Geographic access to transcatheter aortic valve replacement relative to other invasive cardiac services: A statewide analysis.

BACKGROUND: Transcatheter aortic valve replacement (TAVR) received US regulatory approval for treatment of severe symptomatic aortic stenosis (AS) in November 2011. After subsequent approvals for expanded indications, it is now performed throughout Michigan but the distribution of these providers and their impact on access is uncertain. As the number of providers and utilization for TAVR grows, how procedural volume is distributed among providers may significantly impact patient outcomes. METHODS: We determined geographic access to TAVR in Michigan as of October 2014, and compared it to access of other invasive cardiac services; namely, percutaneous coronary intervention (PCI), non-transplant cardiac surgery, and cardiac transplant surgery. A geographic information systems analysis was performed using recent U.S. Census Survey data and statewide inpatient data to construct maps of service areas around hospitals providing TAVR, PCI, non-transplant cardiac surgery, and cardiac transplant surgery. Service areas ranging across multiple driving distances were included in the analysis. Geographic access was calculated as percentage of the population living within the hospital service areas providing invasive cardiac services. RESULTS: In October 2014, 15 hospitals provide TAVR in Michigan. For TAVR sites, the mean number of beds, annual discharges, and annual patient days are 571, 28,946, and 140,859, respectively. Compared to hospitals not offering TAVR, TAVR facilities were more likely to be non-profit (86.7% vs 71.0%), a teaching hospital (93.3% vs 87.1%), and rural (12.1% vs 6.5%). Of the 9,883,640 persons in Michigan, 4,492,941 (45.5%) live within 10 miles, 7,856,455 (79.5%) live within 30 miles, and 9,004,943 (91.1%) live within 50 miles driving distance of TAVR sites. These proportions compare favorably with hospitals providing PCI (8,857,148 [89.6%] living within 30 miles) and non-transplant cardiac surgery (8,814,143 [89.2%] living within 30 miles) as opposed to cardiac transplant surgery (5,481,122 [55.5%] living within 30 miles). For Michigan patients who underwent surgical valve replacement (SAVR) in 2010-2011, the median driving distance to a TAVR site was under 15 miles and under 10 miles to a hospital providing non-transplant cardiac surgery. CONCLUSIONS: Nearly 4 of 5 Michigan residents lived within 30 miles of TAVR services early after its approval, suggesting its wide availability despite initial regulations on its use. These findings may encourage growth in TAVR utilization and limit the development of expertise as procedural volume is distributed among more providers. Given procedural volume tends to relate positively with outcomes, increased access to TAVR may have negative effects on patient outcomes.

Transfer rates from nonprocedure hospitals after initial admission and outcomes among elderly patients with acute myocardial infarction.

IMPORTANCE: It is unknown whether hospital transfer rates for patients with acute myocardial infarction admitted to nonprocedure hospitals (facilities that do not provide catheterization) vary and whether these rates further influence revascularization rates, length of stay, and mortality. OBJECTIVES: To examine hospital differences in transfer rates for elderly patients with acute myocardial infarction across nonprocedure hospitals and to determine whether these rates are associated with revascularization rates, length of stay, and mortality. DESIGN, SETTING, AND PARTICIPANTS: We used Medicare claims data from January 1, 2006, to December 31, 2008, to assess transfer rates in nonprocedure hospitals, stratified according to transfer rates as low (≤ 20%), mid-low (>20%-30%), mid-high (>30%-40%), or high (>40%). Data were analyzed for 55,962 Medicare fee-for-service patients admitted to 901 nonprocedure US hospitals with more than 25 admissions per year for acute myocardial infarction. MAIN OUTCOMES AND MEASURES: We compared rates of catheterization, percutaneous coronary intervention, and coronary artery bypass graft surgery during hospitalization and within 60 days, as well as hospital total length of stay, across groups. We measured risk-standardized mortality rates at 30 days and 1 year. RESULTS The median transfer rate was 29.4% (interquartile range [25th-75th percentile], 21.8%-37.8%). Higher transfer rates were associated with higher rates of catheterization (P < .001), percutaneous coronary intervention (P < .001), and coronary artery bypass graft surgery (P < .001). Median length of stay was not meaningfully different across the groups. There was no meaningful evidence of associations between transfer rates and risk-standardized mortality at 30 days (mean [SD], 22.3% [2.6%], 22.1% [2.3%], 22.3% [2.4%], and 21.7% [2.1%], respectively; P = .054) or 1 year (43.9% [2.3%], 43.6% [2.2%], 43.5% [2.4%], and 42.8% [2.2%], respectively; P < .001) for low, mid-low, mid-high, and high transfer groups. CONCLUSIONS AND RELEVANCE: Nonprocedure hospitals vary substantially in their use of the transfer process for elderly patients admitted with acute myocardial infarction. High-transfer hospitals had greater use of invasive cardiac procedures after admission compared with low-transfer hospitals. However, higher transfer rates were not associated with a significantly lower risk-standardized mortality rate at 30 days. Moreover, at 1 year there was only a 1.1% difference (42.8% vs 43.9%) between hospitals with higher and lower transfer rates. These findings suggest that, as a single intervention, promoting the transfer of patients admitted with acute myocardial infarction may not improve hospital outcomes.