摘要
目的 介绍隐函数Delta法计算调整人群归因危险度 (PAR)可信区间的基本原理和SAS编程方法。方法 实例介绍用SASPROCIML过程 ,实现病例 -对照研究中调整PAR及其可信区间的估计。结果 利用此程序计算了上海市男性食管癌病例 -对照研究资料吸烟、饮酒和水果摄入过少因素的调整PAR及其 95 %可信区间分别为 46 5 9%( 3 4 92 %～ 5 8 2 6%) ,2 0 87%( 10 2 8%～ 3 1 47%)和 3 4 5 4%( 19 81%～ 49 2 9%)。结论 对符合条件的病例 -对照资料 。
Objective The paper introduced basic idea of estimating the confidence intervals of the adjusted population attributable risk based on the implicit delta method and sas program. Methods Using the example illustrated how to write the program with the proc iml procedure in sas environment to carry out the estimation of the adjusted PAR and confidence intervals with the case-control data. Results The adjusted PAR and 95% confidence intervals for the cigarette smoking, alcohol drinking and less fresh fruits intake attributed to the esophageal cancer among the male in Shanghai urban city were calculated as 46.59%(34.92%～58.26%), 20.87%(10.28%～31.47%) and 34.54%(19.81%～49.29%), respectively.Conclusion Through the implicit delta method, the accurate estimation of the PAR and confidence intervals were more likely to acquire provided the case control data satisfied the corresponding condition. It has the practical significance for the method to the prevention from the chronic disease.
出处
《中国卫生统计》
CSCD
北大核心
2004年第3期142-145,共4页
Chinese Journal of Health Statistics
