Selecting participants that raise a clinical trial's population attributable fraction can increase the treatment effect within the trial and reduce the required sample size

Background

Detection of modest but worthwhile treatment effects in randomized controlled trials (RCTs) demands trials of large sample size. Approaches to decreasing required size of RCTs while maintaining power are needed.

Objective

The epidemiological concept of population attributable fraction (AF_p) was applied to the population selected for an RCT to assess its role in determining the size of treatment effect and the required sample size. The additional effect of efficacy of treatment specifically among participants at risk for attributable target events (relative risk reduction_{at risk} RRR_{at risk}]) was also examined.

Results

A model is described which accounts for size of treatment effect in an RCT based on AF_p and RRR_{at risk}: RRR_trial = (AF_p) (RRR_at risk). The increase in RRR_trial resulting from raising AF_p exceeds that possible under the traditional high risk/high response approach to trial design and allows a reduction in required trial sample size. AF_p can be estimated from studies of causation that determine both risk and attributable risk (AR) associated with specific risk factors.

Conclusion

Larger treatment effects within RCTs are enabled by choosing a target outcome having a specific cause and selecting participants at specific risk for that outcome. Using information about phenotypic and genetic predictors of AR may increase our capacity to select trial populations having high AF_p.