Hierarchical ventilator-free days with win method in acute respiratory distress syndrome treatment

Ventilator-free days (VFDs), a composite endpoint combining survival and mechanical ventilation duration, are increasingly used in acute respiratory distress syndrome (ARDS) trials to capture patient-centred outcomes beyond mortality. However, its interpretation and analysis are challenged by methodological limitations of conventional statistical methods. This study aims to investigate the interpretability and statistical performance of the hierarchical VFDs compared with the conventional versions for ARDS-related trials.

We applied a hierarchical composite endpoint framework, incorporating mortality and ventilation duration prioritised in sequence, to reconstruct VFDs censored at 28 days. Using the win ratio method, we carried out a post hoc analysis based on data from 10 high-quality randomised controlled trials and conducted simulation studies to assess its statistical performance.

Analysis revealed bimodal VFD distributions across trials, with peaks near zero and around 18-20 days. The win ratio method detected all the significant treatment effects that were identified by conventional analyses. Simulation studies suggested that the win ratio demonstrated greater statistical power, particularly when mortality was the primary driver of treatment effect, outperforming traditional methods.

The hierarchical VFDs endpoint, analysed using win statistics, provides a more sensitive and interpretable approach by distinguishing the contribution of mortality and ventilation duration components to the overall treatment effect in ARDS trials compared to the conventional approach.