A paper earlier this year detailed some pretty poor performance in reporting clinical trials, but the reporting on it was pretty poor as well. Reporters got tangled up in the data on trials that weren’t even legally obligated to report, missed the distinctions between industrial and academic trials, and more. Now the team at Stat has taken a more focused look at the problem. Clinicaltrials.gov has about 200,000 records, but 95% of those are not legally obligated to report data, as it turns out. Narrowing down on the 9,000 trials that are supposed to be reporting, though, provides some interesting figures.
Drug companies have long been castigated by lawmakers and advocacy groups for a lack of openness on research, and the investigation shows just how far individual firms have gone to skirt the disclosure law. But while the industry generally performed poorly, major medical schools, teaching hospitals, and nonprofit groups did worse overall — many of them far worse.
Indeed they did. The only two organizations in the entire set who complied with the legal obligations more than 50% of the time, as it happens, were big drug companies (Eli Lilly and AbbVie). At the other end of the scale, you have tiny, obscure outlets like Sloan-Kettering, who have a perfect record of being late on all 39 trials they’re listed for, with an average delay of 651 days. That figure also brings up another difference between the big drug companies on the list and the academic institutions – not on only are the former doing a (relatively) better job of reporting trial data, but they’re doing it with a much heavier workload as well, since they run many more trials. Lilly, for example, is down for 139 trials in the database that are required to report, and GSK for 242.
To be sure, many of the drug companies may come out looking better because they’ve been more careful to file requests for extensions. But that’s more than most of the other institutions on the list have bothered to do. The article quotes Ben Goldacre as saying the “FDA and NIH penalties” might be necessary in order to get things enforced, but those two agencies would have to start by fining themselves. The article says that the FDA’s own staff scientists have violated the reporting laws 75% of the time, although compared to (say) Stanford, at a solid 95% noncompliance, perhaps that’s not so bad.