Data monitoring committees should be used more frequently and appropriately to monitor paediatric clinical trials, particularly those involving neonates.

Abstract

Randomised controlled trials (RCTs) are the principal research method used to assess the effectiveness of healthcare interventions. Approval by an ethics committee that has appropriately weighed the risks of the planned trial protocol is a prerequisite for any interventional research. Clinical trials may take place over a long period of time. For ethical reasons, it is desirable to ensure that there is no unavoidable increased risk that would harm patients participating in such trials. If that is the case, the protocol of the study may need to be modified or the study stopped prematurely. On the other hand, it is also important to ensure that a trial continues for an adequate period of time and is not stopped too early to answer its scientific questions. An independent data monitoring committee (DMC), consisting of a group of experts not involved in a study, and reviewing accumulating data from an ongoing clinical trial can play such a role. In general, safety monitoring is the major task for a DMC, and for this reason, it may also be called a safety monitoring committee. A DMC is usually indicated in trials of life-threatening diseases. In a clinical trial performed in a paediatric population, even with noncritical indications, a DMC might be needed in order to detect any potential harm to the patients as early as possible. This is particularly true in trials involving neonates, who are the most vulnerable patients. In this issue of Acta Paediatrica, Perrem et al. 1 report on the DMCs involved in neonatal randomised controlled trials published in four high-impact general medical journals from 2003 to 2013. The authors specifically looked at safety issues, including the documented involvement of a DMC, the stated interim analysis and the rules for stopping a trial or terminating it early. They found that 61.4% of the studies reported the presence of a DMC and 51.4% explicitly mentioned interim analysis, while stopping rules were only reported in 21.4% of RCTs. The results were better than in previous studies that looked at the presence of a DMC in paediatric trials in 2005–2007 in eight general medical and paediatric journals 2 or earlier in 1996–2002 in a larger selection of journals 3. Perrem et al. 1 wisely note that the studies in the four high-impact general medical journals may not be representative of the majority of neonatal trials. The authors conclude that there is still room for improvement, even though the majority of neonatal RCTs included in their study reported on DMC involvement and interim analysis. One of the unavoidable weaknesses of any such study looking at the use of DMCs in clinical trials, by assessing reports of trials published in scientific journals, is that it is not possible to reliably determine how the DMCs were really used. What was the composition, was the DMC really used or just nominated but not really involved in active monitoring of the study and was the monitoring carried out properly? While few paediatric studies involving DMCs exist, it seems that practices in such committees vary widely, and there is room for improvement and not just in reporting 1, 2. A DMC should be considered for all interventional trials in vulnerable populations, not just in drug trials. Children, particularly neonates, are inherently vulnerable populations. Critical indications, such as life-threatening diseases and prior knowledge or strong suspicions that a treatment being considered has the potential to harm patients, are other indications that a DMC is needed. Other possible reasons for setting up a DMC may relate to study design, including preplanned interim analyses for stopping the trial early, either because continuing is futile or positive efficacy has been demonstrated, or complex study designs where a possible modification of the study design based on unblinded interim data is intended 4. A DMC may not be beneficial if a clinical study needs to be performed in a short time frame that does not allow for the researchers to prepare appropriate information for a DMC. Other situations where a DMC might not be needed are clinical studies in noncritical indications, where patients are treated for a relatively short time and the drugs under investigation are well characterised and known for not harming patients 4. A DMC should be set up in parallel to finalising the study protocol, and it should be fully functional before the study starts enrolling patients. Many important factors related to the composition, size and operating procedures of the DMC must be considered. The members have to have experience and, for example, provide the clinical, statistical and trial methodology expertise needed. The members have to be independent from the study team and the sponsor and not have any nonfinancial or financial conflicts of interest. Working procedures have to be planned in advance. In blinded studies, the DMC may need to access unblinded intermediate trial results without communicating these to the trial staff. In such cases, the DMC meetings need to be held privately, while the results of the DMC deliberations need to be communicated as recommendations to the trial staff and possible sponsor. The working procedures should be set in writing, and the DMC should preferably meet before the start of the study. Recommendations on setting up and running a DMC are available 4, 5. The need for a DMC is not principally different between sponsored clinical trials intended for regulatory submission and academic trials, but the requirements in academic trials may be simpler. Despite this, all DMCs require resources to cover the extra work required from trial staff and funding to cover the expenses of the DMC members. The resources needed for a safety monitoring committee are not different from a DMC set up just to monitor safety, while a DMC that is required to assess interim analyses, on efficacy as well as safety, may require more work and resources. As a consequence, the recommendation to have a DMC, especially in paediatric trials, adds to the costs of these already expensive trials. It is advisable, therefore, to already have a plan for a DMC included in the proposal for study funding. The research on reporting on DMCs in neonatal or paediatric clinical trials, like the paper by Perrem et al. 1 in this issue, is important, but reporting the presence of a DMC in the final publication of a trial is just the end of the process. As well as the need to report on the use of DMCs, work is needed to increase the use of DMCs in neonatal or paediatric clinical trials, and, just as importantly, education and training is needed on how a DMC should be set up and operated properly. It might even be worth considering developing paediatric guidance concerning DMCs in neonatal and paediatric clinical trials.