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
| Approach to handling missing data in the principal analysis | |
| Available cases | 96 (69%) |
| Multiple imputation | 9 (6%) |
| Worst-case/best-case scenario | 18 (13%) |
| Last observation carried forward | 2 (1%) |
| Other1 | 1 (1%) |
| Unclear | 14 (10%) |
| Performance of appropriate2 sensitivity analysis for missing data | 17 (12%) |
1In one study, missing outcomes were imputed by independent assessors using a pre-defined set of rules provided in a supplementary appendix
2Defined as an analysis that varies the assumptions made about the underlying missing data mechanism