Number needed to treat: difference between revisions
Diff·revision 14 → 15·05:45, 17 Jun 2025
Difference between revision 14 and revision 15 of Number needed to treat. 7 lines changed; the page grew by 867 bytes.
| Revision 14 — 09:51, 20 May 2025 ArchiveBot (talk) bot: expand PMID to full citation 3,631 bytes ±0 | Revision 15 — 05:45, 17 Jun 2025 StudentSindhu (talk) expand §Number needed to harm 4,498 bytes +867 | ||
|---|---|---|---|
| 30 | Presenting the two side by side is more informative than either alone, and it makes explicit that the outcomes being traded are of different kinds: a cardiovascular event and a discontinuation are not commensurable, and no arithmetic makes them so.{{r|altman1998}} | 30 | Presenting the two side by side is more informative than either alone, and it makes explicit that the outcomes being traded are of different kinds: a cardiovascular event and a discontinuation are not commensurable, and no arithmetic makes them so.{{r|altman1998}} |
| 31 | 31 | ||
| + | 32 | == Limitations == | |
| + | 33 | The measure applies to the population studied. Baseline risk drives it, so a number needed to treat computed in an enriched high-risk trial population does not transfer to a lower-risk group; the relative effect transfers more readily than the absolute one.{{r|laupacis1988}} | |
| + | 34 | ||
| + | 35 | It also assumes a binary outcome. For a continuous endpoint such as weight change, a number needed to treat requires a threshold — for example, the number needed to treat for one additional person to lose 10% of body weight — and the choice of threshold changes the number.{{r|altman1998}} Where the threshold is not pre-specified, the resulting figure is a post hoc construction.{{r|ich_e9}} | |
| + | 36 | ||
| 32 | == References == | 37 | == References == |
| 33 | {{reflist}} | 38 | {{reflist}} |
| ⋮ | ⋮ | ||
| 35 | <ref name="altman1998">Altman DG. "Confidence intervals for the number needed to treat." ''BMJ'' 317(7168):1309–1312 (1998). PMID 9804726.</ref> | 40 | <ref name="altman1998">Altman DG. "Confidence intervals for the number needed to treat." ''BMJ'' 317(7168):1309–1312 (1998). PMID 9804726.</ref> |
| 36 | <ref name="lincoff2023">Lincoff AM, Brown-Frandsen K, Colhoun HM, et al. "Semaglutide and cardiovascular outcomes in obesity without diabetes." ''New England Journal of Medicine'' 389(24):2221–2232 (2023). PMID 37952131.</ref> | 41 | <ref name="lincoff2023">Lincoff AM, Brown-Frandsen K, Colhoun HM, et al. "Semaglutide and cardiovascular outcomes in obesity without diabetes." ''New England Journal of Medicine'' 389(24):2221–2232 (2023). PMID 37952131.</ref> |
| + | 42 | <ref name="ich_e9">International Council for Harmonisation, ''E9(R1): Estimands and Sensitivity Analysis in Clinical Trials'' (2019).</ref> | |
| 37 | 43 | ||
| 38 | == See also == | 44 | == See also == |
| ⋮ | ⋮ | ||
| 41 | * [[SELECT trial]] | 47 | * [[SELECT trial]] |
| 42 | * [[Surrogate endpoint]] | 48 | * [[Surrogate endpoint]] |
| + | 49 | * [[Randomised controlled trial]] | |
| 43 | 50 | ||
| 44 | {{DEFAULTSORT:Number needed to treat}} | 51 | {{DEFAULTSORT:Number needed to treat}} |