Evaluation of an EMR-Based Clinical Decision Support Tool for the Implementation of Guideline-Directed Therapies for Prevention of Heart Failure Among High-Risk Patients With Type 2 Diabetes

Type 2 diabetes mellitus (T2DM) is an independent risk factor for heart failure (HF) and is associated with significant morbidity and mortality. Even despite adequate glycemic control, individuals with T2DM face considerable risk of HF even in individuals without other significant risk factors. Moreover, individuals with both atherosclerotic cardiovascular disease and T2DM face up to a five-fold increased risk of HF and experience higher rates of mortality compared to age-matched controls. Thankfully, recent therapeutic advances in pharmacotherapies, such as sodium-glucose cotransporter-2 inhibitors (SGLT2i), have shown to be beneficial in preventing HF among patients with T2DM. Current guidelines by the American Diabetes Association and the joint American College of Cardiology/American Heart Association (ACC/AHA) both provide class I/A recommendations in initiating SGLT2i medication in individuals with T2DM and cardiovascular comorbidities for prevention of HF. Similarly, the Food and Drug Administration now indicates SGLT2i as a method to reduce the risk of HF hospitalization in adults with T2DM and established CV risk factors.

Unfortunately, SGLT2i are underused in patients with T2DM at risk for HF with ~5% of eligible patients treated with the medication. Risk-based approaches to identify patients who are at increased risk of developing adverse events is key to improve the use of evidence-based therapies and for efficient and cost-effective allocation of preventive strategies. Previous methods, such as the Pooled Cohort Equation, have been effective in guiding prescription of statin medications to at-risk patients. Similarly, alert-based clinical decision support tools have been used to help guide anticoagulation strategies in patients with atrial fibrillation. However, no such risk-based approach exists for implementation of goal-directed medical therapy for HF prevention in patients with T2DM.

The WATCH-DM risk score (Weight [body mass index], Age, hyperTension, Creatinine, HDL-C, Diabetes control [fasting plasma glucose] and QRS Duration, MI and CABG) is one such machine learning-based tool that was developed among participants of the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial.

The investigators used machine-learning methods and readily available clinical characteristics to derive the risk prediction model and has had excellent discrimination and calibration for estimating HF risk. For each risk factor level, patients are given a specific number of points. The sum of the points accounting for all risk factors included in the model is associated with 5-year risk of HF. There is a graded, dose-response relationship between the WATCH-DM risk score and risk of HF. For example, patients who had a WATCH-DM risk score of at least 11 had a 5-year risk of incident HF ≥9.2%.

This proposed trial will test the efficacy of a computer-based electronic alert (clinical decision support) notifying the provider that the patient is at an increased risk of developing heart failure. There currently are no developed or implemented alert systems notifying the provider that the patient is at an increased risk of heart failure. Similarly, there is no risk-based approach to implementation evidence-based T2DM therapies in patients at risk for HF. Currently, SGLT2i use is underutilized with ~5% of eligible patients current prescribed the medication. Clinical decision support tools may inform providers about a patient's risk of HF and may be useful to improve the use of SGLT2i therapies. Previous implementation strategies have been useful to guide statin medications in patients at risk for atherosclerotic cardiovascular events and anticoagulation strategies in patients with atrial fibrillation.