Hope: advancing hfpef management through evidence-based treatment optimization and outcome prediction

The National Heart, Lung, and Blood Institute awarded Mayo Clinic $801,170 on August 1, 2026, under the Cardiovascular Diseases Research program (CFDA 93.837) to conduct HOPE: Advancing HFPEF Management Through Evidence-Based Treatment Optimization and Outcome Prediction, a project grant focused on heart failure with preserved ejection fraction.

Mayo Clinic will develop therapeutic discovery and biomarker prediction models for HFPEF patients through two integrated research strategies. First, the recipient will conduct drug screening to identify therapeutic benefits from existing guideline-directed medical therapy and non-GDMT drug combinations for HFPEF, combining drug-target predictions with EHR-based emulated trials to assess changes in mortality and morbidity endpoints. Second, Mayo Clinic will develop biomarker response prediction models using temporal coherence and intra-visit relationships within electronic health record data, learning stratified phenogroups and building personalized biomarker response prediction models for each group. The recipient plans to validate models and outputs through collaboration with Florida State University using the One-Florida+ EHR network via PCORnet, comparing results against Mayo Clinic EHR data.

Place of performance is Rochester, Minnesota. The period of performance extends from August 1, 2026, through April 30, 2031.

Abstract

Heart failure with preserved ejection fraction (HFPEF), a common subtype of heart failure (HF), presents a unique clinical challenge. Although HFPEF is characterized by a normal ejection fraction (EF), patients still experience HF symptoms. This paradox results from normal pumping efficiency being offset by impaired filling capacity, which complicates treatment in clinical settings. Limited strong clinical evidence further constrains treatment options, with most therapies rated only as class IIA or IIB recommendations by AHA/ACC/hfsa guidelines. Adding to this challenge is HFPEF's complex and often elusive etiology, involving interrelated comorbidities such as hypertension and diabetes. Managing these comorbidities is particularly difficult due to limited guidelines that do not fully address the complexity of patients. To tackle these issues, we propose hope, a comprehensive approach to advance HFPEF management through two innovative strategies that evaluate the benefit of the existing guideline-directed medical therapy (GDMT) drugs and non-GDMT drugs. First, we aim to discover therapeutic benefits from existing GDMT and non-GDMT combinations for HFPEF patients. This will be achieved through a drug screening model that assesses changes in endpoint (e.g., mortality and morbidity) by combining the results of drug-target predictions with EHR-based emulated trials. Second, we will develop biomarkers response prediction models that use temporal coherence and intra-visit relationships within EHR data. We plan to learn to learn stratified pheno-groups and develop a personalized biomarkers response prediction model for each group. To validate our models and their potential benefits, we plan to collaborate with florida state university (FSU) by using the one-florida+ EHR network via pcornet, comparing outputs derived from mayo clinic EHR. Hope applies standardized methodologies within the pcornet common data model, ensuring consistent data extraction and processing to support widespread application and validation. The scope of hope is substantial, encompassing analysis of over 390,000 HFPEF patients. With successful implementation, hope is poised to significantly enhance patient care, offering tailored treatment plans that meet individual patient needs and setting a new standard in HFPEF treatment.