Edited by: Prof Matthew Sydes (MRC Clinical Trials Unit at UCL)
Published in Trials
Independent Data Monitoring Committees (IDMCs), often called Data and Safety Monitoring Committees (DSMC) or Boards (DSMB), are the only trial oversight body which routinely sees accumulating, comparative data in clinical trials. For most late-phase randomised controlled trials (RCTs), this will include efficacy data by arm. These committees are asked to give an opinion on whether the trial should continue as planned, to be stopped early or to be amended in some important way. Their recommendations need to recognise the implications for the safety and treatment of patients within the trial, and patients that could yet be recruited to the trial, and patients whose treatment decisions will be informed by the published results in the future.
Most researchers are aware that interim data can pose challenges: by nature, the dataset contains less information than intend at the end at the end of the trial: fewer patients, less follow-up; and potentially the dataset will be less clean. Decision-making maybe guided by statistical “stopping rules” which may be seen as guidelines or rules which can be broken if the implications are understood.
However, many researchers do not get any formal training or informal exposure to these issues before joining an IDMC. Statisticians commonly report to these committees for years before sitting as members of these committees for other trials and so are exposed to the potential issues. The other specialities typically represented on these committees, which are mostly clinical in nature, have far fewer opportunities for exposure. Most IDMCs never have to make a difficult decision so even people who have attended many IDMC meetings may never have been encountered a difficult decision.
Most IDMCs are advisory rather than executive and make recommendations. Depending on the trial’s set up, this maybe a recommendation to another (semi-) independent committee (Trial Steering Committee), the Sponsor, the trials group director or another individual or group. This group will usually, but not always, follow the IDMC’s recommendations.
There are a small number of helpful papers which serve as published case studies from which researchers can learn. These set out the data that were presented to the IDMC and the recommendations that they made. Some of these papers show the data for a number of interim analyses and the final analyses. We can never know the counterfactuals – what would have happened if another recommendation had been made – but we can learn what recommendations were made, why they were made and what response they received from the recipient of the recommendation.
To support further learning and development of the trials community, Trials invites submission of manuscripts setting out examples of interim (and final) data and the decisions that were made, and of reactions to IDMC recommendations.
IDMC members are not co-authors on trial results papers or protocol papers because this compromises their independence; however, co-authorship would be encouraged for these manuscripts to ensure that insight is captured from all parties. If necessary, Trials could explore how to blind individual trials where sharing lessons could be problematic.