Assuming an estimated successful cleansing rate of 85% for the treatment arms, a 9.0% noninferiority margin, and
a 1-sided significance level of 0.025 powered at 85%, it was determined that 287 patients were necessary for
each treatment arm. The safety population, by definition, included all randomized patients who received treatment. The intent-to-treat (ITT) population included all randomized patients who received treatment and produced efficacy assessments for the Aronchick and/or Ottawa scales. Patients who met the criteria for the ITT population and did not have a protocol deviation during the study were included in the per-protocol (PP) population.
The categorical efficacy endpoints, the Aronchick and Ottawa scales, were summarized by the percentage of patients successfully cleansed for each outcome. Noninferiority was demonstrated for these endpoints if the 1-sided 97.5% confidence interval (CI) for the treatment difference (P/MC minus 2L PEG-3350 and bisacodyl tablets) was greater than "9.0%. If the noninferiority criteria were satisfied, superiority could be demonstrated if the lower bound of the CI for the treatment difference was greater than 0%. A Wilcoxon rank sum test was used to compare the fluid assessment between treatment arms. Results from the patient acceptability and tolerability questionnaire were compared between treatment arms by using a c2 test
for pooled responses. Unless otherwise noted, efficacy and acceptability/tolerability endpoints were summarized by
using the ITT population. Descriptive statistics were used to summarize demographics, AEs, and serious AEs (SAEs)