Mortensen Laura Quitzau, Andresen Kristoffer, Burcharth Jakob, Pommergaard Hans-Christian, Rosenberg Jacob
Center for Perioperative Optimization, Department of Surgery, Herlev Hospital, Herlev, Denmark.
Department of Surgery, University Hospital of Zealand, Koege, Denmark.
Front Big Data. 2019 May 8;2:4. doi: 10.3389/fdata.2019.00004. eCollection 2019.
Matching is frequently used in observational studies, especially in medical research. However, only a small number of articles with matching programs for the SAS software (SAS Institute Inc., Cary, NC, USA) are available, even less are usable for inexperienced users of SAS software. This article presents a matching program for the SAS software and links to an online repository for examples and test data. The program enables matching on several variables and includes in-depth explanation of the expressions used and how to customize the program. The selection of controls is randomized and automated, minimizing the risk of selection bias. Also, the program provides means for the researcher to test for incomplete matching.
匹配在观察性研究中经常被使用,尤其是在医学研究中。然而,只有少数几篇关于SAS软件(SAS Institute Inc.,美国北卡罗来纳州卡里)匹配程序的文章,对于没有经验的SAS软件用户来说,可用的更少。本文介绍了一个用于SAS软件的匹配程序,并提供了一个在线存储库的链接,其中包含示例和测试数据。该程序能够对多个变量进行匹配,并对所使用的表达式以及如何定制程序进行了深入解释。对照的选择是随机且自动化的,将选择偏倚的风险降至最低。此外,该程序还为研究人员提供了测试不完全匹配的方法。