Deparment of Social and Behavioral Sciences, Harvard T.H. Chan School of Public Health, 401 Park Dr, 428F, Boston, MA 02215. Email:
Department of Behavioral and Social Sciences, Brown University School of Public Health, Providence, Rhode Island.
Prev Chronic Dis. 2020 May 28;17:E39. doi: 10.5888/pcd17.190382.
The Framingham risk score (FRS) is widely used to predict cardiovascular disease (CVD), but it neglects to account for social risk factors. Our study examined whether use of a cumulative social risk score in addition to the FRS improves prediction of CVD among South Korean adults.
We used nationally representative data on 19,147 adults aged 19 or older from the Korea National Health and Nutrition Examination Survey 2013-2016. We computed a cumulative social risk score (range, 0-3) based on 3 social risk factors: low household income, low level of education, and single-living status. CVD outcomes were stroke, myocardial infarction, and angina. Weighted logistic regression examined the associations between cumulative social risk, FRS, and CVD. McFadden pseudo-R and area under receiver operating characteristic curve (AUC) assessed model performance. We conducted mediation analyses to quantify the association between cumulative social risk score and CVD outcomes that is not mediated by the FRS.
A unit increase in social risk was associated with 89.4% higher risk of stroke diagnosis, controlling for the FRS (P < .001). The FRS explained 8.0% of stroke diagnosis (R) with fair discrimination (AUC = 0.728), and adding the cumulative social risk score enhanced R and AUC by 2.4% and 0.039. In the association between cumulative social risk and stroke, the proportion not mediated by the FRS was 65% (P < .001). We observed similar trends in myocardial infarction and angina, such that an increase in social risk was associated with increased relative risk of disease and improved disease diagnosis, and a large proportion of the association was not mediated by the FRS.
Controlling for the FRS, cumulative social risks predicted stroke, myocardial infarction, and angina among adults in South Korea. Future research is needed to examine non-FRS mediators between cumulative social risk and CVD.
弗雷明汉风险评分(FRS)被广泛用于预测心血管疾病(CVD),但它忽略了社会风险因素。我们的研究旨在探讨在韩国成年人中,除了 FRS 之外,使用累积社会风险评分是否可以改善 CVD 的预测。
我们使用了来自 2013-2016 年韩国国家健康与营养调查的 19147 名 19 岁及以上成年人的全国代表性数据。我们根据 3 个社会风险因素(家庭收入低、教育水平低和独居状态)计算了一个累积社会风险评分(范围为 0-3)。CVD 结局包括中风、心肌梗死和心绞痛。加权逻辑回归分析了累积社会风险、FRS 与 CVD 之间的关系。麦克法登伪 R 和接收者操作特征曲线下的面积(AUC)评估了模型性能。我们进行了中介分析,以量化累积社会风险评分与 CVD 结局之间的关联,该关联不受 FRS 的影响。
在校正 FRS 后,社会风险每增加一个单位,中风诊断的风险增加 89.4%(P <.001)。FRS 解释了 8.0%的中风诊断(R),具有良好的区分度(AUC = 0.728),而累积社会风险评分增加了 2.4%和 0.039。在累积社会风险与中风之间的关联中,不受 FRS 影响的比例为 65%(P <.001)。我们在心肌梗死和心绞痛方面观察到类似的趋势,即社会风险的增加与疾病的相对风险增加和疾病诊断的改善相关,且该关联的很大一部分不受 FRS 的影响。
在校正 FRS 后,累积社会风险可预测韩国成年人的中风、心肌梗死和心绞痛。未来的研究需要检验累积社会风险与 CVD 之间非 FRS 中介因素。