Skip to main content
. 2018 May 16;13(5):e0197302. doi: 10.1371/journal.pone.0197302

Fig 2. Non-inferiority assessment for the primary outcome.

Fig 2

Point estimate of the difference in retreatment between trial arms (Truncated treatment–Continued treatment) is depicted by solid boxes. Error bars represent two-sided 90% CIs for one-sided alpha of 5% and two-sided 95% CIs for one-sided alpha of 2.5%, upper limits of which correspond to one-sided 95% and 97.5% CIs respectively.