Interpretable (not just posthoc-explainable) heterogeneous survivors bias-corrected treatment effects for assignment of postdischarge interventions to prevent readmissions.