基于SAS NLMIXED的广义线性混合效应模型在发病率数据Meta分析中的应用

doi:10.3969/j.issn.1673-5501.2019.02.010

摘要/Abstract

摘要： 目的介绍利用SAS软件中的PROC NLMIXED过程步实现发病率数据的META分析方法。方法基于广义线性混合效应模型(GLMM)的二项式-正态模型(BN)和泊松-正态模型(PNM)等,可方便地实现发病率数据的随机效应Meta分析,尤其当Meta分析纳入含0事件研究时。以Schutz等发表的血管内皮生长因子受体酪氨酸激酶抑制剂治疗的癌症患者发生致命不良事件风险的系统评价作为实例数据,利用SAS软件实现发病率数据的META分析,并提供编程代码。结果对于含0事件研究,使用PNM模型进行Meta分析,无需进行连续校正法。删除0事件研究对于PNM模型影响较大。与标准正态模型相比,PNM和BNM模型给出的效应值更高,而P值则更小,具有更好的灵敏性。结论基于广义线性混合效应模型,利用SAS的PROC NLMIXED实现发病率数据Meta分析是优选的方法。

关键词: 泊松-正态模型, 二项式-正态模型, 发病率数据, 广义线性混合效应模型, 正态-正态模型

Abstract: Objective To introduce the meta-analysis method for the incidence data using the PROC NLMIXED program in SAS software.Methods A Binomial-Normal model (BNM) or Poisson-Normal model (PNM) based on the generalized linear mixed model (GLMM) was proposed by Stijnen et al., which was extremely convenient to achieve a random-effect meta-analysis of incidence data, especially when the meta-analysis incorporated zero-event studies. A systematic review of the risk of fatal adverse events in cancer patients treated with vascular endothelial growth factor receptor tyrosine kinase inhibitors published by Schutz et al. was used as an example data, and meta-analysis of the incidence data was performed using SAS software and programming code was provided.Results For the zero-event study, the PNM model could be used for meta-analysis without continuity correction. The deletion of the zero-event study could have a greater impact on the PNM model. Compared with the standard normal model, the PNM model or BNM model gave higher effect values, while the P values were smaller, resulting in better sensitivity.Conclusion Based on the generalized linear mixed-effects model, using the PROC NLMIXED in SAS to achieve the meta-analysis of incidence data is the preferred method.

Key words: Binomial-normal model, Generalized linear mixed-effects model, Incidence rate data, Normal-normal model, Poisson-normal model

郑建清, 黄碧芬, 吴敏, 肖丽华. 基于SAS NLMIXED的广义线性混合效应模型在发病率数据Meta分析中的应用[J]. 中国循证儿科杂志, 2019, 14(2): 129-133.

ZHENG Jian-qing, HUANG Bi-fen, WU Min, XIAO Li-hua. The application of the generalized linear mixed model based on SAS NLMIXED in the meta-analysis of incidence rate data[J]. Chinese Journal of Evidence -Based Pediatric, 2019, 14(2): 129-133.

参考文献

[1] Spittal MJ, Pirkis J, Gurrin LC. Meta-analysis of incidence rate data in the presence of zero events. BMC Med Res Methodol, 2015, 15:42
[2] Hasselblad V, McCrory DC. Meta-analytic tools for medical decision making: a practical guide. Med Decis Making, 1995, 15(1):81-96
[3] Guevara JP, Berlin JA, Wolf FM. Meta-analytic methods for pooling rates when follow-up duration varies: a case study. BMC Med Res Methodol, 2004, 4:17
[4] Gibbons RD, Segawa E, Karabatsos G, et al. Mixed-effects Poisson regression analysis of adverse event reports: the relationship between antidepressants and suicide. Stat Med, 2008, 27(11):1814-1833
[5] Stijnen T, Hamza TH, Ozdemir P. Random effects meta-analysis of event outcome in the framework of the generalized linear mixed model with applications in sparse data. Stat Med, 2010, 29(29):3046-3067
[6] Bradburn MJ, Deeks JJ, Berlin JA, et al. Much ado about nothing: a comparison of the performance of meta-analytical methods with rare events. Stat Med, 2007, 26(1):53-77
[7] Sweeting MJ, Sutton AJ, Lambert PC. What to add to nothing? Use and avoidance of continuity corrections in meta-analysis of sparse data. Stat Med, 2004, 23(9):1351-1375
[8] Bhaumik DK, Amatya A, Normand SL, et al. Meta-Analysis of Rare Binary Adverse Event Data. J Am Stat Assoc, 2012, 107(498):555-567
[9] Schutz FA, Je Y, Richards CJ, et al. Meta-analysis of randomized controlled trials for the incidence and risk of treatment-related mortality in patients with cancer treated with vascular endothelial growth factor tyrosine kinase inhibitors. J Clin Oncol, 2012, 30(8): 871-877
[10] Sutton AJ, Higgins JP. Recent developments in meta-analysis. Stat Med, 2008, 27(5): 625-650
[11] Lane P W. Meta-analysis of incidence of rare events. Stat Methods Med Res, 2013, 22(2): 117-132
[12] Higgins JPT, Green S. Cochrane handbook for systematic reviews of interventions. West Sussex: The Cochrane Collabortion and John Wiley & Sons, Ltd, 2008
[13] Vandermeer B, Bialy L, Hooton N, et al. Meta-analyses of safety data: a comparison of exact versus asymptotic methods. Stat Methods Med Res, 2009, 18(4): 421-432
[14] Keus F, Wetterslev J, Gluud C, et al. Robustness assessments are needed to reduce bias in meta-analyses that include zero-event randomized trials. Am J Gastroenterol, 2009, 104(3): 546-551
[15] Kuss O. Statistical methods for meta-analyses including information from studies without any events-add nothing to nothing and succeed nevertheless. Stat Med, 2015, 34(7): 1097-1116