Appendix S1 - Designing and Analyzing Clinical Trials with Composite Outcomes: Consideration of Possible Treatment Differences between the Individual Outcomes

R code to calculate power composite outcome heterogeneity test. The following will calculate the power to detect composite outcome treatment heterogeneity if treatment does not change cardiovascular death (β4 = 0 or hazard ratio = 1.0) and both MI and non-fatal cardiac arrest have a treatment hazard ratio = 0.7 (interaction term β5 = 0): COpower(1500,4100,−0.35667,−1.43508,−2.35138,0.35667,0). The following will calculate the power to detect composite outcome treatment heterogeneity if treatment does not change non-fatal cardiac arrest (β5 = 0 or hazard ratio = 1.0) and both MI and cardiovascular death have a treatment hazard ratio = 0.7 (interaction term β4 = 0): COpower(1500,4100,−0.35667,−1.43508,−2.35138,0,0.35667). The following will calculate the power to detect composite outcome treatment heterogeneity if treatment does not change MI (β1 = 0 or hazard ratio = 1.0) and both non-fatal cardiac arrest and cardiovascular death have a treatment hazard ratio = 0.7 (interaction terms β4 = β5): COpower(1500,4100,0,−1.43508,−2.35138,−0.35667,−0.35667). (DOC)