Addressing confounding through extreme cohort restriction

Addressing confounding through extreme cohort restriction