AHA Scientific Sessions 2026 Abstract · Nov 2, 2026
Same Patients, Different Cardiovascular-Kidney-Metabolic Stages: Comparing Claims and Electronic Health Record-Based Classification
By Cusick M, Kong S, Weiner J et al.
Introduction: Limited research has operationalized cardiovascular-kidney-metabolic (CKM) syndrome staging using real-world administrative claims and electronic health record (EHR) data, a critical step for population health surveillance, risk stratification, and health services research among adults with CKM syndrome. Research Question: How does CKM syndrome stage classification differ across administrative claims, EHR, and linked claims-EHR data sources? Methods: We developed a novel CKM syndrome staging algorithm for real-world data sources (Figure 1). We applied the staging algorithm to a linked claims-EHR database of adults aged 18 years and older who were enrolled in one of three Johns Hopkins Health Plans (Employee Health Plan, Medicaid Managed Care, and Medicare Advantage) and received care within the Johns Hopkins Health System in 2019. We assigned CKM syndrome stages using three data sources: claims only, EHR only, and linked claims-EHR data. We examined each CKM syndrome stage (0-4), as well as advanced CKM syndrome (stages 3 or 4). We estimated prevalence overall and by age group, sex, race, ethnicity, and insurance type, and evaluated agreement in CKM stage classification across data sources. Results: Our study population included 72,870 adults, among whom the mean age was 44 years (SD: 17.1); 65.4% were women, 41.5% were Black or African American, 43.5% were White, and 4.3% were Hispanic (Table 1).The estimated prevalence of CKM syndrome differed across data sources. Advanced CKM syndrome affected 11,599 (15.9%) adults when using claims alone and 12,000 (16.5%) adults when using EHR data alone (Table 1). Prevalence was highest using linked claims-EHR data, with 14,603 (20.0%) adults classified as having advanced CKM syndrome. Across all three data sources, the prevalence of advanced CKM syndrome was higher among adults aged 65 years or older, men, and Medicare Advantage enrollees. Agreement in stage-specific classification between EHR and claims data sources varied across CKM syndrome stages. Concordance was highest for CKM stage 4, where 84.5% of patients classified as stage 4 using EHR data were also classified as stage 4 using claims data (Table 2). Concordance was lowest for CKM stages 1 (20.2%) and 3 (30.1%), stages that depend more heavily on laboratory, biomarker, and flowsheet data available within the EHR. Concordance between EHR-only and linked claims-EHR classifications was at least 89.5% across all CKM stages. Adding claims information to EHR data resulted in upward reclassification, most commonly by one or two CKM stages. Conclusions: CKM syndrome staging can be operationalized using real-world administrative claims and EHR data. However, CKM stage classification may vary meaningfully across data sources, highlighting the need for careful validation before implementation in health system or payer settings.