CLAIMS-BASED ALGORITHM OPTIONS TO IDENTIFY REDUCED LEFT VENTRICULAR EJECTION FRACTION PATIENTS WITH MITRAL OR TRICUSPID VALVE REGURGITATION AND HEART FAILURE
Author(s)
Rebecca A. Horn, PhD1, Lisa S. Kemp, PhD2, Sarah Mollenkopf, BS, MPH3;
1Edwards Lifesciences, Manager, Health Economics & Reimbursement, Wichita, KS, USA, 2Edwards Lifesciences, Irvine, CA, USA, 3Edwards Lifesciences, Newport Beach, CA, USA
1Edwards Lifesciences, Manager, Health Economics & Reimbursement, Wichita, KS, USA, 2Edwards Lifesciences, Irvine, CA, USA, 3Edwards Lifesciences, Newport Beach, CA, USA
OBJECTIVES: Heart failure (HF) is common in patients with valvular heart disease. It is clinically important to categorize HF by preserved or reduced ejection fraction (HFpEF or HFrEF), but these categories do not have billing codes for use in claims-based database research. Published claims-based models generally perform well for identifying HFpEF patients but not HFrEF. This study compares the proportion of reduced ejection fraction (rEF) capture across three simple, claims-based cohort identification algorithms in patients with mitral or tricuspid regurgitation (MR or TR).
METHODS: HF patients in Optum Market Clarity (2007-2025Q2) with either MR or TR and a recorded left ventricular ejection fraction (LVEF) were identified (N=139,120). Three cohort identification algorithm options were then compared to each other on the proportion of patients with rEF (defined as LVEF ≤40%) identified within three months after a systolic HF diagnostic code. Algorithm option 1 included systolic HF and cardiomyopathy diagnoses. Option 2 included option 1 diagnoses plus a history of at least one HF medication. Option 3 included option 1 diagnoses plus a history of pacemaker or cardiac catheterization.
RESULTS: Of patients with systolic HF (n=27,239MR, 15,107TR), 67%MR and 63%TR had rEF. Proportion of identified rEF patients increased with each algorithm option: Option 1 (76%MR, 72%TR of n=12,956MR, 7,474TR), Option 2 (77%MR, 73%TR of n=12,608MR, 14,438 TR), and Option 3 (80%MR, 76%TR of n=5,865MR, 3,630TR).
CONCLUSIONS: A claims-based cohort identification algorithm that included systolic HF and cardiomyopathy diagnostic codes and pacemaker or cardiac catheterization history captured the highest proportion of patients with recorded rEF. The utility of this simplified algorithm as a proxy when ejection fraction values are unavailable should be explored further.
METHODS: HF patients in Optum Market Clarity (2007-2025Q2) with either MR or TR and a recorded left ventricular ejection fraction (LVEF) were identified (N=139,120). Three cohort identification algorithm options were then compared to each other on the proportion of patients with rEF (defined as LVEF ≤40%) identified within three months after a systolic HF diagnostic code. Algorithm option 1 included systolic HF and cardiomyopathy diagnoses. Option 2 included option 1 diagnoses plus a history of at least one HF medication. Option 3 included option 1 diagnoses plus a history of pacemaker or cardiac catheterization.
RESULTS: Of patients with systolic HF (n=27,239MR, 15,107TR), 67%MR and 63%TR had rEF. Proportion of identified rEF patients increased with each algorithm option: Option 1 (76%MR, 72%TR of n=12,956MR, 7,474TR), Option 2 (77%MR, 73%TR of n=12,608MR, 14,438 TR), and Option 3 (80%MR, 76%TR of n=5,865MR, 3,630TR).
CONCLUSIONS: A claims-based cohort identification algorithm that included systolic HF and cardiomyopathy diagnostic codes and pacemaker or cardiac catheterization history captured the highest proportion of patients with recorded rEF. The utility of this simplified algorithm as a proxy when ejection fraction values are unavailable should be explored further.
Conference/Value in Health Info
2026-05, ISPOR 2026, Philadelphia, PA, USA
Value in Health, Volume 29, Issue S6
Code
MSR201
Topic
Methodological & Statistical Research
Disease
SDC: Cardiovascular Disorders (including MI, Stroke, Circulatory)