基于贝叶斯网络模型社区老年人衰弱影响因素网络关系分析及风险预测模型构建

Network Relationship and Risk Prediction Model of Community Elderly Frailty in Hebei Province Based on Bayesian Network

摘要: 目的：探究社区老年人衰弱的影响因素之间的网络关系，构建风险预测模型，为进一步精确预防提供依据。方法：2021年10月至2022年1月间对河北省三所城市的601名社区老年人，采用一般资料调查表、衰弱筛查量表等量表开展横断面调查。结果：在601名老年人中衰弱检出率为44.43%。非条件logistic回归分析表明年龄、锻炼情况、生活方式、孤独感、健康投入、一般自我效能、家庭关怀度是老年人衰弱的独立影响因素。贝叶斯网络模型分析结果显示各因素均与衰弱直接相关。高危组(年龄 ≥ 80岁 + 生活方式一般及以下 + 健康投入水平合适及以下 + 锻炼情况 < 0.5 h + 一般自我效能低 + 家庭关怀障碍 + 孤独感中等水平以上的)发生衰弱的概率达98.6%。结论：社区老年人衰弱发生率较高。构建的衰弱预警模型涵盖生理–心理–特质–行为–家庭五维度，该模型可量化多维风险特征，为社区早期识别及精准干预提供科学依据。

Abstract: Objective: To explore the network relationship between the influential factors of community elderly frailty in Hebei Province, build a risk prediction model, and provide basis for further accurate prevention. Methods: The general data questionnaire and frailty Screening scale were used to conduct a cross-sectional survey on 601 elderly people in three cities in Hebei Province. Results: The detection rate of frailty was 44.43%. Logistic regression analysis showed that age, exercise, lifestyle, loneliness, health involvement, general self-efficacy and family caring were independent factors. Bayesian network model shows that all factors are directly related to frailty. In the high-risk group (age ≥ 80 years + average lifestyle or below + appropriate level of health investment or below + exercise < 0.5 h+ low general self-efficacy + family care disorder + moderate level of loneliness or above), the probability of frailty was 98.6%. Conclusion: The incidence of frailty in community aged people is higher in Hebei Province. The frailty early warning model constructed includes five dimensions: physiological, psychological, trait, behavior and family. The model can quantify multi-dimensional risk characteristics and provide scientific basis for early identification and precise intervention in the community.