Figure 2. Sensitivity Analyses of the Primary Outcome.

One-sided 95% CIs with Bonferroni corrections are reported. The vertical line indicates the noninferiority margin of 0.1. The intention-to-treat analysis of the primary outcome with complete cases yielded 86 of 98 patients in the posterior surgery group and 81 of 106 patients in the anterior surgery group, without and with continuity correction. There were 3 crossovers in total; therefore, per-protocol analysis yielded similar results without and with continuity correction. Scenarios regarding missing data were calculated in full cases with all missing data coded as unsuccessful in both groups without and with continuity correction and in the predefined sample sizes of 140 patients per group with all remaining missing data coded as unsuccessful without and with continuity correction. In addition, 6 different fragility indices were calculated, showing the minimum number of patients whose status (successful Odom score) must change to convert the noninferior outcome to inferior. The fragility index in the complete cases was 14 without and 13 with continuity correction. In the full cases with missing data coded as unsuccessful, the fragility index was 7 without and 6 with continuity correction. In the predefined sample sizes for both groups with all remaining data coded as unsuccessful, the fragility index was 3 without and with continuity correction.