| 64 | 0 | 109 |
| 下载次数 | 被引频次 | 阅读次数 |
系统梳理中医药队列研究中永恒时间偏倚的常见类型及其原因，认为永恒时间偏倚主要包括错误分类与错误排除两类，其产生的原因是研究对象从入组到接受干预的时间间隔内产生了结局事件，这类患者被错误地归入某组或被错误地排除出结局观察。基于上述分析，阐述了控制永恒时间偏倚的具体方法，如研究设计方面，可采用巢式病例对照研究和目标试验模拟；统计学方法方面，时间依赖的COX回归模型、Poisson回归模型、敏感性分析、界标分析可较为有效地监测和控制永恒时间偏倚；硬件技术方面，新的检测技术以及智能手环等可穿戴设备、植入式芯片等连续生物传感器、移动健康应用程序等新技术手段可从根本上缩短永恒时间，减少永恒时间偏倚。通过以上方式，可为永恒时间偏倚的良好控制和中医药高水平临床研究的开展提供方法学借鉴。
Abstract：This article systematically reviewed the common types and causes of immortal time bias in traditional Chinese medicine(TCM) cohort studies. It is suggested that immortal time bias mainly includes misclassification and erroneous exclusion, which occur when outcome events happen during the time interval between participant enrollment and receiving the intervention. These patients are incorrectly assigned to a group or erroneously excluded from outcome observation. Based on the above analysis, specific methods for controlling immortal time bias are discussed. In terms of study design, nested case-control studies and target trial simulations can be used. Regarding statistical methods, time-dependent Cox regression models, Poisson regression models, sensitivity analysis, and landmark analysis can effectively monitor and control immortal time bias. In terms of hardware technology, new detection technologies and wearable devices such as smart wristbands, continuous biosensors like implantable chips, and mobile health applications can fundamentally shorten immortal time and reduce immortal time bias. By using the abovementioned methods, it will provide methodological references for the effective control of immortal time bias and the development of high-quality clinical research in TCM.
[1]LÉVESQUE L E, HANLEY J A, KEZOUH A, et al. Problem of immortal time bias in cohort studies:example using statins for preventing progression of diabetes[J]. BMJ,2010, 340:b5087. doi:10. 1136/bmj. b5087.
[2]YADAV K, LEWIS R J. Immortal time bias in observational studies[J]. JAMA, 2021, 325(7):686-687.
[3]VELDERS B J J, VRIESENDORP M D, GROENWOLD R H H. Preventing immortal time bias in observational studies:a matter of design[J]. Eur Heart J Open, 2023,3(6):oead117.
[4]龙子临，赵厚宇，廖星，等.基于Bibliometrix分析2017—2022年国内外中医药疗效评价队列研究的现状与趋势[J].中医杂志，2024,65(7):737-744.
[5]曾琳，卓琳，赵一鸣.永恒时间偏倚[J].中华儿科杂志，2021,59(2):149.
[6]林翠清，段玉婷，陆丽明.队列研究在中医药领域的应用[J].中华中医药学刊，2021,39(2):189-193.
[7]YANKEY B A, ROTHENBERG R, STRASSER S, et al.Effect of marijuana use on cardiovascular and cerebrovascular mortality:a study using the National Health and Nutrition Examination Survey linked mortality file[J]. Eur J Prev Cardiol, 2017, 24(17):1833-1840.
[8]高武霖，戴国华，吴彬，等.中医药干预心肌梗死合并高脂血症患者的队列研究[J].中华中医药杂志，2018,33(5):2002-2006.
[9]HUBBARD R A, GATSONIS C A, HOGAN J W, et al."Target Trial Emulation"for observational studies:potential and pitfalls[J]. N Engl J Med, 2024, 391(21):1975-1977.
[10]GUPTA S, WANG W, HAYEK S S, et al. Association between early treatment with tocilizumab and mortality among critically ill patients with COVID-19[J]. JAMA Intern Med, 2021, 181(1):41-51.
