Do you see anything about confidence intervals in the article? They’ve just divided one by something in the region of 475,000* and multiplied the answer by a million. The numerator here is an integral with something like a normal distribution, such that the sample size at present isn’t large enough to rule out a prevalence similar to that observed in the EU, or perhaps even a prevalence similar to that observed in the UK.
It seems to me to be another instance of creative accounting, like claiming that the initial R value was 15 (apparently just an artifact due to a sudden increase in testing), and it’s a bit disappointing to see that from the CECC.
(* I assume this is the number of first vaccine doses administered by the date of comparison, which is in itself dubious as it neglects the duration (several weeks?) during which post-vaccination blot clots can occur. In other words, some of the people vaccinated during the recent surge in vaccinations and included in the denominator may also experience blood clots in coming weeks but this hasn’t occurred yet – this is guaranteed to understate the value reported here.)