Clinical outcomes and healthcare expenditures in the real world with left ventricular assist devices - The CLEAR-LVAD study
Francis D Pagani,Mandeep R Mehra,Jennifer A Cowger,Douglas A Horstmanshof,Scott C Silvestry,Pavan Atluri,Joseph C Cleveland Jr,JoAnn Lindenfeld,Gregory J Roberts,Rupinder Bharmi,Nirav Dalal,Robert L Kormos,Joseph G Rogers
DOI: https://doi.org/10.1016/j.healun.2021.02.010
Abstract:Background: Several distinctly engineered left ventricular assist devices (LVADs) are in clinical use. However, contemporaneous real world comparisons have not been conducted, and clinical trials were not powered to evaluate differential survival outcomes across devices. Objectives: Determine real world survival outcomes and healthcare expenditures for commercially available durable LVADs. Methods: Using a retrospective observational cohort design, Medicare claims files were linked to manufacturer device registration data to identify de-novo, durable LVAD implants performed between January 2014 and December 2018, with follow-up through December 2019. Survival outcomes were compared using a Cox proportional hazards model stratified by LVAD type and validated using propensity score matching. Healthcare resource utilization was analyzed across device types by using nonparametric bootstrap analysis methodology. Primary outcome was survival at 1-year and total Part A Medicare payments. Results: A total of 4,195 de-novo LVAD implants were identified in fee-for-service Medicare beneficiaries (821 HeartMate 3; 1,840 HeartMate II; and 1,534 Other-VADs). The adjusted hazard ratio for mortality at 1-year (confirmed in a propensity score matched analysis) for the HeartMate 3 vs HeartMate II was 0.64 (95% CI; 0.52-0.79, p< 0.001) and for the HeartMate 3 vs Other-VADs was 0.51 (95% CI; 0.42-0.63, p < 0.001). The HeartMate 3 cohort experienced fewer hospitalizations per patient-year vs Other-VADs (respectively, 2.8 vs 3.2 EPPY hospitalizations, p < 0.01) and 6.1 fewer hospital days on average (respectively, 25.2 vs 31.3 days, p < 0.01). The difference in Medicare expenditures, conditional on survival, for HeartMate 3 vs HeartMate II was -$10,722, p < 0.001 (17.4% reduction) and for HeartMate 3 vs Other-VADs was -$17,947, p < 0.001 (26.1% reduction). Conclusions: In this analysis of a large, real world, United States. administrative dataset of durable LVADs, we observed that the HeartMate 3 had superior survival, reduced healthcare resource use, and lower healthcare expenditure compared to other contemporary commercially available LVADs.