This is a side point, but since you keep repeating this, sometimes going as far as referring to this as an "official estimate", it bears repeating from the primary source of this estimate, in the words of the authors.
Not many will have actually read the study, relying instead on an interpretation of the abstract not found within the study.
You might be surprised to read these expressly stated "Key Points" extracted verbatim in its entirety, in the words of the authors, for track and field athletes (WCA):
The UQM approach (the specific random-response technique) they used seems to be a work in progress.
While it guarantees anonymity, it does not guarantee compliance from the respondents. Indeed, when reading the study, the study is highly concerned with modeling all the forms of non-compliance among survey respondents. Although UQM is unable to measure any form of non-compliance, the authors still recommended ignoring 30% of the fastest respondents, to eliminate an observable over-estimation bias (approx. 40%, or 12%-points), subjectively suggesting, but not showing, the remaining 70% is not significantly biased.
In a followup study, one of the authors suggests that UQM could be lead to heavily inflated estimates for sensitive questions. In a similar doping prevalence study, they found the UQM method prevalence of 58% did not correspond well with other methods, included for cross-checking the result: SSC (19.8%), social projection (13.8% in own sport and 26.1% in all sports), and network scale-up (19.3%). This inflated effect was not seen for a non-sensitive question on herbal supplements. For more details see:
A potential inflating effect in estimation models: Cautionary evidence from
comparing performance enhancing drug and herbal hormonal supplement use
estimates
Ricky A. James, Tamás Nepusz, Declan P. Naughton, Andrea Petróczi