Skip to main content

Table 1 Study characteristics

From: Intention-to-treat analysis may be more conservative than per protocol analysis in antibiotic non-inferiority trials: a systematic review

  All comparisons within studies
(N = 164)
Primary analysis population as per author
 ITT only 42 (25.6%)
 PP only 45 (27.4%)
 ITT and PP 55 (33.5%)
 Not specified 22 (13.4%)
Sample size per group in the ITT population, Median (IQR) 221.5 (129.5, 326.0)
Proportion of treatment arm in the ITT population that was included in the PP population, Median (IQR) 0.86 (0.79, 0.93)
Proportion of control arm in the ITT population that was included in the PP population, Median (IQR) 0.87 (0.77, 0.92)
Definition of ITT
 Based on assignment alone 54 (32.9%)
 Use of drug at least once 95 (57.9%)
 Other exclusion criteria used 34 (20.7%)
PP population clearly defined 138 (84.2%)
Definition of PP population
 Exclusion based on concomitant therapy 96 (58.5%)
 Exclusion based on incompliance 123 (75.0%)
 Exclusion based on lost to follow-up 118 (72.0%)
 Exclusion based on withdrawn from study due to other reasons 29 (17.7%)
 Exclusion for other reasons 28 (17.1%)
Description of methods for handling missing data 64 (39.0%)
Missing data methods used
 Missing data as failure 58 (35.4%)
 Tipping point analysisa 3 (1.8%)
 Multiple imputation 4 (2.4%)
 Last outcome carried forward 2 (1.2%)
CI reported
 2-sided 95% or 1-sided 97.5% CIb 142 (86.6%)
  1. CI Confidence interval, IQR Interquartile range, ITT Intention-to-treat, PP Per-protocol
  2. aTipping point analysis assumes that all missing patients in the treatment group were failures and all missing patients in the control group were successes
  3. bOther CIs include 1-sided 95% CI (N = 4), 2-sided 90% (N = 9), 2-sided 97.5% (N = 4). Five studies did not report any CI