Its important to remember: more people die from contact with the medical care system than die from breast cancer, prostate cancer and lung cancer combined. Indeed, these numbers are on the conservative end as they count primarily people killed by hospitals.

Numbers on the health care system writ large are hard to come by but some – admittedly controversial – estimates put medical care as the leading cause of death in the US, and by a fairly wide margin.

I often use “Drink Plenty of Fluids” as an example of common medical advice that not only has little scientific basis but might very well kill the patient.

The basic intuition, of course, was that people, when sick, often loose a lot of fluids and may become dehydrated. In turns out that the body is a complex equilibrium system (not unlike the economy) and that dehydration might be exactly what it wants. In particular, in some cases fluid loss is an effort by the body prevent excess build-up in the lungs or brain.

Now Robin Hanson notes the same thing might be true with breaking fevers. While in extreme cases fevers can be deadly, in many (most?) cases they are likely quite helpful and could even save your life

There has … been one randomised trial … in patients in intensive care … In 2005, [researchers] … studied 82 critically ill patients who did not have head injuries or other problems that make a high temperature risky. Patients were randomised to get either the standard treatment of antipyretics if their temperature went past 38.5 °C, or only receiving the drugs if their temperature reached 40 °C. As the trial progressed, there were seven deaths in people getting standard treatment and only one in those allowed to have fever. Although this difference was not quite large enough to be statistically significant, the team felt compelled to call a halt, feeling it would be unethical to allow any more patients to get standard treatment. …