This week, we launched our FDAAA TrialsTracker which gives you a live look at whether individual sponsors and trialists are meeting their responsibility to report the results of clinical trials on ClinicalTrials.gov. A lot of work went into the tracker and making sure we got it right. You can read all about our methods, in detail, here but the short version goes like this…
Certain trials registered on ClinicalTrials.gov, the US trial registry run by the National Institutes of Health (NIH), are required to report their results. This requirement began with the passing of the FDA Amendment Act of 2007 (FDAAA) but different interpretations of the broad requirements laid down under FDAAA led to some confusion over who was required to report and when. In September of 2016, a “Final Rule” was published by the NIH that clarified and expanded how the reporting requirement under FDAAA would operate moving forward. The Final Rule came into effect on January 18, 2017.
Broadly sponsors are now required, under threat of penalty, to report results for certain trials, known as “Applicable Clinical Trials” or ACTs, within 1 year of the primary completion date of their trial. Notably, this requirement is for all trials meeting these criteria, regardless of the approval status of the product under investigation (you can ask for a certificate of delay if your treatment is not yet on the market). Any ACTs ending on or after January 18, 2017 would be beholden to these reporting requirements. Sponsors may only delay the reporting of results under special circumstances.
Currently ClinicalTrials.gov does not publicly note which trials are, or are not, required to post their results. This is problematic. However they do provide a lot of other public information that can allow you to reverse engineer the ACT criteria. This was hard work, as we describe in our paper: but we can now say, with a high degree of confidence, whether each trial is required to report its results.
The FDAAA TrialsTracker presents all this data in one place. It allows you to look at all ACTs and see which ones have or have not reported their results. We’ve built a number of useful filters to help you dive deeper into the data, and assigned each trial to the sponsor who is responsible for making sure results are shared.
There are a number of smaller details and considerations that went into creating the tracker, such as considering a trial “due” to report results after 1 year + 30 days instead of 1 year. You can read about all of the assumptions, decisions, and data wrangling that went into creating the tracker in our paper, available on the bioRxiv preprint server. The site was built by our fabulous DataLab engineer Seb Bacon, building on prior work by the equally fabulous Francis Irving.
We’ll be posting about interesting unreported trials over the next few days. The data updates, every day, so this is a very exciting time! We’re also keen to hear feedback on the tracker and our methods so if you have any insights, comments, questions, suggestions, or anything else you feel we should know about, please get in touch: email@example.com
Nick DeVito, Ben Goldacre