Edward P. Fitts Department of Industrial and Systems Engineering, North Carolina State University, Campus Box 7906, 400 Daniels Hall, Raleigh, NC, 27695, USA.
Poole College of Management, North Carolina State University, Campus Box 7229, Nelson Hall, Releigh, NC, 27695, USA.
Health Care Manag Sci. 2018 Dec;21(4):534-553. doi: 10.1007/s10729-017-9408-4. Epub 2017 Jul 22.
In the US, one in four adults has two or more chronic conditions; this population accounts for two thirds of healthcare spending. Comorbidity, the presence of multiple simultaneous health conditions in an individual, is increasing in prevalence and has been shown to impact patient outcomes negatively. Comorbidities associated with diabetes are correlated with increased incidence of preventable hospitalizations, longer lengths of stay (LOS), and higher costs. This study focuses on sex and race disparities in outcomes for hospitalized adult patients with and without diabetes. The objective is to characterize the impact of comorbidity burden, measured as the Charlson Weighted Index of Comorbidities (WIC), on outcomes including LOS, total charges, and disposition (specifically, probability of routine discharge home). Data from the National Inpatient Sample (2006-2011) were used to build a cluster-analytic framework which integrates cluster analysis with multivariate and logistic regression methods, for several goals: (i) to evaluate impact of these covariates on outcomes; (ii) to identify the most important comorbidities in the hospitalized population; and (iii) to create a simplified WIC score. Results showed that, although hospitalized women had better outcomes than men, the impact of diabetes was worse for women. Also, non-White patients had longer lengths of stay and higher total charges. Furthermore, the simplified WIC performed equivalently in the generalized linear models predicting standardized total charges and LOS, suggesting that this new score can sufficiently capture the important variability in the data. Our findings underscore the need to evaluate the differential impact of diabetes on physiology and treatment in women and in minorities.
在美国，每四个成年人中就有两个或两个以上患有两种或两种以上的慢性疾病；这部分人群占医疗保健支出的三分之二。合并症，即个体同时存在多种健康状况，其发病率正在上升，并已被证明对患者的预后产生负面影响。与糖尿病相关的合并症与可预防的住院率增加、住院时间延长（LOS）和成本增加有关。本研究关注的是患有和不患有糖尿病的住院成年患者的结局中的性别和种族差异。其目的是描述合并症负担（以 Charlson 合并症加权指数（WIC）衡量）对包括 LOS、总费用和处置（特别是常规出院回家的概率）在内的结局的影响。本研究使用来自国家住院患者样本（2006-2011 年）的数据，构建了一个聚类分析框架，该框架将聚类分析与多变量和逻辑回归方法相结合，用于实现以下几个目标：（i）评估这些协变量对结局的影响；（ii）确定住院人群中最重要的合并症；（iii）创建简化的 WIC 评分。结果表明，尽管住院女性的结局优于男性，但糖尿病对女性的影响更糟。此外，非白人患者的住院时间更长，总费用更高。此外，简化的 WIC 在预测标准化总费用和 LOS 的广义线性模型中表现相当，这表明该新评分可以充分捕捉数据中的重要变异性。我们的研究结果强调了需要评估糖尿病对女性和少数民族生理和治疗的差异影响。
