Skip to main content
. Author manuscript; available in PMC: 2016 Jan 21.
Published in final edited form as: Ann Intern Med. 2015 Jul 21;163(2):127–134. doi: 10.7326/M14-1741

Appendix Table 1.

Upper 95% confidence bounds for HR and RD measures from a noninferiority study designed using RMST

Total study size Accrual rate (per day) The entire study time (days) Total number of events observed Estimated upper bound of 95% confidence interval for hazard ratio Estimated upper bound of 95% confidence interval for risk difference at 900 days
2216 5 949 160 1.56 4.4%
2172 10 924 176 1.53 4.0%
2094 30 908 182 1.52 3.6%

Abbreviation: HR = hazard ratio; RD = risk difference; RMST = restricted mean survival time

Estimates of upper 95% confidence bounds for the HR and RD were calculated, assuming the following fixed inputs: total sample size, accrual rate, and total study time. All these configurations were figured out, so that the estimated upper bound of 95% confidence interval for difference in RMST at 900 days can meet a noninferiority margin of 18 days (2% of the 900 days).