The statistical significance of randomized controlled trial results is frequently fragile: a case for a Fragility Index

The statistically significant results of many RCTs hinge on small numbers of events. The Fragility Index complements the P-value and helps identify less robust results.