Silverman Ariel R C, Chung Paul J, Gould Michael K, Ngo-Metzger Quyen, Tauali'i Maile M, Mosen David M, Duggan Mark C, Nocon Robert S
Kaiser Permanente Bernard J. Tyson School of Medicine, Pasadena, CA, USA.
Department of Internal Medicine, University of California, Los Angeles, Los Angeles, CA, USA.
J Gen Intern Med. 2025 Jun 25. doi: 10.1007/s11606-025-09647-8.
While research on health-related social needs (HRSNs) has expanded, important gaps remain in understanding associations between HRSN and healthcare cost, especially across general populations of patients with broad-ranging medical and social needs.
To examine the association between HRSN and healthcare cost in a large, diverse, insured population.
In this cross-sectional study, we evaluated cost differences between patients with different HRSN levels using survey-weighted multivariable generalized linear models. We examined three alternate model specifications: one that included controls for basic demographics; another adding insurance type, race/ethnicity, and social isolation; and a third adding a diagnosis-based medical risk index called DxCG. Variables added in the latter models were assessed separately due to concern for over-correlations with HRSN.
10,226 adult survey respondents (23% response rate) from eight states and Washington, D.C.
The primary exposure was patient-reported HRSN, based on survey questions asking about financial strain, food insecurity, housing instability, and transportation difficulties. HRSN was constructed as a three-level variable. The primary outcome was total direct healthcare cost.
"Moderate HRSN" was not significantly associated with cost in any model. In the model controlling for patient demographics, costs for patients with "Severe HRSN" were 1.27 (95%CI: 1.00-1.60) times those of patients with "No HRSN." In a model that adjusted for medical risk (DxCG), the relationship between HRSN and cost was not statistically significant.
Relationships between HRSN and healthcare cost may vary by HRSN level. Our mixed findings highlight the complex relationship between medical and social risks, which often have bi-directional causal relationships. If measures of medical risk incidentally capture variation in social risk, then models controlling for medical risk may mask relationships between HRSN and cost. Further studies should investigate the extent to which HRSN may be related to cost, even when controlling for specific patient diagnoses.
虽然对健康相关社会需求（HRSN）的研究有所扩展，但在理解HRSN与医疗保健成本之间的关联方面仍存在重大差距，尤其是在具有广泛医疗和社会需求的普通患者群体中。
在一个大型、多样化的参保人群中研究HRSN与医疗保健成本之间的关联。
在这项横断面研究中，我们使用调查加权多变量广义线性模型评估了不同HRSN水平患者之间的成本差异。我们研究了三种替代模型规格：一种包括基本人口统计学控制变量；另一种增加了保险类型、种族/族裔和社会隔离变量；第三种增加了一个基于诊断的医疗风险指数DxCG。由于担心与HRSN的过度相关性，后一种模型中添加的变量进行了单独评估。
来自八个州和华盛顿特区的10226名成年调查受访者（回复率23%）。
主要暴露因素是患者报告的HRSN，基于询问财务压力、粮食不安全、住房不稳定和交通困难的调查问题。HRSN被构建为一个三级变量。主要结局是直接医疗保健总成本。
在任何模型中，“中度HRSN”与成本均无显著关联。在控制患者人口统计学的模型中，“严重HRSN”患者的成本是“无HRSN”患者的1.27倍（95%CI：1.00 - 1.60）。在调整了医疗风险（DxCG）的模型中，HRSN与成本之间的关系无统计学意义。
HRSN与医疗保健成本之间的关系可能因HRSN水平而异。我们的混合研究结果突出了医疗和社会风险之间的复杂关系，这种关系通常具有双向因果关系。如果医疗风险测量偶然捕捉到社会风险的变化，那么控制医疗风险的模型可能会掩盖HRSN与成本之间的关系。进一步的研究应该调查即使在控制特定患者诊断的情况下，HRSN与成本可能相关的程度。
