That's only when you count every unsuccessful intervention as a "cause". The research paper that created this factoid counted every possible way medicine could have prevented the death, and did not weighting of other factors like me underlying illness or injury.

One of the paper's highlighted cases of " doctor-caused death" from its own summary , was a case where a kid had an unprovoked heart failure while running track, and the doctor forgot to tell him that doing that again might kill him, and then he kid ran again, and had another failure, and died. That's ridiculous to blame on faulty care.

If 10 people have heart attacks, and medicine saves 5, suboptimal care fails to save 3, and 2 cases were beyond the reach of even perfect care, that research paper would say medicine killed 3 and heart disease killed 2, ignoring the 5 saved. Medicine has room for improvement, but is an absolutely massive net win.