1972～2001年启东市肺癌发病趋势的时间序列分析及预测模型探讨  

Analysis on time trends forecast and the model comparison for lung cancer incidence rate in Qidong City from 1972 to 2001

作　　者：陈勇[1] 陈建国[2] 朱健[2] 张永辉[2] 沈洪兵[1] 陈峰[1] 徐耀初[1] 

机构地区：[1]南京医科大学公共卫生学院,江苏南京210029 [2]江苏省启东肝癌研究所,江苏启东226200

出　　处：《南京医科大学学报（自然科学版）》2005年第7期514-519,共6页Journal of Nanjing Medical University(Natural Sciences)

基　　金：国家自然科学基金资助项目(30170827)

摘　　要：目的:采用不同的预测模型拟合江苏省启东市1972￣2001年肺癌发病资料,并预测其到2015年的发展趋势,为启东市制订肺癌防制策略提供依据。方法:利用启东市肿瘤登记处积累的1972￣2001年肺癌登记资料,采用时间序列分析方法中的趋势外推法、指数平滑法、ARIMA模型(Box鄄Jenkins法)等,分析该地肺癌发病率的变化趋势。比较各方法预测精度,赋予不同权重,建立组合预测模型。结果:各种分析、预测方法显示:肺癌的发病率呈现上升趋势;保守预测估计至2010年肺癌发病率将突破50/10万,2015年肺癌发病率将达到60/10万,接近周边大城市肺癌发病率水平。其中男性肺癌发病率到2010年接近85/10万,2015年将达到90/10万以上;女性肺癌发病率到2010年接近30/10万,2015年将超过35/10万。到2010年启东市男女肺癌发病率将在2001年基础上分别增长38.85%和25.91%。结论:启东市30年(1972￣2001)肺癌发病率呈现持续增长趋势,应对该地居民肺癌发生影响因素及变动情况作进一步的探讨;探索比较了多种预测模型的拟合优度,在此基础上建立组合模型,可以达到综合预测的目的。Objective: To explore the trends of lung cancer incidence rates from 1972 to 2001 in Qidong city and to forecast the rates before 2015 and provide baseline data for the control and prevention. Methods: Trends on lung cancer incidence rates were analyzed based upon the data from the Qidong City Cancer Registry, and time-series analyses, such as trends extrapolation, exponent smoothness, Box-Jenkins model etc, were employed to compare all kinds of the prognostication precision and endow differ power for founding assembled forecast model. Results: Analysis results showed that incidences of lung cancer in males and females experienced rising trends from 1972 to 2001, and the rates would reach to 85 per 100 000 for males and 30 per 100 000 for females in 2010, and 90 per 100 000, 35 per 100 000, respectively in 2015. The incidence rates of lung cancer in 2010 would have an increase of 38.85% for males and 25.91% for females, respectively, if compared with the rate in 2001. Conclusion: Facing the increasing trends of lung cancer in Qidong area during the past 30 years, further investigation should be made to study the impact factors on the incidence. Assembled time-series models, with better goodness of fittest, may help to predict the trends of the disease.

关 键 词：肺癌 趋势预测 时间序列分析 

分 类 号：R73-31[医药卫生—肿瘤]