[11]REMAP-CAP Investigators, GORDON A C, MOUNCEY P R, et al. Interleukin-6 receptor antagonists in critically ill patients with COVID-19[J]. N Engl J Med, 2021,384(16):1491-1502.
[12]叶冬青.巢式病例对照研究的设计及分析[J].疾病控制杂志，2001,5(1):65-68.
[13]XIA L, NAN B, LI Y. Statistical inference for COX proportional hazards models with a diverging number of covariates[J]. Scand Stat Theory Appl, 2023, 50(2):550-571.
[14]GELLAR J E, COLANTUONI E, NEEDHAM D M,et al. COX regression models with functional covariates for survival data[J]. Stat Modelling, 2015, 15(3):256-278.
[15]LEE M, TROXEL A B, KWON S, et al. Partial-linear single-index COX regression models with multiple timedependent covariates[J]. BMC Med Res Methodol,2024, 24(1):311.
[16]YELLAND L N, SALTER A B, RYAN P. Performance of the modified Poisson regression approach for estimating relative risks from clustered prospective data[J].Am J Epidemiol, 2011, 174(8):984-992.
[17]ZOU G. A modified poisson regression approach to prospective studies with binary data[J]. Am J Epidemiol,2004, 159(7):702-706.
[18]REA F, MORABITO G, CORRAO G, et al. The hazard of using the Poisson model to cope with immortal time bias in the case of time-varying hazard[J]. BMC Med Res Methodol, 2024, 24(1):272.
[19]MOORE D B, FOLSOM A R, MINK P J, et al. Physical activity and incidence of postmenopausal breast cancer[J]. Epidemiology, 2000, 11(3):292-296.
[20]HERNÁN M A, ROBINS J M. Estimating causal effects from epidemiological data[J]. J Epidemiol Community Health, 2006, 60(7):578-586.
[21]HERNÁN M A, SAUER B C, HERNÁNDEZ-DÍAZ S,et al. Specifying a target trial prevents immortal time bias and other self-inflicted injuries in observational analyses[J]. J Clin Epidemiol, 2016, 79:70-75. doi:10. 1016/j. jclinepi. 2016. 04. 014.
[22]ZHANG X, GUO Q, LI C, et al. Immortal time biascorrected effectiveness of traditional Chinese medicine in non-small cell lung cancer(C-EVID):a prospective cohort study[J]. Front Oncol, 2022, 12:845613. doi:10. 3389/fonc. 2022. 845613.
[23]ANDERSON J R, CAIN K C, GELBER R D. Analysis of survival by tumor response[J]. J Clin Oncol, 1983, 1(11):710-719.
[24]DAFNI U. Landmark analysis at the 25-year landmark point[J]. Circ Cardiovasc Qual Outcomes, 2011, 4(3):363-371.
[25]MORGAN C J. Landmark analysis:a primer[J]. J Nucl Cardiol, 2019, 26(2):391-393.
[26]DE CAROLIS E, VELLA A, VACCARO L, et al. Application of MALDI-TOF mass spectrometry in clinical diagnostic microbiology[J]. J Infect Dev Ctries, 2014, 8(9):1081-1088.
[27]ZHU H, FOHLEROVÁZ, PEKÁREK J, et al. Recent advances in lab-on-a-chip technologies for viral diagnosis[J]. Biosens Bioelectron, 2020, 153:112041. doi:10. 1016/j. bios. 2020. 112041.
[28]章浩月，叶红华.可穿戴设备在心力衰竭长期管理中的应用进展[J].临床荟萃，2024,39(10):945-952.
基本信息:
DOI：10.13288/j.11-2166/r.2026.08.006
中图分类号:R2-03
引用信息:
[1]陈俞含,李润庭,袁滋艺,等.中医药队列研究中永恒时间偏倚的识别与控制策略[J].中医杂志,2026,67(08):838-844.DOI:10.13288/j.11-2166/r.2026.08.006.
基金信息:
国家自然科学基金（82374298）; 北京中医药大学后备学科带头人资助项目（90010960920033）
2026-04-16
2026-04-16