This is a linkpost for /r/discussion/lw/nzi/80_of_data_in_chinese_clinical_trials_have_been/
Missing link: should point to Science Alert.
Article says the Chinese State Food and Drug Adminstration (CFDA or SFDA) conducted an internal review of the drugs currently pending approval, and found out that in more than 80%:
the data failed to meet analysis requirements, were incomplete, or totally non-existent. [Also], many clinical trial outcomes were written before the trials had actually taken place. [...] The report found that pretty much everyone involved was guilty of some kind of malpractice of fraud. [...] even third party independent investigators tasked with inspecting clinical trial facilities are mentioned in the report as being "accomplices in data fabrication due to cut-throat competition and economic motivation".
There's no matching news item on the SFDA site; it probably doesn't have an official version in English. The article linked relies on this and that.
Compare and contrast with Scott Alexander's idea of making the American FDA regulate less. Two ends of a spectrum? Different cultures and markets leading to different outcomes? Similar situations but better hidden in the American case?
Compare and contrast with Scott Alexander's idea of making the American FDA regulate less.
The FDA actually does regulate less than the CFDA in this case. The FDA doesn't disapprove 80% of the drug seeking approval for misbehavior.
If you look at the Ranbaxy case the FDA is quite bad at detecting data forgery of generics companies.