Atrial fibrillation (AF) is the most preventable cause of stroke. However, despite widely available risk stratification tools, five options for oral anticoagulation, and evidence-based practice guidelines, anticoagulation for stroke prevention in AF is consistently under-prescribed. Data from this center (Brigham and Women's Hospital [BWH]) (1) demonstrate that fewer than 50% of outpatients with AF at high-risk for stroke according to 2012 Focused Update of the European Society of Cardiology Guidelines for the Management of AF (2) receive anticoagulation.

Aim #1: To determine the impact of electronic alert-based computerized decision support (CDS) on prescription of anticoagulation in high-risk AF patients in the outpatient setting who are not being prescribed anticoagulation for stroke prevention.

Hypothesis #1: Electronic alert-based CDS will increase prescription of anticoagulation by 80% in high-risk AF patients in the outpatient setting who are not being prescribed anticoagulation for stroke prevention.

Aim #2: To determine the impact of electronic alert-based computerized decision support (CDS) on the frequency of stroke and systemic embolic events in high-risk AF patients in the outpatient setting who are not being prescribed anticoagulation for stroke prevention.

Hypothesis #2: Electronic alert-based CDS will reduce the frequency of stroke and systemic embolism in high-risk AF patients in the outpatient setting who are not being prescribed anticoagulation for stroke prevention. Data acquired through this study regarding the frequency of stroke and systemic embolism will be used to calculate sample size requirements for a future clinical end-point driven randomized controlled trial of electronic alerts to prevent stroke in high-risk AF patients.

AF is associated with an increased risk of fatal and disabling ischemic stroke. Despite widely accessible evidence-based clinical practice guidelines, risk stratification tools (CHA2DS2-VASc and HAS-BLED) and five effective oral antithrombotic agents for stroke prevention (22), at least 40% of high-risk patients with AF worldwide remain unprotected because of failure to prescribe anticoagulation. The failure to prevent stroke in AF has become a critical international patient safety crisis. The root causes of underutilization include lack of provider and patient education about the importance of stroke prevention in AF, inadequate risk stratification, and concerns regarding the bleeding risk with anticoagulant therapy. The investigators were surprised to learn that even at BWH, underutilization of anticoagulation for stroke prevention in AF continues to be a concern (1).

The failure to prevent stroke in AF is similar to the crisis in VTE prevention among hospitalized patients from a decade ago. To address this critical patient safety problem, investigators evaluated the impact of alert-based CDS on VTE prevention (18). First, investigators designed software linked to our Electronic Health Record (EHR) and provider order entry program to identify hospitalized patients at risk for VTE using a weighted risk score and for whom prophylaxis was not ordered. Patients were randomized to the intervention group, in which the responsible physician received an electronic alert regarding the risk of VTE and recommendation regarding prophylaxis, or to the control group, in which no alert was issued. Compared with the control group, electronic alerts more than doubled the rate of thromboprophylaxis orders (from 14.5% to 33.5%; p<0.0001). The risk of symptomatic VTE was reduced by 41% (hazard ratio, 0.59; 95% confidence interval 0.43-0.81; p=0.001) among patients for whom an electronic alert was issued. Investigators have also shown that "human" (person-to-person) alert-based decision support for the prevention of VTE in at-risk hospitalized patients more than doubled the rate of VTE prophylaxis compared with controls during hospitalization (21) and after discharge (19).

The current study will determine the impact of electronic alert-based CDS on prescription of anticoagulation in high-risk AF patients in the outpatient setting who are not being prescribed anticoagulation for stroke prevention. Because this is a Quality Improvement initiative, investigators will not mandate a specific antithrombotic agent, regimen, or duration. Investigators will provide options for anticoagulation to prevent stroke in AF and allow the provider to make the best choice based on their clinical judgment. If there is a contraindication to anticoagulation or if the risks outweigh the benefits of antithrombotic therapy, the provider can elect to omit anticoagulation but will need to provide the rationale for doing so. The current study will also determine the impact of electronic alert-based CDS on the frequency of stroke and systemic embolic events in high-risk AF patients in the outpatient setting who are not being prescribed anticoagulation for stroke prevention. Data acquired through this study regarding the frequency of stroke and systemic embolism will be used to calculate sample size requirements for a future clinical end-point driven randomized controlled trial of electronic alerts to prevent stroke in high-risk AF patients.

Inclusion Criteria:

• Men and women
• Age ≥ 18 years
• Problem list entry of AF or atrial flutter
• CHA2DS2VASc score ≥2

Exclusion Criteria:

• Active prescription for anticoagulant therapy
• CHA2DS2VASc score <2