Trial analysis and interpretation in critical care using the evidential (likelihood) approach: Rationale and practical considerations

Selecting the optimal methodological framework for evidence synthesis presents a fundamental challenge in contemporary clinical research. In critical care, where many interventions yield inconclusive results under traditional p-value-based analyses, complementary analytical approaches can enhance our understanding of trial data. While frequentist statistics remain predominant and Bayesian methods have recently experienced a resurgence of interest, the evidential (or likelihood) framework offers a methodological perspective that potentially bridges these two inferential paradigms. In this Concise Translational Review, we introduce readers to the evidential approach. To present the evidential approach as an analytical tool for critical care trials, we demonstrate its application using data from two mechanical ventilation trials (the Alveolar Recruitment Trial [ART, n=1,010] and the STrAtegy for coMmunIty acquired pNeumoniA trial [STAMINA, n=214]) and one trial evaluating balanced solutions (Balanced Solutions in Intensive Care Study – BaSICS, n=10,520). We focus on how concepts and terminology translate across paradigms, the framework’s measures of effect (i.e., likelihood ratios, support values, and S intervals), proposals for its use in sequential analysis and trial monitoring, and how to report results from this framework in research articles. We propose that the evidential framework provides a clinically intuitive approach to trial interpretation by focusing on the relative evidence between competing hypotheses, thereby offering additional and complementary insights that align with clinical reasoning processes. To facilitate implementation by the scientific community, we have developed an interactive Shiny (open-source web-based) application (https://fzampier.shinyapps.io/Likelihood_Shiny/).

National Heart, Lung, and Blood Institute

Fernando G Zampieri, Peter M B Cahusac, Israel S Maia, Nadir Yehya, Nuala J Meyer, Fan Li, Michael O Harhay