PERSEVERE-PEF: optimizing medical therapy saves lives in heart failure with preserved ejection fraction

Need. In the US, heart failure (HF) is the contributing cause of 1 in 8 deaths. HF with preserved ejection
fraction (HFpEF) affects close to 50% of all HF patients. The 5-years mortality is 35%. HFpEF multi-organ
syndrome clinical care management is complex and time consuming. In addressing HFpEF, the American
College of Cardiology guidelines directed medical therapy (GDMT) references the medical therapy decision to
the individual disease guidelines [hypertension (HTN), coronary artery disease (CAD), atrial fibrillation (AFib)].
Providing concerted multi-disease HFpEF management is a major unmet clinical need.
Solution. In response to this need, we (OPTIMA) developed and demonstrated the feasibility of a clinical
analytic intelligence (AI) for the management of HFpEF multi-organ syndrome, optima4PEF AI. The solution
adds significant value to OPTIMA’s HF Management Service currently addressing the GDMT management of
HTN (optima4BP AI) and HFrEF (optima4heart AI). optima4PEF deconstructs a complex set of disease-
specific clinical guidelines and re-assembles them into a concerted multi-disease GDMT that is patient-
personalized, explainable, and actionable.
Objectives. PERSEVERE-PEF [optimizing medical therapy saves lives in heart failure with preserved ejection
fraction] proposes to complete the AI development of optima4PEF product concept, and to validate its efficacy
using contemporary, diverse, retrospective patient cohorts.Aim 1. Build optima4PEF AI to address the GDMT management of HFpEF multi-organ syndrome.
Hypothesis. optima4PEF deconstructs a complex set of disease-specific clinical guidelines and re-assembles
them into a concerted multi-disease GDMT that is patient-personalized, explainable, and actionable.
The product concept work built the optima4PEF AI system architecture and developed the decision logic to
address GDMT management for patients experiencing HFpEF + volume overload + HTN. optima4PEF product
concept will be extended to include GDMT management of AFib and of CAD. An end-to-end algorithm
regression test will be performed to verify that each decision logic step performs its intended function.Aim 2. Validate optima4PEF AI in recommending the most relevant GDMT. Hypothesis. In ≥ 90% of
patient cases, optima4PEF case-specific treatment recommendation is ACCEPTED as the appropriate next
step in the process of multi-disease GDMT treatment optimization of patients diagnosed with HFpEF.
Unidentified patient records will be collected from 4 clinical partner sites. A randomization algorithm will select
n=840 patient records. optima4PEF will generate a Treatment Action (TA) for each patient record. A simple
majority rule of pharmacology and cardiology experts (n=5) will adjudicate the optima4PEF TA.
optima4PEF averts loss of lives by assisting in the delivery of HFpEF multi-disease management.
optima4PEF surveillance and personalized care support the emerging digital-first clinical care practices.