Applying Intersectionality to Better Characterize Healthcare Disparities for Critically Ill Adults.
Applying Intersectionality to Better Characterize Healthcare Disparities for Critically Ill Adults.
复制标题
应用交叉性更好地描述危重成人的医疗保健差异。
DOI：
10.1164/rccm.202301-0153le
复制
发表时间：
2023
影响因子：
24.7
通讯作者：
Mehta,AnujB
中科院分区：
文献类型：
--
作者：
Day,GwenythL;Havranek,EdwardP;Campbell,EricG;Mehta,AnujB
方法我们使用2018年至2019年7个地理和种族多样化的州住院患者数据库（亚利桑那州、加州州、佛罗里达州、爱荷华州、马里兰州、密西西比州和纽约）进行了一项回顾性队列研究，确定了在入院时接受MV治疗的肺炎或脓毒症的成人非手术患者，并提供出院账单代码（9，10）。创建了基于种族和性别的8类交叉变量（男性白色、男性黑人、男性亚洲人、男性其他人、女性白色、女性黑人、女性亚洲人、女性其他人）。为了确定交叉研究方法是否识别出更大的变异性，我们分别根据性别、种族和交叉研究身份确定了医院死亡率。我们通过两种方式确定风险调整后的死亡率。在三个独立的模型中，我们使用分层逻辑回归，每个人口统计学类别（性别，种族和交叉身份）作为随机截距，以计算风险调整的死亡率百分比。在第二种方法中，人口统计学类别（性别、种族和交叉身份）在三个独立的分层回归模型中被视为固定效应，医院作为随机截距，以确定死亡的调整优势比。模型根据患者年龄、38例Elixhauser合并症和入院时出现的急性器官衰竭进行了调整（11）。确定为女性或亚裔的患者是年龄最大的组，而确定为其他人种的患者是年龄最小的组（表1）。一般来说，男性、黑人或亚洲人的急性和慢性疾病更多。风险调整后，风险调整后的死亡率没有显着差异，观察性别或种族的基础上，单独的，虽然男性患者的调整后的死亡率比女性患者略低。然而，当使用交叉身份时，发现医院死亡率的变异性更大（表2）。经风险调整的死亡率最高的患者为黑人女性（40.0%; 95%置信区间[CI]，38.8-41.1）和其他种族女性（39.4%; 95% CI，38.1-40.7）。确定为男性白色（36.9%; 95% CI，36.1-37.8）和男性黑人（37.6%; 95% CI，36.6-38.7）的患者的风险调整死亡率最低。确定为黑人女性的患者校正后的死亡几率高于其他女性组（女性白色和亚洲女性）以及确定为黑人男性的患者。
MethodsWe conducted a retrospective cohort study using seven geographically and racially diverse state inpatient databases (Arizona, California, Florida, Iowa, Maryland, Mississippi, and New York) from 2018 to 2019, identifying adult nonsurgical patients treated with MV with pneumonia or sepsis on admission with discharge billing codes (9, 10). An eight-category intersectional variable based on race and sex was created (male White, male Black, male Asian, male Other, female White, female Black, female Asian, female Other). To determine if an intersectional approach identified greater variability, we determined hospital mortality on the basis of sex, race, and intersectional identity separately. We determined risk-adjusted mortality in two ways. In three separate models, we used hierarchical logistic regression with each demographic category (sex, race, and intersectional identity) as a random intercept to calculate risk-adjusted mortality percentages. In the second approach, demographic categories (sex, race, and intersectional identity) were treated as fixed effects in three separate hierarchical regression models with the hospital as a random intercept to determine the adjusted odds ratio for death. Models were adjusted for patient age, 38 individual Elixhauser comorbidities, and acute organ failures present on admission (11).ResultsWe identified 161,560 adult nonsurgical patients with pneumonia or sepsis who received MV. Patients identified as female or Asian were the oldest groups, whereas patients identified as Other race were the youngest (Table 1). In general, patients identified as male, Black, or Asian were more acutely and chronically ill. After risk adjustment, no significant difference in risk-adjusted mortality rates was observed on the basis of sex or race alone, although male patients had slightly lower adjusted odds of mortality than female patients. However, when intersectional identities were used, greater variability in hospital mortality was identified (Table 2). The highest risk-adjusted mortality was among patients identified as female Black (40.0%; 95% confidence interval [CI], 38.8–41.1) and female Other race (39.4%; 95% CI, 38.1–40.7). The lowest risk-adjusted mortality rates were among patients identified as male White (36.9%; 95% CI, 36.1–37.8) and male Black (37.6%; 95% CI, 36.6–38.7). Patients identified as female Black had higher adjusted odds of death than Other female groups (female White and female Asian), as well as patients identified as male Black.