贵州省苗岭以南地区气象因素对心脑血管疾病影响的分析与预测  

Analysis and prediction of meteorological factors affecting cardiovascular and cerebrovascular diseases in the southern region of the Miao Mountains,Guizhou Province

作　　者：尚媛媛 段莹 龙杰琦 彭波 杜正静 李光一 SHANG Yuan-yuan;DUAN Ying;LONG Jie-qi;PENG Bo;DU Zheng-jing;LI Guang-yi(Guizhou Provincial Center for Ecological and Agricultural Meteorology,Guiyang,Guizhou 550002,China;不详)

机构地区：[1]贵州省生态与农业气象中心,贵州贵阳550002 [2]贵州省山地气象科学研究所

出　　处：《现代预防医学》2024年第19期3594-3601,共8页Modern Preventive Medicine

基　　金：低纬山区心脑血管疾病气象预报技术研究与应用(CXFZ2022J071)。

摘　　要：目的 探究贵州省凯里市气象因素与心脑血管疾病(CVD)发病率的关联性,并评估气象因素的滞后效应。方法收集2018-2022年凯里市CVD发病数据及同期气象数据,进行了关联性分析。采用广义线性模型(GLM)来评估气象因素对CVD发病的滞后效应。结果 凯里市CVD年平均发病率为348.75/10万,无明显季节性峰值。Pearson相关性分析显示,CVD发病与气压、相对湿度、日照时数和降水量呈显著正相关(P<0.05),而与气温和风速呈显著负相关。以气温为例,日均气温升高8℃时,滞后1 d的相对风险(RR)值超过1.2;日均气温下降超过11℃时,滞后3 d的RR值增加至高于1.2,滞后7 d时RR值超过1.5。以气压为例,气压下降8hPa时,滞后1 d的RR值超过1.2;气压下降12hPa时,滞后1 d的RR值超过1.5。以降水量为例,累积降水量超过170 mm时,RR值超过1.2;累积降水量超过200 mm时,RR值超过1.7。机器学习模型在预测CVD高发病风险方面表现出色,准确率超过90%,其中随机森林和支持向量机模型表现尤为突出。结论 气温、气压和降水量与CVD发病存在非线性关系,并具有显著的滞后效应。机器学习方法能够有效预测CVD高发病风险,为公共卫生决策提供了有力的支持。Objective To explore the correlation between meteorological factors and the incidence of cardiovascular diseases(CVD)in Kaili city,Guizhou Province,and to assess the lag effects of these meteorological factors.Methods Data on CVD incidence and corresponding meteorological data from 2018 to 2022 in Kaili city were collected for correlation analysis.A generalized linear model(GLM)was employed to evaluate the lag effects of meteorological factors on CVD incidence.Results The annual average incidence of CVD in Kaili city was 348.75 per 100000,with no significant seasonal peaks.Pearson cor-relation analysis revealed a significant positive correlation between CVD incidence and atmospheric pressure,relative humidity,sunshine duration,and precipitation(P<0.05),while a significant negative correlation was found with temperature and wind speed.For instance,when the daily average temperature increased by 8°C,the relative risk(RR)value lagging by one day ex-ceeded 1.2;when the daily average temperature decreased by more than 11°C,the RR value increased to above 1.2 after a lag of three days,and exceeded 1.5 after a lag of seven days.In terms of atmospheric pressure,a decrease of 8 hPa resulted in an RR value exceeding 1.2 after one day of lag;a decrease of 12 hPa resulted in an RR value exceeding 1.5 after one day.Re-garding precipitation,when the cumulative precipitation exceeded 170 mm,the RR value exceeded 1.2;when it exceeded 200 mm,the RR value surpassed 1.7.Machine learning models demonstrated excellent performance in predicting the risk of high incidence of cardiovascular diseases,achieving an accuracy rate exceeding 90%,with particularly outstanding results from the random forest and support vector machine models.Conclusion There is a nonlinear relationship between temperature,atmo-spheric pressure,and precipitation with CVD incidence,accompanied by significant lag effects.Machine learning methods can effectively predict the risk of high incidence of CVD,providing strong support for public health decision-making.

关 键 词：气象因素 心脑血管病 关联分析 滞后效应 机器学习 

分 类 号：R54[医药卫生—心血管疾病] R122[医药卫生—内科学]