Background: Pharmacist-led anticoagulation management services (AMS) for direct oral anticoagulants (DOACs) reduce prescribing errors and enhance adherence, but have not demonstrated lower rates of stroke or bleeding compared to usual care, and their cost-effectiveness is unknown. We evaluated four anticoagulant strategies for patients with atrial fibrillation initiating therapy: warfarin AMS, DOAC usual care, DOAC population management tool (PMT), and DOAC AMS. Methods: We developed a Markov model with monthly cycles simulating lifetime risk of ischemic stroke, major bleeding, death, disability, and costs from a US healthcare sector perspective. Costs and outcomes were discounted 3% annually. Model probabilities were derived from a prior Kaiser Permanente comparative-effectiveness analysis. Other inputs from published literature and national data. Primary outcomes were direct healthcare costs (2025 USD), quality-adjusted life years (QALYs), and incremental cost-effectiveness ratios (ICERs). Sensitivity analyses assessed parameter uncertainty. Results: DOAC-based strategies yielded greater QALYs than warfarin AMS and were cost-effective at standard willingness-to-pay thresholds. Compared with warfarin AMS, DOAC usual care gained 0.4 QALYs (ICER $89,200/QALY), DOAC PMT gained 0.6 QALYs (ICER $66,700/QALY), and DOAC AMS gained 0.6 QALYs (ICER $64,500/QALY). DOAC usual care and DOAC PMT were extendedly dominated by DOAC AMS. At $120,000/QALY, DOAC AMS was preferred in 50.4% of probabilistic iterations, DOAC PMT in 36.3%, DOAC usual care in 11.0%, and warfarin AMS in 2.3%. Results were most sensitive to DOAC program effectiveness and DOAC costs. Conclusions: Pharmacist-led DOAC management is cost-effective compared with warfarin AMS for AF patients. These findings support broader adoption of structured DOAC management programs to optimize anticoagulation therapy.
Toward terminological clarity in digital biomarker research
Digital biomarker research has generated thousands of publications demonstrating associations between sensor-derived measures and clinical conditions, yet clinical adoption remains negligible. We identify a foundational



