Javier A Valle1, Zhuokai Li2, Andrzej S Kosinski2, Adam J Nelson2, Sreekanth Vemulapalli3, Joseph Cleveland4, David Fullerton4, John C Messenger4, Jessica Y Rove4, Rory S Bricker4, Steven M Bradley5, Frederick A Masoudi4, Robert W Yeh6, Ehrin J Armstrong7, Stephen W Waldo7, John D Carroll4. 1. Department of Medicine, Rocky Mountain Regional Veterans Affairs Medical Center, Denver, Colorado, USA; University of Colorado School of Medicine, Aurora, Colorado, USA; Michigan Heart and Vascular Institute, Ann Arbor, Michigan, USA. Electronic address: Javier.valle@cuanschutz.edu. 2. Duke Clinical Research Institute, Durham, North Carolina, USA. 3. Duke Clinical Research Institute, Durham, North Carolina, USA; Duke University School of Medicine, Durham, North Carolina, USA. 4. University of Colorado School of Medicine, Aurora, Colorado, USA. 5. Minneapolis Heart Institute, Minneapolis, Minnesota, USA. 6. Richard A. and Susan F. Smith Center for Outcomes Research in Cardiology, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA. 7. Department of Medicine, Rocky Mountain Regional Veterans Affairs Medical Center, Denver, Colorado, USA; University of Colorado School of Medicine, Aurora, Colorado, USA.
Abstract
BACKGROUND: Societal guidelines and payor coverage decisions for transcatheter aortic valve replacement (TAVR) attempt to strike a balance between providing access and maintaining quality. The extent to which dissemination of TAVR has achieved these ideals remains unknown. OBJECTIVES: This study sought to define patterns of TAVR dissemination in the United States and their influence on outcomes. METHODS: Using data from the TVT (Transcatheter Valvular Therapy) registry, this study identified TAVR sites from 2011 to 2018 and calculated drive-times from existing to new sites. In a contemporary cohort, this study compared site and patient characteristics by annual case volume and density of sites per million Medicare beneficiaries. Using hierarchical regression and Cox methods, this study determined the association between case volumes, site density, and changes in volume and density with patient risk profiles and outcomes. RESULTS: TAVR sites participating in the TVT registry increased from 198 to 556 from 2011 to 2018. Median drive-time from existing to new sites decreased from 403 minutes (interquartile range: 211-587 minutes) to 26 minutes (interquartile range: 17-48 minutes). In a contemporary cohort, higher site density was associated with lower procedural risk as well as with an increased hazard of 30-day risk-adjusted mortality (P = 0.017). Similarly, longitudinal increases in site density over time were associated with a higher hazard of 30-day (P = 0.011) and 1-year (P = 0.013) mortality. CONCLUSIONS: TAVR has expanded significantly over time, but with regional clustering of sites. Although procedural risk is lower at higher density sites, these sites demonstrate an increased hazard of mortality. These findings suggest that the expansion of TAVR services in the United States may have had unintended consequences on procedural quality.
BACKGROUND: Societal guidelines and payor coverage decisions for transcatheter aortic valve replacement (TAVR) attempt to strike a balance between providing access and maintaining quality. The extent to which dissemination of TAVR has achieved these ideals remains unknown. OBJECTIVES: This study sought to define patterns of TAVR dissemination in the United States and their influence on outcomes. METHODS: Using data from the TVT (Transcatheter Valvular Therapy) registry, this study identified TAVR sites from 2011 to 2018 and calculated drive-times from existing to new sites. In a contemporary cohort, this study compared site and patient characteristics by annual case volume and density of sites per million Medicare beneficiaries. Using hierarchical regression and Cox methods, this study determined the association between case volumes, site density, and changes in volume and density with patient risk profiles and outcomes. RESULTS: TAVR sites participating in the TVT registry increased from 198 to 556 from 2011 to 2018. Median drive-time from existing to new sites decreased from 403 minutes (interquartile range: 211-587 minutes) to 26 minutes (interquartile range: 17-48 minutes). In a contemporary cohort, higher site density was associated with lower procedural risk as well as with an increased hazard of 30-day risk-adjusted mortality (P = 0.017). Similarly, longitudinal increases in site density over time were associated with a higher hazard of 30-day (P = 0.011) and 1-year (P = 0.013) mortality. CONCLUSIONS: TAVR has expanded significantly over time, but with regional clustering of sites. Although procedural risk is lower at higher density sites, these sites demonstrate an increased hazard of mortality. These findings suggest that the expansion of TAVR services in the United States may have had unintended consequences on procedural quality.