健康期望寿命是衡量人口生存质量的重要指标，“健康中国2030”和各省如“健康江西2030”等规划目标，都明确要求“人均健康期望寿命逐步提高”。然而，由于国内健康数据的匮乏，很多地方健康期望寿命的研究是空白，而对同一地方连续多年的估计和预测几乎没有。本研究拟用全球疾病负担研究的相关缺失数据估计方法，估计2008-2018年江西省每年不同性别的城市和农村自评健康期望寿命和患有不同慢病的健康调整期望寿命，并以此预测。为此，我们首先拟用死因监测等数据，用模型寿命表等方法，估计每年不同人群的寿命表。然后，拟用卫生服务调查等数据，用Bayes广义混合效应模型等方法，估计相应的年龄别自评健康率和慢病患病率。最后，用Sullivan法计算各健康期望寿命，并用Lee-Carter模型预测。研究结果，一方面将为相关政策提供直接参考依据，另一方面也为缺失数据下国人健康期望寿命估计和预测的方法体系提供了示范。

Healthy life expectancy is an important index to measure the quality of life of the population. The planning goal of program outline in "Health China 2030" as well as in the provincial programs such as "health Jiangxi 2030" has explicitly mentioned that "Healthy life expectancy will gradually increase". However, because of the lack of the domestic health data, there is no research on healthy life expectancy in many areas, and even there is no area in which many continuous years of healthy life expectancy had been estimated and predicted. In this project, we intends to estimate and predict the self-rated health life expectancy respectively for urban and rural population by sexes in Jiangxi province from 2008 to 2018, as well as life expectancy without chronic morbidities, based on the methods of missing data used in the Global Burden of Disease Study. Therefore, we will first use model life table and other methods to estimate each year of life tables for different population based on the data of Disease Surveillance Points system. Then, we will combine Bayesian generalized mixed-effects model with other models to estimate age-specific self-assessed health rate and prevalence of chronic diseases for the same population based on the data of National Health Service Surveys. Finally, Sullivan method will be used to calculate each of the healthy life expectancies, and Lee-Carter model will be used to predict their trends. The results will provide a direct reference for relevant policies on the one hand. On the other hand, it will demonstrate the system methods how to estimate the healthy life expectancy and its trend when the healthy data is lack.