Prevention trials: Design and analysis considerations

The design and analysis features of prevention trials are common to treatment trials. However, since prevention trials are usually performed in well populations who have a low incidence of disease, and require treatment and follow-up that typically last years, the operation, philosophy and ethics of prevention trials are different from those of treatment trials. Compared with the typical treatment trial, prevention trials have larger sample sizes and longer periods of follow-up. The potential for harm in a treatment trial arises simultaneously with the potential for benefit making it fairly straightforward to interpret, and to justify the balance of risks and benefits. However, for long-term prevention trials, the potential for harm begins to accrue with the start of the treatment, while the potential for benefit, if any, may be far in the future and then only realized by the delay or absence of disease. Trialists struggle to maintain patient retention and compliance in prevention trials that require long periods of follow-up. Factorial designs or large, simple designs may improve the efficiency of prevention trials. Analysis of prevention trials may be complicated by a lower than expected incidence of disease, performing and interpreting many subgroup analyses and dealing with missing data.