Intention-to-treat analysis: difference between revisions
Diff·revision 6 → 7·04:42, 17 Nov 2024
Difference between revision 6 and revision 7 of Intention-to-treat analysis. 10 lines changed; the page grew by 625 bytes.
| Revision 6 — 18:13, 23 Oct 2024 ForestPlotFinn (talk) expand §Why analysing as treated fails 2,691 bytes ±0 | Revision 7 — 04:42, 17 Nov 2024 TriumphThad (talk) add the trial acronym expansion at first use 3,316 bytes +625 | ||
|---|---|---|---|
| 20 | The practical consequence is that a striking difference between an intention-to-treat and a per-protocol result is informative about differential dropout rather than about the treatment.{{r|ich_e9}} | 20 | The practical consequence is that a striking difference between an intention-to-treat and a per-protocol result is informative about differential dropout rather than about the treatment.{{r|ich_e9}} |
| 21 | 21 | ||
| + | 22 | == Missing data == | |
| + | 23 | Strict intention-to-treat requires outcome data on every randomised participant, which trials rarely have. How missing data are handled determines what an "intention-to-treat" analysis actually is, and the label alone does not say.{{r|ich_e9}} | |
| + | 24 | ||
| + | 25 | | Approach | Assumption | | |
| + | 26 | |---|---| | |
| + | 27 | | Complete-case | Missingness unrelated to outcome — usually implausible | | |
| + | 28 | | Last observation carried forward | Outcome froze at withdrawal — usually implausible | | |
| + | 29 | | Multiple imputation | Missingness explained by observed variables | | |
| + | 30 | | Reference-based imputation | Withdrawers behave like the control group thereafter | | |
| + | 31 | ||
| 22 | == References == | 32 | == References == |
| 23 | {{reflist}} | 33 | {{reflist}} |