SAS交互式矩阵语言在膳食调查营养素数据处理中的应用  

Application of SAS interactive matrix language in the nutrient data processing of dietary investigation

作　　者：张彬艳 米白冰 王予童 刘卉萌 党少农 颜虹 Zhang Binyan;Mi Baibing;Wang Yutong;Liu Huimeng;Dang Shaonong;Yan Hong(Department of Epidemiology and Biostatistics, School of Public Health, Xi′an Jiaotong University Health Science Center, Xi′an 710061, China)

机构地区：[1]西安交通大学医学部公共卫生学院流行病与卫生统计学系,陕西西安710061

出　　处：《中国医院统计》2022年第2期150-155,共6页Chinese Journal of Hospital Statistics

基　　金：国家自然科学基金(81230016);国家重点研发计划资助项目(2017YFC0907200,2017YFC0907201);陕西省卫计委出生缺陷防治项目(Sxwsjswzfcght2016-013)。

摘　　要：目的介绍基于SAS交互式矩阵语言计算半定量食物频率问卷(SQFFQ)营养素摄入量,以期应用于营养流行病学膳食调查。方法以陕西大规模出生缺陷调查数据为基础,采用SAS 9.4 PROC SURVEYSELECT程序中的简单随机抽样法随机选取3000份SQFFQ作为训练集,运用交互式矩阵语言PROC IML编程演示计算SQFFQ营养素摄入量。结果根据矩阵运算原理,采用PROC IML获得计算模块,并运用3000例样本进行验证获得25种营养素摄入量;同时,SAS日志输出窗口看到PROC IML过程执行时间为0.59 s,高效快速。结论应用SAS交互式矩阵语言计算SQFFQ营养素摄入量高效、简洁且准确度高,适合SQFFQ分析应用。Objective To introduce the calculation of nutrient intake based on SAS interactive matrix language for semi-quantitative food frequency questionnaire(SQFFQ)and its application in nutritional epidemiological dietary investigation.Methods Three thousand women who completed SQFFQ were randomly selected from a large-scale birth defect investigation dataset in Shaanxi Province as a training dataset using the simple random sampling method by PROC SURVEYSELECT program with SAS 9.4.Next,the interactive matrix language PROC IML programming was used to demonstrate the calculation of SQFFQ nutrient intake.Results According to the principle of matrix operation,PROC IML was used to obtain the calculation module and 3000 samples were validated to obtain 25 nutrients intake.Meanwhile,in the SAS log output window,the execution time of the PROC IML process was 0.59 s,which was efficient and fast.Conclusion The SAS interactive matrix language was used to calculate the nutrient intake of SQFFQ with high efficiency,simplicity and accuracy,which was suitable for SQFFQ analysis.

关 键 词：SAS 交互式矩阵语言 半定量食物频率 营养素 

分 类 号：R195.1[医药卫生—卫生统计学]