Simulation Code and Results for "The Modified-arm Fragility Quotient: Normalizing Statistical Fragility for Unequal Randomization"

Simulation study supporting manuscript on the Modified-arm Fragility Quotient (MFQ) This deposit contains R simulation code and complete results for a Monte Carlo study comparing fragility metrics (FI, FQ, IFQ, nIFQ) across 96 trial configurations with 10,000 replications each. Important Terminology Note: The manuscript uses MFQ (Modified-arm Fragility Quotient), which corresponds to nIFQ in the simulation output files: MFQ = nIFQ = FI / (2 × n_mod) where n_mod = number of subjects in the arm receiving outcome toggles during Fragility Index (FI) calculation. Simulation Design: Allocation ratios: ρ ∈ {0.5, 1.0, 2.0, 3.0} (1:2, 1:1, 2:1, 3:1 randomization) Intervention arm sizes: n_I ∈ {50, 100, 200} Baseline event rates: p_C ∈ {0.2, 0.3, 0.4} Effect sizes: Odds ratio ∈ {1.5, 2.0, 3.0} Total conditions: 96 Replications per condition: 10,000 Files Included: mfq_nIFQ_simulation.R - Complete R simulation code using Fisher's exact test mfq_simulation_results.csv - Results table with columns: n_I, rho, p_C, OR (design parameters) n_significant, power (trial statistics) mean_FI, mean_FQ, mean_IFQ, mean_nIFQ (fragility metrics) cv_FQ, cv_IFQ, cv_nIFQ (coefficients of variation) Key Finding: MFQ (nIFQ in output) maintains relative stability across allocation ratios (2.5-fold variation) compared to FQ (5.7-fold variation), making it superior for comparing trials with different randomization schemes. Related Publication: Heston TF. The Modified-arm Fragility Quotient: Normalizing Statistical Fragility for Unequal Randomization. Zenodo. 2025. DOI:10.5281/zenodo.17386377 Software: R version 4.5.1 License: MIT License