基于机器学习的老年人医疗服务利用影响因素分析

Analysis of factors influencing medical service utilization among older adults based on machine learning

  • 摘要: 目的: 探讨医养结合机构老年人医疗服务利用现状及其影响因素,为优化机构内医疗服务资源配置、提升老年人健康管理水平提供实证依据。方法: 于2024年10—12月,采用分层抽样法对南宁、桂林、玉林3市10家医养结合机构的395例老年人进行问卷调查。运用随机森林模型对影响因素进行重要性排序,结合LASSO回归进行变量筛选,最终将显著变量纳入多因素logistic回归模型进行分析。结果: 门诊服务利用率为42.5%,住院服务利用率为32.4%。多因素分析结果显示,健康体检(OR=2.107,P=0.007)、疼痛(OR=2.447,P=0.000)是门诊服务利用的显著促进因素,而到医院时长<10 min组(OR=1.988,P=0.026)显著提升门诊服务利用;摔倒(OR=2.822,P=0.000)是住院服务利用的强影响因素,女性(OR=0.480,P=0.006)住院服务利用概率较低。门诊服务利用中家族常见病、遗传病在随机森林模型排名靠前,住院服务利用中摔倒、文化程度在随机森林模型排名靠前。结论: 广西医养结合机构老年人医疗服务利用受健康需求、服务可及性、健康信念及社会人口特征共同影响。建议从强化健康管理、优化服务布局、开展健康教育、关注弱势群体4个方面入手,提升服务利用效率与公平性。

     

    Abstract: Objective: To investigate the current status and influencing factors of medical service utilization among older adults in integrated care institutions, aiming to provide an empirical basis for optimizing the allocation of medical resources and enhancing health management for older adults. Methods: A cross-sectional survey was conducted from October to December 2024, employing stratified sampling to recruit 395 elderly residents from 10 integrated care institutions in the cities of Nanning, Guilin, and Yulin. The random forest model was used to rank the importance of influencing factors, combined with LASSO regression for variable selection. Significant variables were subsequently incorporated into a multivariate logistic regression model for analysis. Results: The utilization rates for outpatient and inpatient services were 42.5% and 32.4%, respectively. Multivariate analysis revealed that health checkups (OR=2.107, P=0.007) and pain (OR=2.447, P=0.000) were significant promoting factors for outpatient service utilization, whereas a travel time to the hospital of <10 minutes (OR=1.988, P=0.026) significantly increased outpatient utilization. A history of falls (OR=2.822, P=0.000) was a strong influencing factor for inpatient service utilization, while female gender (OR=0.480, P=0.006) was associated with a lower probability for inpatient service utilization. In terms of outpatient service utilization, common family diseases and genetic diseases ranked high in the random forest model, while for inpatient service utilization, a history of falls and education level ranked prominently in the model. Conclusion: Medical service utilization among older adults in integrated care institutions in Guangxi is jointly influenced by health needs, service accessibility, health beliefs, and sociodemographic characteristics. To enhance the efficiency and equity of service utilization, it is recommended to strengthen health management, optimize service layout, implement health education, and focus on vulnerable groups.

     

/

返回文章
返回