A cluster randomised study in the primary care setting to evaluate a electronic clinical decision tool for stroke prophylaxis in patients with atrial fibrillation.

Atrial fibrillation is the most common form of arrhythmia, affecting more than three percent of the population. The condition carries an increased risk of thromboembolism, in particular stroke. In 2013 approximately 25000 acute strokes were diagnosed in Sweden, with the result of death or severe disability in nearly half of the cases. Numerous studies have shown that the risk for stroke can be reduced by approximately 60-70 % with the use of anticoagulant therapy to patients with one or several risk factors for stroke and concurrent atrial fibrillation. The European Society of Cardiology recommends use of the CHA2DS2VASc algorithm to identify persons at increased risk for stroke in the setting of atrial fibrillation. Prophylaxis is recommended if CHA2DS2VASc ≥ 1 (not only female gender).

The agents used include warfarin and the more recently developed non-vitamin k oral anticoagulants (NOACs). Despite good evidence and recommendations in current guidelines, however, there remains a substantial undertreatment in this group of patients.

The Swedish Association of Local Authorities and Regions have shown that only 63% of patients with a CHA2DS2VASc score ≥ 2 are prescribed anticoagulant therapy [7]. The reasons for this is most likely multifactorial, including ignorance of the CHA2DS2VASc algorithm as well as reluctance of the use of potent drugs from both doctors and patients. Furthermore, the high pace in modern medicine increase the risk of missing the diagnosis of atrial fibrillation and/or the conditions constituting the CHA2DS2VASc algorithm.

Clinical decision tools is a relatively new phenomena in modern medicine showing promising results, but evidence for clinical outcome are sparse. The clinical decision tool for stroke prevention (CDSS) has been developed in collaboration between Cambio Cosmic (the supplier of the electronic journal in the county of Östergötland), the Cardiology Department at Linköping University hospital and primary care professionals. The decision tool is activated when a patient is being logged into the electronic journal. If the patient has a diagnosis of atrial fibrillation (or atrial flutter) and a CHA2DS2VASc score ≥ 1 (not only female gender) without current anticoagulant therapy, a screen warning will appear. By clicking on the warning, the responsible physician will get an overview of the patient's diagnosis according to the CHA2DS2VASc algorithm. Furthermore, a calculation of the estimated stroke risk for the coming year will appear, and links to national guidelines be provided. The physician can thereafter decide to prescribe anticoagulant therapy in accordance with current guidelines or, alternatively, postpone the decision/make a decision to refrain from medication. In case the choice is made to refrain from medication the physician is asked to choose between a set of predetermined reasons in order to monitor the main reasons for deviation from guidelines.

If, on the other hand, the patient has anticoagulant therapy in accordance with current guidelines, no screen warning will appear.

CDSS has shown promising results in a pilot study conducted at five units in the County of Östergötland during the fall of 2014. The aim of the present study is to investigate this computerized decision tool in a large randomized trial in the primary care setting.

Study design The present study is a cluster randomized study in the primary care setting in the County of Östergötland, Sweden. The investigators intend to include all primary care units (n = 43) in the County of Östergötland. Participation is non-compulsory. The population in the County of Östergötland is 442 105 (December 2014)

At the time of inclusion all primary care units will be stratified in four strata based on the number of patients listed on each unit and current adherence to guidelines (i.e. the percentage of patients with a diagnosis of atrial fibrillation and CHA2DS2VASc ≥ 1 (not only female gender) that are currently prescribed anticoagulant therapy). They will thereafter be randomized to intervention with the CDSS application or serve as a control unit (continue with usual care, randomized 1:1). Prior to randomization all the participating general practitioners will receive an education about atrial fibrillation and the associated risk of stroke, including the CHA2DS2VASc algorithm and an overview of anticoagulant therapy. Furthermore, all units receiving the CDSS will have a briefing about the technical aspects of using the application.

Inclusion Criteria:

• Primary care centers in the county of Östergötland, Sweden.

Exclusion Criteria:

• None