In this study, Karwath et al. determined β-blocker efficacy in reducing all-cause mortality in patients with heart failure and reduced left ventricular ejection fraction (LVEF). 13,659 patients from nine double-blind trials of β-blockers with LVEF less than 50% were divided into sinus rhythm or atrial fibrillation cohorts based on ECG or pace rhythm. Baseline characteristics, such as age, LVEF, and treatment with β-blocker or other therapies, were determined in order to assign meaningful phenotypic traits to clusters during the clustering process. All-cause mortality of patients was tracked over one to three or more years. Hierarchical clustering was employed to find the optimal number of clusters and number of dimensions, and the model with the highest gap statistic (i.e. the model least likely to have clusters assigned at random) was chosen. This model’s robustness and validation were evaluated using bootstrapping and leave-one-trial-out methods. For the sinus rhythm cohort, six clusters were identified. For the whole cohort, all-cause mortality was reduced with β-blockers, with an odds ratio (OR) of 0.74. However, while most clusters presented with reduced all-cause mortalities, cluster SR4 displayed no significant reduction in response to β-blockers. Cluster interrogation revealed SR4 likely consisted of older patients with less severe symptoms and lower heart rates. In contrast to the sinus rhythm cohort, the atrial fibrillation cohort did not display a reduced all-cause mortality in response to β-blockers, with an OR of 0.94. Out of the six cohort clusters the cohort, only cluster AF2, composed of younger patients with fewer frequencies of MI, displayed a reduction in all-cause mortality with β-blockers (OR = 0.57)