Modeling the Association of Volume vs Composite Outcome Thresholds With Outcomes and Access to Transcatheter Aortic Valve Implantation in the US.

IMPORTANCE: Professional societies and the Centers for Medicare & Medicaid Services suggest volume thresholds to ensure quality in transcatheter aortic valve implantation (TAVI). OBJECTIVE: To model the association of volume thresholds vs spoke-and-hub implementation of outcome thresholds with TAVI outcomes and geographic access. DESIGN, SETTING, AND PARTICIPANTS: This cohort study included patients who enrolled in the US Society of Thoracic Surgeons/American College of Cardiology Transcatheter Valve Therapy registry. Site volume and outcomes were determined from a baseline cohort of adults undergoing TAVI between July 1, 2017, and June 30, 2020. EXPOSURES: Within each hospital referral region, TAVI sites were categorized by volume (<50 or ≥50 TAVIs per year) and separately by risk-adjusted outcome on the Society of Thoracic Surgeons/American College of Cardiology Transcatheter Valve Therapy 30-day TAVI composite during the baseline period (July 2017 to June 2020). Outcomes of patients undergoing TAVIs from July 1, 2020, to March 31, 2022, were then modeled as though the patients had been treated at (1) the nearest higher volume (≥50 TAVIs per year) or (2) the best outcome site within the hospital referral region. MAIN OUTCOMES AND MEASURES: The primary outcome was the absolute difference in events between the adjusted observed and modeled 30-day composite of death, stroke, major bleeding, stage III acute kidney injury, and paravalvular leak. Data are presented as the number of events reduced under the above scenarios with 95% bayesian credible intervals (CrIs) and median (IQR) driving distance. RESULTS: The overall cohort included 166 248 patients with a mean (SD) age of 79.5 (8.6) years; 74 699 (47.3%) were female and 6657 (4.2%) were Black; 158 025 (95%) were treated in higher-volume sites (≥50 TAVIs) and 75 088 (45%) were treated in best-outcome sites. Modeling a volume threshold, there was no significant reduction in estimated adverse events (-34; 95% CrI, -75 to 8), while the median (IQR) driving time from the existing site to the alternate site was 22 (15-66) minutes. Transitioning care to the best outcome site in a hospital referral region resulted in an estimated 1261 fewer adverse outcomes (95% CrI, 1013-1500), while the median (IQR) driving time from the original site to the best site was 23 (15-41) minutes. Directionally similar findings were observed for Black individuals, Hispanic individuals, and individuals from rural areas. CONCLUSIONS AND RELEVANCE: In this study, compared with the current system of care, a modeled outcome-based spoke-and-hub paradigm of TAVI care improved national outcomes to a greater extent than a simulated volume threshold, at the cost of increased driving time. To improve quality while maintaining geographic access, efforts should focus on reducing site variation in outcomes.

Duke Scholars

Author Sreekanth Vemulapalli Medicine, Cardiology

Author Andrzej Stanislaw Kosinski Biostatistics & Bioinformatics, Division of Biostatistics

Author Sean Michael O'Brien Biostatistics & Bioinformatics, Division of Biostatistics