Why it is wrong

Wrong journal and wrong decade. The study is Ganzini and colleagues, published in the BMJ in 2008, not in the American Journal of Psychiatry in 2016. The debate itself later refers to «the 15% figure from Ganzini et al.», so it has the author right and has invented the rest — and the invented part is what lends the number its weight. That 15% is then used four more times as the measure of how often the safeguards fail.

Source: Ganzini et al., «Prevalence of depression and anxiety in patients requesting physicians' aid in dying», BMJ (2008)