field notes from an uncertain science, Siddhartha Mukherjee

I still have not finished Mukherjee's The Emperor of all maladies, but this short tome was a quick, easy, and pleasant read.

Spoiler: here are the two laws.

1. A strong intuition is much more powerful than a weak test.
2. "Normals" teach us rules; "outliers" teach us laws.

The idea of Law 1 is that the result of a test in isolation is usually not very useful, and conversely that the result of a test in context, and, crucially, a test administered because the situation suggested it, is very powerful. The intuition here is that false positive rates may screw up the first case (especially when the true positives are extremely rare in the population). So test to confirm, not simply for the hell of it. (There is a strong connexion here with p-hacking or data dredging.)

The idea of Law 2 is that medicine is slowly moving in the direction of true science. For a true science, one needs a good idea of what phenomena exist—a catalogue of observations. Only with a fairly full catalogue can one write laws rather than guidelines or rules: one needs the outliers to prove (i.e., test) the law. (There is a strong connection here with the recent trends of highly specialized, or personalized, medicine.)