Life is a bunch of statistics and probabilities and we humans have a strong tendency to want to simplify those away. We are further very naturally bad at stats (probably due to our amazing pattern finding abilities).
Heck, I think what makes a programmer good is something that can easily get in the way of fields like medicine. Good programmers like to create abstractions to put things into neat boxes. Programming is an exercise in generalization and specialization and, unfortunately, that can drive people to thinking "Oh, these diseases are alike so lets put them in the same box". That particularly gets in the way because MOST people won't experience complications from illness. Consider measles blindness, 30 million people get measles a year, 60,000 will get blindness. That's a 0.2% chance of developing blindness as a result of measles (1 in 500). That can lead to unfounded skepticism because your observed reality "I don't know anyone that's been blinded by measles!" might make you think that the risks are lower than they are.
And, heck, as a programmer if you have a method that fails at 1 in 500 cases you might even be justified in punting fixing that thing.
