Bayesian mixed treatment comparisons meta-analysis for correlated outcomes subject to reporting bias

被引:0
作者
Liu, Yulun [1 ]
DeSantis, Stacia M. [2 ]
Chen, Yong [1 ]
机构
[1] Univ Penn, Philadelphia, PA 19104 USA
[2] Univ Texas Hlth Sci Ctr Houston, Houston, TX 77030 USA
基金
美国国家卫生研究院;
关键词
Bayesian model; Mixed treatment comparison; Multivariate meta-analysis; Network meta-analysis; Publication bias; Systematic review; MULTIVARIATE METAANALYSIS; NETWORK METAANALYSIS; SENSITIVITY-ANALYSIS; PUBLICATION BIAS; JOINT SYNTHESIS; MODEL; ALCOHOLISM;
D O I
暂无
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
Many randomized controlled trials report more than one primary outcome. As a result, multivariate meta-analytic methods for the assimilation of treatment effects in systematic reviews of randomized controlled trials have received increasing attention in the literature. These methods show promise with respect to bias reduction and efficiency gain compared with univariate meta-analysis. However, most methods for multivariate meta-analysis have focused on pairwise treatment comparisons (i.e. when the number of treatments is 2). Current methods for mixed treatment comparisons meta-analysis (i.e. when the number of treatments is more than 2) have focused on univariate or, very recently, bivariate outcomes. To broaden their application, we propose a framework for mixed treatment comparisons meta-analysis of multivariate (two or more) outcomes where the correlations between multivariate outcomes within and between studies are accounted for through copulas, and the joint modelling of multivariate random effects respectively. We consider a Bayesian hierarchical model using Markov chain Monte Carlo methods for estimation. An important feature of the framework proposed is that it allows for borrowing of information across correlated outcomes. We show via simulation that our approach reduces the effect of outcome reporting bias in a variety of missing outcome scenarios. We apply the method to a systematic review of randomized controlled trials of pharmacological treatments for alcohol dependence, which tends to report multiple outcomes potentially subject to outcome reporting bias.
引用
收藏
页码:127 / 144
页数:18
相关论文
共 42 条
  • [1] Network meta-analysis of multiple outcome measures accounting for borrowing of information across outcomes
    Achana, Felix A.
    Cooper, Nicola J.
    Bujkiewicz, Sylwia
    Hubbard, Stephanie J.
    Kendrick, Denise
    Jones, David R.
    Sutton, Alex J.
    [J]. BMC MEDICAL RESEARCH METHODOLOGY, 2014, 14
  • [2] [Anonymous], COCH DATABASE SYST R
  • [3] [Anonymous], 2021, Bayesian data analysis
  • [4] [Anonymous], 2008, SYSTEMATIC REV HLTH
  • [5] [Anonymous], 2002, STAT INFERENCE
  • [6] Simultaneous comparison of multiple treatments: combining direct and indirect evidence
    Caldwell, DM
    Ades, AE
    Higgins, JPT
    [J]. BMJ-BRITISH MEDICAL JOURNAL, 2005, 331 (7521): : 897 - 900
  • [7] Graphical Tools for Network Meta-Analysis in STATA
    Chaimani, Anna
    Higgins, Julian P. T.
    Mavridis, Dimitris
    Spyridonos, Panagiota
    Salanti, Georgia
    [J]. PLOS ONE, 2013, 8 (10):
  • [8] Using network meta-analysis to evaluate the existence of small-study effects in a network of interventions
    Chaimani, Anna
    Salanti, Georgia
    [J]. RESEARCH SYNTHESIS METHODS, 2012, 3 (02) : 161 - 176
  • [9] Chen M.-H., 2000, Monte Carlo Methods in Bayesian Computation
  • [10] An alternative pseudolikelihood method for multivariate random-effects meta-analysis
    Chen, Yong
    Hong, Chuan
    Riley, Richard D.
    [J]. STATISTICS IN MEDICINE, 2015, 34 (03) : 361 - 380