Skip to main content

Table 4 Handling of missing data in the principal analysis of the 140 studies that reported some missing data in their primary outcome and performed an analysis

From: Current practice in analysing and reporting binary outcome data—a review of randomised controlled trial reports

Approach to handling missing data in the principal analysis
 Available cases96 (69%)
 Multiple imputation9 (6%)
 Worst-case/best-case scenario18 (13%)
 Last observation carried forward2 (1%)
 Other11 (1%)
 Unclear14 (10%)
Performance of appropriate2 sensitivity analysis for missing data17 (12%)
  1. 1In one study, missing outcomes were imputed by independent assessors using a pre-defined set of rules provided in a supplementary appendix
  2. 2Defined as an analysis that varies the assumptions made about the underlying missing data mechanism