肝癌发病时期趋势与年龄效应的贝叶斯图解模型估计  

Application of Bayesian Graphical Modelling to Analysis of the Incidence Data of Primary Liver Cancer in Qidong County

作　　者：宗占红[1] 沈其君[2] 陈建国[3] 李文广[3] 姚红玉[3] 

机构地区：[1]南京人口管理干部学院,210009 [2]宁波大学医学院,315211 [3]启东肝癌防治研究所

出　　处：《中国卫生统计》2005年第1期13-15,共3页Chinese Journal of Health Statistics

基　　金：江苏省卫生科研基金资助

摘　　要：目的　为了较为系统地介绍贝叶斯图解模型 ,并用于肝癌发病资料分析和发病率预测。方法　利用贝叶斯图解模型和年龄 队列模型对肝癌发病的年龄、出生队列效应的参数进行估计。结果　三个区肝癌发病自 195 8～ 1962年出生队列有明显下降 ,下一个时期的发病率整体上将适度下降 ,并且贝叶斯图解模型估计的参数稳定 ,标准误小。结论　肝癌发病的时期趋势体现在出生队列的变化上 ;贝叶斯图解模型估计参数具有较好的稳定性 。Objective This article introduces Bayesian graphical modelling and analyze incidence data of primary live cancer in Qidong county,and predict incidence rate of Primary live cancer next period.Methods Age cohort models was used to fit data of twenty five years PLC incidence in Qidong to estimate parameters for birth cohort and age effect by means of Bayesian graphical modeling.Results A significant decrease trend in incidence risk begin with birth cohort around 1958～1962,the estimation of parameters is stable.Conclusion The eatimation of parameters using Bayesian graphical modelling was robust better than the MLE,which could reveal the trends of incidence well.

关 键 词：肝癌 发病率 出生 发病时期 发病资料 发病趋势 年龄 模型估计 下降 队列 

分 类 号：R735.7[医药卫生—肿瘤]