S. Singh1, S. Armenia1, A. Merchant1, D. H. Livingston1, N. E. Glass1 1New Jersey Medical School,Department Of Surgery,Newark, NJ, USA
Introduction:
Evidence supports index over interval cholecystectomy for patients admitted with acute cholecystitis. Though studies have shown poorer outcomes at safety-net hospitals (SNH), the effect of SNH has been considered as a single variable and not been stratified by other factors such as payer status and hospital characteristics. Our hypothesis was that for the underinsured, SNH will provide a higher level of care compared to non-SNH. We compared the treatment of patients with acute cholecystitis using index cholecystectomy and length of stay (LOS) as proxies for quality and cost of care between SNH and non-SNH controlling for insurance status, hospital region, size, and teaching status.
Methods:
National Inpatient Sample 2012-2014 was queried for all patients ≥18 years with acute cholecystitis. The primary and secondary outcomes of interest were index cholecystectomy and LOS, respectively. Proportion of Medicaid and uninsured discharges were used to define SNH (highest quartile) and non-SNH (lowest quartile). Multivariate logistic regression was used to calculate associations between outcomes and the effect of SNH designation, stratified by insurance status and hospital characteristics, while controlling for other factors such as demographics and comorbidities.
Results:
403,370 discharges with acute cholecystitis were identified (241,505 SNH; 161,865 non-SNH). SNH discharges were younger, more male, less white, and had fewer comorbid conditions than those at non-SNH (all p < 0.0001). SNH and non-SNH discharges had similar rates of index cholecystectomy, except at SNH in the northeast (OR: 0.74, p = 0.001). Patients at SNH had longer LOS for acute cholecystitis regardless of treatment; cholecystectomy or no surgery. When controlling for insurance status, patients at SNH had longer lengths of stay compared with non-SNH: Medicare (OR: 1.17, p = 0.0076), private insurance (OR 1.18, p = 0.0489), and uninsured discharges (OR: 1.47, p = 0.0159). There was also increased LOS in SNH relative to non-SNH in the Midwest, in urban non-teaching and teaching hospitals, and in large hospitals (all as defined by NIS; 1.36, p = 0.0195; 1.28, p = 0.0111; 1.29, p = 0.0037; 1.25, p = 0.0202).
Conclusion:
Quality of care disparities cannot be explained by simply examining SNH status as a single independent variable. Instead the effect of SNH on outcomes is a complex relationship between other variables like insurance status, region, teaching status of hospital, etc. While our data did not demonstrate superiority of care for uninsured patients at SNH we did show general equivalence of care with respect to index cholecystectomy. Regional differences in cholecystectomy in the northeast require further exploration. The variability and increased LOS at SNH highlight potential opportunities to further improve quality and decrease cost of care at our most vulnerable hospitals.