Re: Compliance with mandatory reporting of clinical trial results on ClinicalTrials.gov: cross sectional study
We thank Huser for his interest in our work, and aim to clarify some important points which he raises.
In our paper, we present an analysis of ClinicalTrials.gov which aimed to identify which clinical trials were compliant with the FDA Amendment Act 2007 (FDAAA 2007).[1, 2] In order to do this, we identified trials which completed in 2009, which were conducted with one site in the USA, and which were of drugs previously approved by the FDA. We conducted our search on the 19th Jan 2011. We found a low overall reporting rate of 22%. Huser questions our use of the field “Primary Completion Date”, and the type of results available.
We did not look at which results were available for each individual study. We note that the FDAAA 2007 requires information on demographics, primary and secondary outcomes, and information regarding agreements between the sponsor and chief investigator restricting discussion of results from the trials. In our study, we investigated if any results were available, and did not analyse the results for particular trials.
As we were assessing compliance with the FDAAA 2007, we used the definition of completion date from the Act: “the date that the final subject was examined or received an intervention for the purposes of final collection of data for the primary outcome, whether the clinical trial concluded according to the prespecified protocol or was terminated.” This definition maps onto the field “Primary Completion Date” within the ClinicalTrials.gov record. Whilst it is true the study may well continue data collection after this date for secondary outcomes (and so the completion date would be later than the “Primary Completion Date”), the regulations requiring reporting results clearly apply to the “Primary Completion Date”. We therefore used the “Primary completion date” to identify when a trial had completed from the point of view of FDAAA reporting requirements. However, where this was missing, we used “completion date”, as stated in the methods. This affected 19/738 (2.5%) of the “mandatory reporting required” records. If a trial did not give either, we could not include it in our study, as we did not know when it had completed by either definition.
Huser asserts that had we used “completion date” (when the trial actually ended), our findings would have been significantly different. We argue that the “completion date” is not referenced in the legislation, and therefore we did not use it in the analysis which we present in the paper. However, if we restrict our analysis to the trials which we categorized as subject to mandatory reporting, and which list a date in 2009 in the “Completion date” field, 138/527 (26.2%) of the mandatory reporting required trials had reported results by the date of our search in early 2011. This is not a large difference.
It would appear to us that however you look at it, neither the letter nor the spirit of the law is being followed.
1. Prayle, A.P., M.N. Hurley, and A.R. Smyth, Compliance with mandatory reporting of clinical trial results on ClinicalTrials.gov: cross sectional study. BMJ, 2012. 344: p. d7373.
2. Food and Drug Administration Amendments Act of 2007.