广州市社区老年人跌倒发生率及影响因素的前瞻性队列研究

A prospective cohort study on the prevalence and influencing factors of falls among community-dwelling older adults in Guangzhou City

目的探究广州市社区老年人跌倒发生率及影响因素,并分析不同影响因素之间的协同模式和影响路径。方法采用前瞻性队列研究设计,对2021年广州市≥65岁的4950名社区老年人开展基线调查,并于2022年进行随访,收集跌倒事件的发生及健康情况信息。采用Cox比例风险回归模型分析跌倒的影响因素,并利用关联规则和贝叶斯网络模型分析影响因素的协同模式和影响路径。结果最终纳入有效样本3393例,中位随访时间为1.53年,跌倒发生率为25.49%(95%CI:24.02%~26.96%)。Cox比例风险回归模型结果显示,女性(HR=1.246,95%CI:1.088~1.426)、年龄≥75岁(HR=1.343,95%CI:1.133~1.592)、,有医疗保险(HR=1.440,95%CI:1.038~1.997)、患糖尿病(HR=1.309,95%CI:1.143~1.498)和脑卒中(HR=1.914,95%CI:1.309~2.799)是老年人跌倒的独立危险因素;而高中及以上学历(HR=0.861,95%CI:0.750~0.987)、每天锻炼(HR=0.684,95%CI:0.580~0.807)和健康状态自我评估满意(HR=0.484,95%CI:0.278~0.841)与跌倒风险降低相关。关联规则分析揭示了糖尿病、女性与跌倒的协同模式(支持度:19.07%,置信度:32.77%,提升度:1.29)。贝叶斯网络模型分析结果显示,锻炼频率、年龄、文化程度、健康状态自我评估、性别、患糖尿病和脑卒中是影响老年人跌倒的关键因素,模型受试者工作特征曲线下面积(area under the curve,AUC)为0.612。结论广州市社区老年人跌倒发生率较高,且影响因素多样。建议针对缺乏锻炼、高龄、女性及慢性病患者等高风险群体,制定和实施综合性的干预措施。

Objective To understand the prevalence and influencing factors of falls among community-dwelling older adults in Guangzhou City,and to explore the synergistic patterns and influencing pathways among different factors.Methods The prospective cohort study method was used for a baseline survey of 4950 community-dwelling older adults aged 65 years and above in Guangzhou City in 2021;the occurrence of falls and other health conditions among baseline older adults was followed up in 2022.Cox proportional hazards model was used to analyze the influencing factors of falls.Association rules and Bayesian network modeling were used to analyze the synergistic patterns and influence pathways of the influencing factors.Results A final valid sample of 3393 older adults was included with a median follow-up of 1.53 years.The prevalence of falls was 25.49%(95%CI:24.02%-26.96%).Cox proportional hazards model results showed that females(HR=1.246,95%CI:1.088-1.426),≥75 years(HR=1.343,95%CI:1.133-1.592),insured(HR=1.440,95%CI:1.038-1.997),diabetes(HR=1.309,95%CI:1.143-1.498)and stroke(HR=1.914,95%CI:1.309-2.799)increased the risk of falls occurring in older adults;high school and above(HR=0.861,95%CI:0.750-0.987),daily exercise(HR=0.684,95%CI:0.580-0.807),and satisfaction with self-assessment of health status(HR=0.484,95%CI:0.278-0.841)reduced the risk of falls among older adults.The results of association rules showed that the most common collaboration mode was diabetes,females,and falls(support:19.07%,confidence:32.77%,lift:1.29).The result of Bayesian Network model showed that exercise frequency,age,education,self-assessment of health status,gender,diabetes,and stroke were the important influencing factors of falls,and the area under the curve of the model was 0.612.Conclusions The prevalence of falls among community-dwelling older adults in Guangzhou City was high,and the influencing factors are complex.It is recommended to develop and implement comprehensive intervention measures for high-risk groups,including those who are physically inactive,elderly,female,and those with chronic diseases.