Opportunities for enhancing efficiency and reducing cost in large scale disease prevention trials: a statistical perspective.

Randomized intervention trials play an important role in the identification of practical approaches to reducing major chronic diseases in our society. Very few such trials are likely to be possible, however, in coming years unless we adopt procedures less demanding and less costly than those used in the past. Towards this end some elements of an efficient prevention trial are discussed, including aspects of study design, criteria for eligibility and exclusion, procedures for data collection, follow-up, endpoint ascertainment, and data analysis, and for ancillary studies. For example, possibilities for enhancing efficiency in the use of endpoint events include the selection of an appropriately weighted test statistic, the use of case-control or case-cohort sampling procedures in the processing of covariate data and specimens, and the notion of replacing a disease endpoint by an earlier or less expensive surrogate endpoint. Ideas for enhancing the efficiency of prevention trials will be illustrated in the context of a proposed Dietary Fat Intervention Trial for disease prevention in women.