Abstract 16521: Using Electronic Health Record Data to Identify and Prioritize Patients With Heart Failure With Reduced Ejection Fraction for a Remote Medication Optimization Program

Abstract

Introduction: Remote medication management programs (RMMP) might be effective in optimizing guideline directed medical therapy (GDMT) for HFrEF. Identification of patients that can be safely managed remotely and prioritization of those with the highest need for medication optimization is crucial, especially in the setting of constrained resources. Methods: We analyzed records from our institutional enterprise data warehouse to identify patients with HFrEF in whom GDMT can be safely optimized by an RMMP within the Mass General Brigham health system between May 2021 and May 2023. After identifying a population with HFrEF, we excluded patients with contraindications to GDMT and with higher risk of adverse medication reactions. Next, we identified active and past medications by using medication discontinuation data (Figure). We performed chart review on a random sample of 100 patients to validate our analysis. Results: We have identified 1771 patients with HFrEF who are actively receiving care and are eligible for remote medication optimization. Only 7% (129/1771) of patients were prescribed all four GDMT, however in 4% (78/1771) at least one of these prescriptions was discontinued. Of all patients, 45% (801/1771) had stopped at least one GDMT, including 50% of SGLT2i, 30% of MRA, and 46% of ARNI. Patients who were never prescribed GDMT were identified as top priority for the RMMP. Conclusion: EHR data can effectively identify and prioritize HFrEF patients for medication optimization. Similar to prior HFrEF registries, our study showed poor utilization of GDMT. However, it also revealed a significant number of patients who had their GDMT prescriptions discontinued. Investigating the reasons for discontinuations may elucidate barriers to GDMT optimization. This study can be expanded to patients with HFpEF and HFmEF, and refined further by using natural language processing, machine learning, and by incorporating data for medication refill, claims, and dosages.