Hierarchical composite end point analysis methods, such as pairwise win time and expected win time, allow for combining different clinical end points by accounting for time in clinical state.
This methodological review describes hierarchical composite endpoint analysis methods, such as win ratio and pairwise win time, for evaluating treatments in heart failure clinical trials.
Much work has been done on developing hierarchical composite end point analysis methods, which meaningfully measure the effect of a treatment for patients with heart failure. Two motivations for this work have been as follows: (1) trying to ensure that more severe outcomes are weighted more heavily in the analysis; (2) combining different types of end points such as death, number of recurrent hospitalizations, and continuous functional or biologic end points. Such methods include the win ratio, the win odds, and the proportion in favor of treatment. In this article, our focus is when all components are clinical end points such as death or hospitalizations and do not include continuous end points. We review these methods using HF-ACTION (Heart Failure: A Controlled Trial Investigating Outcomes of Exercise Training). We also describe recent methods for combining different clinical end points, which take into account the time a subject is in a particular clinical state. These include the pairwise win time, the restricted mean time in favor of treatment, the expected win time, and the expected win time against reference. We discuss the US Food and Drug Administration guidances and make general recommendations.
Leifer et al. (Thu,) conducted a review in Heart failure. Hierarchical composite end point analysis methods was evaluated. Hierarchical composite end point analysis methods, such as pairwise win time and expected win time, allow for combining different clinical end points by accounting for time in clinical state.