摘要
目的:探讨倾向性评分匹配在横断面研究中的应用,明确饮酒与高血压前期的相关性。方法:基于无锡市锡山区居民慢性病基线调查数据,利用倾向性评分匹配法均衡饮酒组与不饮酒组间的混杂因素效应,对匹配前后饮酒与高血压前期的相关性进行分析。结果:年龄、性别、婚姻状况等因素在匹配前组间分布不均衡,成功匹配649对后,所有混杂因素在组间分布均达到均衡。匹配前,饮酒组和不饮酒组高血压前期患病率差异有统计学意义(χ2=70.212,P<0.05);匹配后,饮酒组高血压前期患病率为71.96%,不饮酒组高血压前期患病率为68.88%,差异无统计学意义(P>0.05)。匹配后,各亚组间饮酒组和不饮酒组高血压前期患病率差异均无统计学意义(P>0.05)。结论:尚不能证实饮酒与高血压前期发生情况存在统计学关联;倾向性评分匹配法能够有效控制混杂偏倚,在处理横断面数据方面有一定应用前景。
OBJECTIVE To explore the application of propensity score matching in cross-sectional study,and to clarify the correlation between drinking and prehypertension.METHODS Based on the baseline survey data of chronic diseases in Xishan Dis⁃trict,the mixed factor effect between drinking group and no drinking group was balanced by propensity score matching,and the correlation between drinking and prehypertension before and after matching was analyzed.RESULTS The distribution of age,gender,marital status and other factors before matching was unbalanced.After 649 pairs were matched successfully,all the confounding factors were evenly distributed among groups.Before matching,the prevalence of prehypertension in drinking group and no drinking group was statistically significant(χ2=70.212,P<0.05).After matching,the prevalence of prehypertension in drinking group was 71.96%,and the prevalence of prehypertension in no drinking group was 68.88%,the difference was not statistically significant(P>0.05).After matching,there was no significant difference in prehypertension prevalence between drinking group and no drink⁃ing group(P>0.05).CONCLUSIONS There is no statistical correlation between drinking and prehypertension.Propensity score match⁃ing can effectively control the confounding bias,and it has a certain application prospect in processing cross-sectional data.
作者
李杰
顾月
LI Jie;GU Yue(Center for Disease Control and prevention in Xishan District,Wuxi,Jiangsu,214000,China)
出处
《中国初级卫生保健》
2021年第7期49-52,共4页
Chinese Primary Health Care
关键词
倾向性评分
饮酒
高血压前期
相关性
Propensity score
Drinking
Prehypertension
Correlation
