基于随机滤波模型的老年人剩余寿命预测

Residual life prediction of the elderly based on stochastic filtering model

自理能力是老年人健康状况的重要指标,但其与剩余寿命之间的关系一直缺乏严格的量化研究.本文根据中国营养健康调查中不小于55岁人群的自理能力面板数据,构建随机滤波模型预测其剩余寿命概率分布,并采用拟合优度检验测试模型的准确度.研究结果表明,自理能力和健康状况有着密切的关系,是老年人剩余寿命的重要指征;在生化指标难以得到时,构建的模型可以通过观察老年人的自理能力有效地预测其寿命分布.这为人群健康管理所倡导的关注老年人自理能力提供了科学依据,并对人寿保险公司根据老年人自理能力来确定寿险的费率提供了重要参考.

Although activities of daily living （ADL） is an important indicator of the health status of the elderly, the rigorous quantitative studies of the relationship between it and the residual life has been well studied. According to the panel survey data of the elderly aged from 55 upward from China health and nutrition survey, we constructed a stochastic filtering model to predict the residual life based on ADL. Then, the goodness-of-fit test was carried out to test the model. The result demonstrates that the ADL which is closely related to health status can be well used in residual life prediction when biochemical markers are not available in most cases. It provides a scientific basis for paying attention to the ADL of the elderly which is advocated by population health management and an important reference for determining premiums by life insurance companies through the ADL of the elderly.