The Win Ratio Should Be Complemented by Other Win Statistics to Provide a Comprehensive Picture of Relative and Absolute Treatment Effects: The Case Study of the REPRIEVE Trial
Bibliographic record
Abstract
TO THE EDITOR—Davies Smith et al [1] report on a win ratio analysis of the REPRIEVE trial [2] to provide further insights into the effect of pitavastatin on cardiovascular outcomes in patients living with human immunodeficiency virus (HIV) infection. Although the primary analysis of the REPRIEVE trial used a simple composite outcome of major adverse cardiovascular events (MACE), this post hoc analysis used a hierarchical composite endpoint that ranks outcomes in order of importance. The analysis resulted in a win ratio at 8-year follow-up of 1.55 (95% confidence interval [CI]: 1.20–1.99), suggesting that patients treated with pitavastatin have better cardiovascular outcomes compared to those with placebo. However, the estimated win ratio should be interpreted considering the large number of ties due to the very low event rate in the trial, with 94.7% (14 298 034 of 15 089 328) of pairwise comparisons in the analysis having tied outcomes. The win ratio is a relative measure that does not account for ties and, in the presence of a large number of ties, can overestimate the treatment effect [3, 4]. In this report, we note that other win statistics that account for ties were not reported. The win odds is a relative measure that takes into account all patients including those tied with no events and may be a more appropriate statistic to report when there are a large number of tied pairs [5]. The net treatment benefit (NTB) is an absolute measure, which also takes into account ties, and its inverse is the number needed to treat (NNT). The NTB can thus be viewed as similar to the absolute risk reduction (ARR) typically calculated for conventional binary or composite outcomes as the difference in event rates between treatment and control arms. Previous authors have suggested that these complementary win statistics should be reported concurrently to provide a more nuanced overview of the treatment effect [6, 7]. Based on the reported raw data in the supplementary appendix, we have calculated the win odds, NTB, and NNT for each year of follow-up in the REPRIEVE trial (Table 1). For MACE by year, at year 8, the win odds is 1.02, NTB is 0.011, and the NNT is 89. This marked difference between the win ratio and win odds is attributable to the low event rate and large number of ties—that is, for most patient-pairs compared, pitavastatin did not lead to better outcomes. Reporting the win ratio alone without the win odds, NTB, and NNT may overrepresent the benefit with pitavastatin.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.156 | 0.415 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.005 | 0.003 |
| Bibliometrics | 0.003 | 0.003 |
| Science and technology studies | 0.001 | 0.005 |
| Scholarly communication | 0.005 | 0.011 |
| Open science | 0.003 | 0.003 |
| Research integrity | 0.005 | 0.009 |
| Insufficient payload (model declined to judge) | 0.015 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".