Li Yue, Cai Xueya, Glance Laurent G
Departments of *Public Health Sciences, Division of Health Policy and Outcomes Research †Biostatistics and Computational Biology ‡Anesthesiology, University of Rochester Medical Center, Rochester, NY.
Med Care. 2015 Dec;53(12):1058-65. doi: 10.1097/MLR.0000000000000441.
To examine racial and site-of-care disparities in all-cause and potentially avoidable 30-day rehospitalization rates among a national cohort of Medicare skilled nursing facility (SNF) residents.
We analyzed the 2012 Minimum Data Set, Medicare inpatient claims, and other data. Multivariable logistic regressions were used to adjust for resident demographic, functional, and diagnostic characteristics, as well as observed SNF and geographic factors. Conditional fixed effects for SNFs were further used to adjust for both observed and unobserved factors. Independent effects of black race and site-of-care groups were estimated, where sites were defined using proportions of black Medicare admissions to the SNF.
The 30-day all-cause and potentially avoidable rehospitalization rates were 21.9% and 8.8%, respectively, for black residents (n=120,508), and 17.7% and 7.9% for white residents (n=1,182,003). Racial disparities persisted after adjustment for resident characteristics. Moreover, risk-adjusted disparities were essentially related to the type of SNFs to which residents were admitted; after controlling for SNF sites, significant racial disparity disappeared for potentially available rehospitalizations. Black residents and white residents admitted to SNFs with high proportions of black admissions (>25%) were 31% and 19%, respectively, more likely to be rehospitalized than white residents admitted to SNFs caring for only a small percentage of black postacute residents (<3%).
Compared with white SNF residents, black SNF residents are more likely to be rehospitalized even after adjusting for patient risk factors. Black-white disparities, especially in potentially preventable rehospitalizations, are largely due to the fact that black residents tend to be admitted to the small number of SNFs with very high rehospitalization rates.
在全国医疗保险熟练护理机构(SNF)居民队列中,研究全因及潜在可避免的30天再住院率方面的种族差异和护理地点差异。
我们分析了2012年最小数据集、医疗保险住院索赔及其他数据。采用多变量逻辑回归来调整居民的人口统计学、功能和诊断特征,以及观察到的SNF和地理因素。进一步使用SNF的条件固定效应来调整观察到的和未观察到的因素。估计黑人种族和护理地点组的独立效应,护理地点根据SNF中黑人医疗保险入院比例来定义。
黑人居民(n = 120,508)的30天全因再住院率和潜在可避免再住院率分别为21.9%和8.8%,白人居民(n = 1,182,003)分别为17.7%和7.9%。在调整居民特征后,种族差异依然存在。此外,风险调整后的差异主要与居民入住的SNF类型有关;在控制SNF地点后,潜在可避免再住院方面的显著种族差异消失。入住黑人入院比例高(>25%)的SNF的黑人居民和白人居民,比入住仅照顾一小部分黑人急性后期居民(<3%)的SNF的白人居民再住院的可能性分别高31%和19%。
与白人SNF居民相比,即使在调整患者风险因素后,黑人SNF居民再住院的可能性仍更高。黑人和白人之间的差异,尤其是在潜在可预防的再住院方面,很大程度上是因为黑人居民往往入住少数再住院率非常高的SNF。