Unfortunately the study that claims this rate blatantly states that they massaged the data and took the amount of reported incidents and divided it by 6.6%, claiming that was approximately the rate of infection versus the population... when the database used was representative of the whole population already; they did discover that it was underreported by about 50% compared to the end result claimed on insurance, though. If we assume that these insurance claim values are accurate (remember this is not the US; they have a much better claim and treatment system there) and not cases of lazy doctors or fraud or whatnot, it's still higher in Japan than anywhere else in the world, sure, but it's closer to a rate of 2,000 cases annually (as rates of claims were about 900-800 in 2018-2019), which is significantly lower.