Methodological study
Issues in data monitoring and interim analysis of trials (the DAMOCLES project)
Doug Altman with Adrian Grant, Marion Campbell, Sharon McLeer, Anne Walker, Sheila Wallace (Aberdeen), Diana Elbourne, Stuart Pocock, Felicity Clemens, Matthew Sydes, Abdel Babiker, Janet Darbyshire, Max Parmar (London), and David Spiegelhalter (Cambridge)
Randomised controlled trials are essential to identify the best way to manage a particular health condition, but they are costly and they can raise difficult ethical issues. Clinical trials are only worthwhile if their findings are sufficiently robust to enhance future health care; and they are only ethical if the possibility of being harmed as a result of being in the trial is avoided. To ensure that this is the case, many clinical trials incorporate a data monitoring committee (DMC) to assess the data as they accumulate and recommend whether or not a trial should continue. The quality of the decision made by a data monitoring committee is crucial, and concern has been raised that ‘mistaken’ decisions may occur because of the ways in which data monitoring committees are conducted.
The project aimed to develop recommendations for the conduct of data monitoring committees in UK clinical trials in order to minimise the risk that ‘mistaken’ decisions will occur. The project was divided into four components.
A) A systematic review of available information about the conduct of DMCs. This section of the project included a systematic review of published papers regarding DMCs and their functionality; a review of published reports in which DMCs undertook trial-related decisions; and a cross-sectional study of the use of DMCs in the main reports of randomised trials published in high impact journals in a range of disease areas.
B) A review of the relevant literature on decision-making processes in other types of task-orientated small groups. The aim of this part of the project was to systematically review the empirical psychosocial literature in order to identify factors that make errors more or less likely in small task-oriented groups.
C) A survey of the plans and practices of DMCs in current and recently completed trials in the UK. This part of the project consisted of three inter-related surveys of: policies about DMCs as stated by major funding agencies; policies about DMCs as stated by a sample of protocols for ongoing trials; and recent practice about DMCs as stated by a sample of published papers.
D) Case studies of particular trials drawn from the sample survey and experiences of members of the study group. Empirical case studies were used to explore the decision-making processes and issues raised when DMCs make these difficult decisions.
On the basis of these findings the group developed recommendations for the construction and process of future DMCs. In particular, a template was produced to enable a DMC to develop a charter outlining how it intends to operate.
Publications: 114, 115, 116, 170, 171, 173
