In observational research, a primary objective is to accurately and reliably assess the causal impact of exposure on outcomes. Identifying and properly adjusting for confounding factors is a key prerequisite and central challenge to achieving this goal. Ineffective management of confounders, whether by neglecting significant ones, (leading to residual confounding), or by over-adjusting for irrelevant factors, (introducing collider bias), can distort effect estimates and lead to erroneous scientific conclusions and clinical decisions. Therefore, it is essential to develop and implement systematic, transparent, and reproducible methods for identifying and selecting confounding factors to enhance the validity and reliability of causal inferences in observational studies. This paper provides a systematic review of directed acyclic graphs (DAGs), a robust visual causal modeling tool, and offers a detailed examination of three prominent criteria for selecting confounding factors based on DAGs: the Pre-exposure criterion, the Common cause criterion, and the Modified disjunctive cause criterion. The aim is to equip researchers with a structured and theoretically grounded framework for identifying and selecting confounding factors, thereby improving the process of estimating causal effects in observational studies.

1.Dahabreh IJ, Bibbins-Domingo K. Causal inference about the effects of interventions from observational studies in medical journals[J]. JAMA, 2024, 331(21): 1845-1853. DOI: 10.1001/jama.2024.7741.

2.Martin L, Hutchens M, Hawkins C, et al. How much do clinical trials cost?[J]. Nat Rev Drug Discov, 2017, 16(6): 381-382. DOI: 10.1038/nrd.2017.70.

3.袁驰, 周祎灵, 曹雨滋, 等. 基于真实世界数据的观察性研究因果推断——目标试验模拟实施要点及案例分析[J]. 中国胸心血管外科临床杂志, 2024, 31(12): 1743-1752. [Yuan C, Zhou YL, Cao YZ, et al. Causal inference in observational studies based on real-world data: key points and case studies for target trial emulation[J]. Chinese Journal of Clinical Thoracic and Cardiovascular Surgery, 2024, 31(12): 1743-1752.] DOI: 10.7507/1007-4848.202407072.

4.Benson K, Hartz AJ. A comparison of observational studies and randomized, controlled trials[J]. N Engl J Med, 2000, 342(25): 1878-1886. DOI: 10.1056/NEJM200006223422506.

5.Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic research[J]. Epidemiology, 1999,10(1):37-48. https://pubmed.ncbi.nlm.nih.gov/9888278/.

6.Dickerman BA, García-Albéniz X, Logan RW, et al. Avoidable flaws in observational analyses: an application to statins and cancer[J]. Nat Med, 2019, 25(10): 1601-1606. DOI: 10.1038/s41591-019-0597-x.

7.Hammerton G, Munafò MR. Causal inference with observational data: the need for triangulation of evidence[J]. Psychol Med, 2021, 51(4): 563-578. DOI: 10.1017/S0033291720005127.

8.Pearl J, Glymour M, Jewell NP. Causal inference in statistics: a primer[M]. New Jersey: John Wiley & Sons, 2016: 1-160.

9.VanderWeele TJ, Shpitser I. On the definition of a confounder[J]. Ann Stat, 2013, 41(1): 196-220. DOI: 10.1214/12-aos1058.

10.Inoue K, Sakamaki K, Komukai S, et al. Methodological tutorial series for epidemiological studies: confounder selection and sensitivity analyses to unmeasured confounding from epidemiological and statistical perspectives[J]. J Epidemiol, 2025, 35(1): 3-10. DOI: 10.2188/jea.JE20240082.

11.Chan GCK, Sun T, Stjepanović D, et al. Designing observational studies for credible causal inference in addiction research-Directed acyclic graphs, modified disjunctive cause criterion and target trial emulation[J]. Addiction, 2024, 119(6): 1125-1134. DOI: 10.1111/add.16442.

12.郝允逸, 夏雪, 胥芹, 等. 有向无环图的构建规范: ESC-DAGs方法介绍及脑血管病研究案例解读[J]. 中国卒中杂志, 2024, 19(10): 1221-1229. [Hao YY, Xia X, Xu Q, et al. Construction specification of directed acyclic graphs:an introduction to the ESC-DAGs method and case interpretation in cerebrovascular disease research[J]. Chinese Journal of Stroke, 2024, 19(10): 1221-1229.] DOI: 10.3969/j.issn.1673-5765. 2024.10.016.

