Carlos Guerreiro (npub12ld…d8ag) npub16ewvqtsleg8qnf6gdc5xldvjstjv5lmddj8um9s7tqxegdf5m6fqgw6fmh (npub16ew…6fmh) npub1s69n4td37hpm4q2hf6ar4dj2e52lnyp0n3y50htyd9vqkw406v8qjql9l2 (npub1s69…l9l2)
I understand what you're saying, but not everything is a scientific paper. This may be part of a larger study, but by itself I wouldn't expect that to be published as a lone review of insurance data. My guess is that he's using the insurance data to check his team's projections, but that's just a guess at this point. As you can see Akimasa Hirata has done other research on COVID infection/reinfection projections. For example:
https://pubmed.ncbi.nlm.nih.gov/36992217/
Personally, I'm taking it for exactly what it says until further notice one way or another. Insurance codes show that people are testing positive, on average, every ~4 months in a country where they're still testing. The earlier wave(s) data shows pretty much exactly what we've seen and expected. Dropping to 4 months looks pretty dramatic, but it's definitely not unexpected. If we weren't going about this so backwards in the US we'd be able to verify it with another data set, but that ship's sailed.