Six million Americans live with atrial fibrillation (AF), a heart rhythm abnormality that is a major risk factor for stroke and only half of those AF patients receive oral anticoagulation (AC). Within the context of an ongoing collaboration between two large learning health systems, the investigators propose to study the impact of a shared clinical decision support tool embedded within the electronic health record designed to enhance guideline-based AC prescription. The hypothesis the investigators are testing is that the number of AC starts in patients seen by intervention providers will be higher than in patients seen by control providers. Additionally, the investigators are testing the hypothesis that the risk of not persisting on AC will be higher in the patients of control providers compared with patients of intervention providers.

The proposed project will examine the impact of a novel anticoagulant (AC) clinical decision support tool within the electronic health record (EHR) on AC prescribing behaviors among healthcare providers from two distinct health systems who receive it. Using a mixed-methods approach, the investigators will also examine how healthcare providers interact with the decision support tool to further refine the tool and measure its impact on anticoagulation prescribing behaviors. The investigators will also examine clinician engagement through provider interviews and detailed study of behaviors using electronic health record access logs. Finally, the investigators will refine the alert, develop a comprehensive dissemination plan, and create an implementation toolkit to widen the use of the alert at other learning health systems nationally.

After identifying a list of all cardiology providers and primary care providers (PCPs), the investigators will send an email to all providers notifying them of our research study and request for participation. The investigators will send this email through our data collection platform- Redcap-with a link to one item survey allowing providers to opt out. This follows the prior SUPPORT-AF II protocol which the investigators' Institutional Review Board (IRB) approved. Based on discussions with the IRB, the investigators do not anticipate requiring informed consent from providers or patients for executing this aim of the proposal. After removing providers who opted out, the investigators will randomize the remaining providers to either intervention or control in a 1:1 ratio, using randomly permuted blocks; randomization will be stratified jointly by provider type (cardiologist / PCP) and study site (University of Massachusetts/University of Florida). The allocation of the provider conducting the first visit for a patient will dictate the allocation for the patient.

Inclusion Criteria:

• Providers caring for patients aged 18 years and older with international classification of disease 10th revision (ICD-10) diagnostic code consistent with atrial fibrillation (AF) or atrial flutter who had a visit with a primary care provider or cardiovascular medicine specialist in the previous one year with the diagnosis of AF present as an active diagnosis in the electronic health record (EHR).
• Patients aged 18 years and older with ICD-10 diagnostic code consistent with AF or atrial flutter who had a visit with a primary care provider or cardiovascular medicine specialist in the previous one year with the diagnosis of AF present as an active diagnosis in the EHR.

Exclusion Criteria:

• Left atrial appendage closure
• Hospice status
• Providers who do not care for patients aged 18 years and older with ICD-10 diagnostic code consistent with AF or atrial flutter who had a visit with a primary care provider or cardiovascular medicine specialist in the previous one year with the diagnosis of AF present as an active diagnosis in the EHR.
• Patients under the age of 18; patients without a ICD-10 diagnostic code consistent with AF or atrial flutter who had a visit with a primary care provider or cardiovascular medicine specialist in the previous one year with the diagnosis of AF present as an active diagnosis in the EHR. Patients will also be excluded from this study if they are known to be pregnant or a prisoner.