More than one-third of Medicaid beneficiaries receiving care coordination had at least one documented administrative burden. Together, these obstacles accounted for an estimated 18,822 hours of patient time.

In acute myeloid leukemia, clinically important biology may be visible in DNA methylation even when a defining genetic alteration is not detected. A large cohort study shows that…

Medication burden was the rule rather than the exception in this large memory-clinic cohort. Nearly two-thirds of patients had polypharmacy, and more than six in ten received at…

A deep-learning analysis of standard 12-lead electrocardiograms produced estimates of left atrial structure and function that were associated with future atrial fibrillation, heart failure, and cardioembolic stroke. The…

Alpha-1 antitrypsin deficiency–associated liver risk is not confined to cirrhosis or the Pi*ZZ genotype. In a large veteran cohort, the adjusted hazard of major adverse liver outcomes increased…

Recent AI & Computer Science coverage points to a stricter definition of progress: reusable representations and hidden-disease biomarkers matter only when data, outputs, evidence, and workflow fit are…

This week's strongest studies did more than improve prediction: they unified EEG interpretation, extracted biomarkers from routine data, and tested risk models across populations. The unresolved question is…

A machine learning model predicted death or heart failure hospitalization in transthyretin amyloid cardiomyopathy more accurately than established staging systems. Across held-out cohorts, its 3-year AUC reached 0.73…

A deep learning system analyzing routine 12-lead ECGs distinguished transthyretin amyloid cardiomyopathy (ATTR-CM) across diverse clinical cohorts. Used as a triage test rather than a diagnostic endpoint, it…

Historically marginalized patients were less likely to receive care-team responses to portal messages, particularly from the clinician they intended to contact. Differences in writing style accounted for substantially…