Small group processes relevant to data monitoring committees in controlled clinical trials: an overview of reviews

Abstract

The quality of the decisions reached by data monitoring committees (DMCs) is crucial. The aim of this paper is to identify factors that may make errors more or less likely in small, task-oriented, decision-making expert groups and to consider the implications of these factors for data monitoring committees. A systematic overview was carried out of reviews of empirical studies of small group processes and decision errors in small, task-oriented decision-making groups in laboratory or real-world settings, published between 1950 and 2002 (n = 57 included reviews). These reviews suggest that a number of factors may increase the likelihood that small groups will make poor and potentially erroneous decisions. The most important of these, in terms of empirical support, are: biased or overly directive leadership, expression of a limited range of opinions during group discussion, poor procedures for identifying or appraising the available information, and presentation of the available information in a way that is likely to result in biased perception of it. The main implications for DMCs relate to membership, the role of the chairperson, the information provided for DMCs and training for DMC members. Selection methods that encourage a degree of diversity within the DMC are recommended. Chairs of DMCs should be experienced members, who have the skills to facilitate a discussion, can manage conflict effectively and can be impartial. Adherence to a prespecified analysis plan is recommended to reduce the risk of error associated with strong evidence or excess information. Training in the use of methodical decision-making procedures, education about the factors that influence decision quality and an opportunity to participate in mock DMC discussions may be of benefit for new members.