Abstract P269: Linkage of Patients in a National Heart Failure Clinical Registry to Medicare Claims Data: An Analysis of Generalizability and Longitudinal Outcomes from ADHERE

Abstract

Background: Registries are increasingly being used to analyze quality, outcomes, and clinical effectiveness in HF patients. The degree to which registry patients reflect the population of real-world HF patients is unclear. Methods: We linked data from the Acute Decompensated Heart Failure National Registry (ADHERE) to Medicare claims. We compared patient characteristics and inpatient mortality of ADHERE hospitalizations linked to Medicare, with ADHERE hospitalizations not linked to Medicare. We also compared patient characteristics, 1-year readmission, inpatient mortality, and 1-year mortality of linked ADHERE patients to a 20% sample of Medicare patients hospitalized for HF. Hospital characteristics among Medicare sites participating and not participating in ADHERE were also compared. Results: Among 135,667 ADHERE hospitalizations for HF patients ≥ 65 years, we linked 104,808 (77.3%) to Medicare claims, representing 82,074 unique patients. Within ADHERE, Medicare-linked hospitalizations were more likely to involve patients who were female and white, but there were no clinically relevant differences in other patient characteristics. In-hospital mortality was identical for ADHERE hospitalizations linked and not linked to Medicare (4.5%). ADHERE hospitals were more likely than non-ADHERE hospitals to be teaching hospitals, to have a higher volume of HF patients, and to offer advanced cardiac care. Within Medicare, ADHERE patients and non-ADHERE patients had similar demographics and comorbidities. ADHERE patients had slightly lower observed mortality and readmission rates (Table). After risk adjustment, modest but statistically significant differences remained. Conclusion: There is a high degree of similarity between ADHERE patients and Medicare patients hospitalized with HF. While there are differences in hospital characteristics, ADHERE data are representative of the real-world elderly HF population. Mortality/readmission outcomes for ADHERE and non-ADHERE Medicare patients hospitalized with HF Outcomes variable Patients within Medicare Non-ADHERE (N=571,649) ADHERE (N=79,508) 1 P-value 2 Adjusted OR/HR (95% CI) 3 Mortality, In-hospital 27,779 (4.9%) 3,516 (4.4%) <.001 0.94 (0.90-0.97) Mortality, 1-year 218,703 (38.3%) 28,579 (36.0%) <.001 0.95 (0.94, 0.96) Readmission, 1-year all-cause 4 349,990 (67.9%) 48,802 (65.8%) <.001 0.94 (0.93, 0.94) Readmission, 1-year cardiovascular 3 223,430 (43.4%) 31,553 (42.6%) <.001 0.96 (0.95, 0.97) 1 Total number of unique, linked ADHERE patients with 1 year prior claims eligible for analysis 2 P-values based on chi-square tests stratified by discharge year 3 In-hospital mortality results shown as odds ratios (OR); all other results shown as hazard ratios (HR) 4 Among patients eligible for readmission analysis (N=518,548 Non-ADHERE patients; N=74,450 ADHERE patients)