Ardabili Ahad Khaleghi, Sadr Alireza Vafaei, Abedi Vida, Bonavia Anthony S
Department of Anesthesiology and Perioperative Medicine, Penn State Milton S Hershey Medical Center, Hershey, PA 17036, USA.
Critical Illness and Sepsis Research Center (CISRC), Penn State College of Medicine, Hershey, PA 17036, USA.
medRxiv. 2024 Dec 20:2024.12.19.24319343. doi: 10.1101/2024.12.19.24319343.
To determine whether neighborhood-level social determinants of health (SDoH) influence mortality following sepsis in the United States.
Retrospective analysis of data from 4.4 million hospitalized patients diagnosed with sepsis, identified using International Classification of Diseases-10 codes, across the United States.
De-identified, aggregated data were sourced from the TriNetX Diamond Network. SDoH variables included income, housing cost burden, broadband access, park proximity, racial/ethnic diversity, and the Area Deprivation Index (ADI). The primary outcome was mortality, assessed using univariate and multivariate binomial generalized linear models. Predictors with high multicollinearity (Variance Inflation Factor > 5) were excluded to enhance model stability.
Lower median income, higher ADI scores, limited park access, and lack of broadband connectivity were strongly associated with increased sepsis mortality. Unexpectedly, greater racial/ethnic diversity was negatively associated with mortality, possibly reflecting regional disparities in healthcare access and socioeconomic conditions. Multivariate analyses revealed that the inclusion of SDoH variables attenuated some effects observed in univariate models, highlighting their complex interplay. Random Forest analysis identified park access as the most important predictor of sepsis mortality, emphasizing its role as a potential proxy for broader neighborhood resources.
Neighborhood-level SDoH are critical for risk stratification in sepsis prognostic models and should be systematically integrated into predictive frameworks. These findings highlight the need for targeted public health interventions to address social vulnerabilities, enhance access to green spaces, and reduce disparities in sepsis outcomes across diverse populations.
确定美国邻里层面的健康社会决定因素（SDoH）是否会影响脓毒症后的死亡率。
对美国440万例确诊为脓毒症的住院患者数据进行回顾性分析，这些数据通过国际疾病分类第10版代码识别。
去识别化的汇总数据来源于TriNetX钻石网络。SDoH变量包括收入、住房成本负担、宽带接入、公园距离、种族/族裔多样性以及地区贫困指数（ADI）。主要结局为死亡率，使用单变量和多变量二项式广义线性模型进行评估。排除具有高多重共线性（方差膨胀因子>5）的预测变量以增强模型稳定性。
较低的收入中位数、较高的ADI分数、有限的公园可达性以及缺乏宽带连接与脓毒症死亡率增加密切相关。出乎意料的是，更大的种族/族裔多样性与死亡率呈负相关，这可能反映了医疗保健可及性和社会经济状况的地区差异。多变量分析显示，纳入SDoH变量减弱了单变量模型中观察到的一些效应，突出了它们之间复杂的相互作用。随机森林分析确定公园可达性是脓毒症死亡率最重要的预测因素，强调了其作为更广泛邻里资源潜在替代指标的作用。
邻里层面的SDoH对于脓毒症预后模型中的风险分层至关重要，应系统地纳入预测框架。这些发现凸显了有针对性的公共卫生干预措施的必要性，以解决社会脆弱性、增加绿地可达性并减少不同人群脓毒症结局的差异。
