Implementation of mitigated fraction estimators in SAS based on existing R package

Abstract

Mitigated fraction is frequently used to evaluate the effect of an intervention in reducing the severity of a particular outcome, a common measure in vaccines study. It utilizes rank of the observations and measures the overlap of the two distributions using their stochastic ordering. In a vaccine trial, mitigated fraction is used to estimate the relative increase in probability that a disease will be less severe in the vaccinated group. SAS macros have been developed using SAS/IML in equivalence with existing R functions in MF package to estimate the mitigated fraction both for independent and clustered data. The macros also provide asymptotic and bootstrap-based confidence interval. The macros were evaluated using real life data from a vaccine study and were validated by comparing output generated by the equivalent existing R functions available in MF package.