PROJECT SUMMARY/ABSTRACT: The use of direct oral anticoagulants (DOACs) for stroke prophylaxis in patients with atrial fibrillation (AF) has markedly increased in the U.S. over the past 5 years, now accounting for 2 in 3 oral anticoagulant prescriptions. This increase is driven in part by a perception that DOACs are simpler to use than traditional therapy (i.e., warfarin) given their monitoring convenience. However, effect monitoring is not the only relevant consideration. Incorrect dosing and poor adherence are both common with DOACs, contributing to increased rates of adverse events, including bleeding, stroke, and death. To address the substantial clinical concerns associated with safe and effective ambulatory DOAC use, Kaiser Permanente Colorado enrolls patients treated with a DOAC in the anticoagulation management services originally developed to monitor warfarin therapy. In this proactive DOAC management model, all patients prescribed a DOAC are comprehensively evaluated and longitudinally monitored by anticoagulation specialists. Although this approach may prevent adverse events, it is very resource-intensive. It is possible that targeting patients who are at risk of dosage errors or adherence problems could provide the same benefit more cost-effectively. In 2016, Kaiser Permanente Southern California adopted such an approach. Their data-driven DOAC management model relies on administrative reports to detect potential dosing and adherence concerns and refers those at risk to anticoagulant specialists. This difference in practice amongst otherwise highly similar health care organizations provides a unique natural experiment to evaluate the effectiveness and efficiency of these management approaches. Using robust methods from pharmacoepidemiology and computer simulation methodologies, and with Kaiser Permanente Northwest, which does not provide system-level DOAC management services, as a reference population, we propose to determine the comparative (1) safety, (2) effectiveness, and (3) cost-effectiveness of each of these models of DOAC care. Preliminary evidence suggests both proactive and data-driven care models may improve stroke and bleed outcomes relative to usual care. DOAC management services must be studied at the health system level as health systems bear most of the cost burden for DOAC management and are responsible for making decisions about implementing care models. Our proposal and setting are uniquely suited for such an assessment. This study will provide reliable evidence on the comparative effectiveness of these care models more quickly and efficiently than would be possible with a randomized trial, and results will have greater generalizability due to lack of selection bias. Results from our careful analysis of these innovative programs will answer important and timely questions concerning optimal strategies for ambulatory DOAC care. They will have immediate impact on clinical practice, leading to safer use of DOACs and imp...