The Fragility Index Should Not Be Used For Sample Size Calculations in Clinical Trials.

Baer et al. recommend accounting for potentially “fragile” results when calculating a trial's sample size.1 “Fragility” is defined by the fragility index (FI), calculated as the minimum number of swaps from non-event to event in the treatment arm needed to undo the significance of a trial, with smaller FI values representing higher fragili ty. The FI lacks a theoretical basis and promotes an incorrect intuition of probability, leading users to make the wrong decision.

Source: Journal of Clinical Epidemiology - October 21, 2021 Category: Epidemiology Authors: Gail E Potter Source Type: research