Win Ratio: Enhancing Outcomes Assessment in Orthopaedic RCTs

Orthopaedic trial design often forces a choice between multiple important outcomes. Imagine a trial where avoiding a devastating complication counts the same as sidestepping a minor setback—an analytic quirk rooted in traditional composite, time-to-event analyses. In such study designs, only the first event is counted, regardless of its clinical weight, so a superficial infection may eclipse a later nonunion or even death1. Moreover, patient-reported functional outcome measures often stand apart, analyzed in isolation rather than folded into the primary end point. These shortcomings misalign statistical evaluation with clinical priorities and patient experiences, highlighting the need for a more nuanced approach. The win ratio answers this call by embedding outcome importance directly into the analysis. Investigators rank outcomes from most to least critical (e.g., in order: death, major reoperation, minor complication, then functional recovery) and compare every treated patient with every control patient. Each pair “wins,” “loses,” or “ties” according to their first differing outcome; ties are resolved by proceeding down the hierarchy. The final win ratio is simply the number of treatment wins divided by control wins2. By integrating time-to-event data, recurrent events, categorical outcomes, and continuous measures, the method utilizes all collected data and prevents trivial events from diluting overall effects. This comprehensive use of information can boost statistical power and reduce required sample sizes—an advantage in orthopaedics, where recruiting large cohorts is challenging. A core strength of the win ratio is its alignment with clinical judgment. In the TRILUMINATE trial, a hierarchical composite of death or surgery, hospitalization for heart failure, and quality-of-life improvements yielded a win ratio of 1.48, underscoring patient-centered gains that conventional analyses might have missed3. Similarly, re-analyses of the DAPA-HF and DELIVER trials showed an approximate win ratio of 1.3 when hospitalizations, symptom scores, and mortality were assessed hierarchically4. These high-profile examples demonstrate how the win ratio captures multifaceted treatment effects in a single, interpretable metric. Bzovsky et al. introduce the win ratio to orthopaedics with clarity and practicality. The authors explain the mechanics of the method, contrast it with standard techniques, and reference its use so far in orthopaedic trials, including its use in a 2024 JBJS-published trial on distal femoral fractures—which ranked radiographic union, clinical healing, and functional recovery5. Their step-by-step guidance on building outcome hierarchies, managing ties, and limiting the number of components demystifies the approach for clinicians and researchers alike. Nevertheless, the win ratio has limitations. It relies on correctly ordering outcomes, as missteps in creating the order can bias results. Overly complex hierarchies or excessive tie-breaking rules risk obscuring interpretation, and current methods offer limited adjustment for confounders, confining its use to randomized trials. Adoption will also require education: trialists and clinicians must learn to trust and interpret win-ratio outputs. Despite these challenges, the win ratio represents a patient-focused advancement in trial methodology. By aligning analysis with clinical importance and leveraging all available data, the win ratio overcomes key limitations of conventional composites. The balanced roadmap put forward by Bzovsky et al. highlights both the method’s considerable strengths and its practical requirements. As orthopaedic research evolves to holistically measure patient benefit, the win ratio merits serious consideration—and thoughtful application—in future studies.