It is surprisingly hard to track down what is meant exactly. It is not either of the options you listed here, but closer to the first. Chasing a very long chain

It is surprisingly hard to track down what is meant exactly. It is not either of the options you listed here, but closer to the first. Chasing a very long chain of citations to other citations, it appears this paper contains the original explanation of where the data come from: https://sci-hub.st/10.1007/s11205-015-1196-y https://sci-hub.st/10.1007/s11205-015-1196-y. They survey all of the possible healthcare goods and services available across OECD nations, make their best attempt to select a representative basket that is both available across all nations and reasonably similar, then estimate what they call a "quasi-price" per unit of good and/or service, to account for the fact that the actual charged price is often artificially suppressed or set to zero by government fiat. This seems to be done by scouring management accounting databases to figure out what the payers and providers consider to be reasonable reimbursement rates for accounting purposes, whether or not that is what they actually receive. I get what they're trying to do, but this probably explains some of the counterintuive results, because mostly people are probably thinking more along the lines of "add up all premiums paid to insurers, out of pocket expenses paid directly by consumers to providers, and all government outlays classified as healthcare" and that's how much your country spends on healthcare. That's a reasonable comparison to make, but as the blog and the OECD report both point out, it does nothing to account for differences in quantity and quality of healthcare goods being paid for. The problem is this discourse then inevitably leads to "well the US gets worse outcomes," but to what extent is that fair? The only reason I can walk today is because of US healthcare. If you incur a musculoskeletal injury that requires intervention in various different countries, how likely are you to fully recover? If you get cancer, how likely are you to go into remission? I don't necessarily know exactly what should be measured, but I know that when the discussion goes straight to lifespan, that is heavily confounded. Americans drive more, own more guns, are fatter. There has been tremendous industrial pollution in various places, though I don't know how that compares to the rest of the OECD. I wouldn't be surprised if we have more backyard pools. There are many, many reasons we might live shorter lives that have nothing at all to do with the quality of the healthcare we receive.