Still reading digging in. In particular one reference in the second link[1]

Still not clear to me how they are generating the numbers for putting it at 3rd or 4th. I might have to read the paper rather than listen the author interview in my link above.

That said 98,000 dead from medical error in 2000, from the first link, would put it at 9th in the list that I linked:

https://www.cdc.gov/nchs/data/databriefs/db492-tables.pdf#4

from 2020. So even with that lower estimate it would put it in the top ten.

The definition of a death caused by medical error from [1] seem too board from the likely simplified explanation at least:

"Medical error has been defined as an unintended act (either of omission or commission) or one that does not achieve its intended outcome,"

That "or does not achieve its intended outcome" seems like it would count cases I would not want in a statistic like this. For example surgery to remove cancer to save the patients life did not achieve the intended outcome of saving the patients life so it is counted as death via medical error.

Probably have to look at the full paper to see how they applied the standard, but the pdf is not free on the site I linked. I might come back later and look for a free copy or another source.

[1] https://www.bmj.com/content/353/bmj.i2139/