13.Feeney T, Hartwig FP, Davies NM. How to use directed acyclic graphs: guide for clinical researchers[J]. BMJ, 2025, 388: e78226. DOI: 10.1136/bmj-2023-078226.

14.Tennant PWG, Murray EJ, Arnold KF, et al. Use of directed acyclic graphs (DAGs) to identify confounders in applied health  research: review and recommendations[J]. Int J Epidemiol, 2021, 50(2): 620-632. DOI: 10.1093/ije/dyaa213.

15.姜弘扬. 面向生物医学数据的因果推断方法研究[D]. 长春: 吉林大学, 2024. DOI: 10.27162/d.cnki.gjlin.2024.007467.

16.刘慧鑫, 汪海波, 汪宁. 有向无环图在混杂因素识别与控制中的应用及实例分析[J]. 中华流行病学杂志, 2020, 41(4): 585-588. [Liu HX, Wang HB, Wang N. Application of directed acyclic graphs in identifying and controlling confounding bias[J]. Chinese Journal of Epidemiology, 2020, 41(4): 585-588.] DOI: 10.3760/cma.j.cn112338-20190729-00559.

17.刘子言, 吴小丽, 解美秋, 等. 在因果推断中应用有向无环图识别和控制选择偏倚[J]. 中华疾病控制杂志, 2019, 23(3): 351-355. [Liu ZY, Wu XY, Xie MQ, et al. Application of directed acyclic graphs in identification and control of selection bias in causal inference[J]. Chinese Journal of Disease Control & Prevention, 2019, 23(3): 351-355.] DOI: 10.16462/j.cnki.zhjbkz.2019.03.022.

18.Rubin DB. Should observational studies be designed to allow lack of balance in covariate distributions across treatment groups?[J]. Stat Med, 2009, 28(9): 1420-1423. https://doi.org/10.1002/sim.3565.

19.潘雨奇. 基于改进匹配算法的因果效应估计[D]. 武汉: 华中科技大学, 2023. DOI: 10.27157/d.cnki.ghzku.2023.000185.

20.北京爱琴海乐之技术有限公司. NoteExpress V3.0功能图解[Z]. 201436.

21.Cole SR, Platt RW, Schisterman EF, et al. Illustrating bias due to conditioning on a collider[J]. Int J Epidemiol, 2010, 39(2): 417-420. DOI: 10.1093/ije/dyp334.

22.Glymour MM, Weuve J, Chen JT. Methodological challenges in causal research on racial and ethnic patterns of  cognitive trajectories: measurement, selection, and bias[J]. Neuropsychol Rev, 2008, 18(3): 194-213. DOI: 10.1007/s11065-008-9066-x.

23.VanderWeele TJ, Shpitser I. A new criterion for confounder selection[J]. Biometrics, 2011, 67(4): 1406-1413. DOI: 10.1111/j.1541-0420.2011.01619.x.

24.Ding P, VanderWeele TJ, Robins JM. Instrumental variables as bias amplifiers with general outcome and confounding[J]. Biometrika, 2017, 104(2): 291-302. DOI: 10.1093/biomet/asx009.

25.杨宇琳, 沈瑞博, 李振翔, 等. 多不饱和脂肪酸与结直肠癌的关系: 一项两样本孟德尔随机化研究[J]. 数理医药学杂志, 2024, 37(7): 490-499. [Yang YL, Shen RB, Li ZX, et al. The relationship between polyunsaturated fatty acids and colorectal cancer: a two-sample Mendelian randomization study[J]. Journal of Mathematical Medicine, 2024, 37(7): 490-499.] DOI: 10.12173/j.issn.1004-4337.20240500.

26.VanderWeele TJ. Principles of confounder selection[J]. Eur J Epidemiol, 2019, 34(3): 211-219. DOI: 10.1007/s10654-019-00494-6.