Abstract 356: Hospital Safety Net Status and Clinical Outcomes After Trans-catheter Aortic Valve Replacement

Abstract

Introduction: Safety-net hospitals (SNH) treat a large population of un-insured and low income patients; several prior studies report worse outcome at these centers. Trans-catheter valve replacement (TAVR) is emerging as first-line therapy for aortic stenosis irrespective of surgical risk scores. However, results of TAVR performed at these centers is limited. Objective: To determine whether post-procedural outcomes of TAVR are comparable at safety-net (SNH) and non-safety net hospitals (non-SNH). Methods: We conducted a retrospective, cohort study with propensity-matched analysis. Complex survey data from the Agency for Healthcare Quality and Research containing weighted sample of all hospital admissions nationwide was utilized for this study. Adults undergoing TAVR at US hospitals participating in the National In-patient sample (NIS) database from January 2014 - December 2015 were included. A 1:1 propensity-matched cohort of patients operated at safety-net hospitals (SNH) and non-SNH institutions was analyzed. Propensity-matching was performed on the basis of sixteen demographic and clinical confounding co-variates. Main outcome studied was all-cause post-procedural mortality. Secondary outcomes compared were stroke, acute kidney injury and length of post-operative stay. Results: Between 2014 - 2015, 41410 patients (mean age 80 +/- 0.11 years, 46% female) underwent TAVR at 731 centers nationwide; 6996 (16.80 %) procedures were performed at safety net centers. SNH comprised 135/731 (18.4%) of all centers performing TAVR. SNH patients were more likely to be female (49 % vs 46 %, p <0.001); admitted emergently (31% vs 21%; p <0.001) and at the lowest quartile for household income (25% % vs 20 %; p <0.001). A large proportion of SNH patients were minorities (Blacks 5.9% vs 3.9%; Hispanic 7.2% vs 3.2%). Adjusted logistic regression was performed on 6995 propensity-matched patient pairs. Post-procedural mortality [OR 0.99 (0.98 - 1.007); p = 0.43], stroke [OR 1.009 (0.99-1.02); p = 0.08], and acute kidney injury [OR 0.99 (0.96 - 1.01); p = 0.5] were comparable in both cohorts. Overall length of stay was also similar (6.9 +/- 0.1 vs 7.1 +/- 0.2 days; p = 057). Conclusion: Post-procedural outcomes after TAVR at SNH are comparable to national outcomes. Our study provides preliminary evidence that wider adoption of TAVR may not adversely influence outcomes at SNH.