Started a book called Evidence-Based Diagnosis: An Introduction to Clinical Epidemiology tonight. It’s pretty well-written for this sort of work.

It does have a great section on No Treat–Test and Test–Treat Probability Thresholds and how all that functions, which I have never understood — until tonight! So I do recommend it, but it requires a good bit of background knowledge. As do most books I read, so I should probably stop writing that. But as such things go, it’s not too bad. It’s mainly a book (unsurprisingly) about how to avoid deceiving yourself with bad data or bad presuppositions.