Quantifying heterogeneity in individual participant data meta-analysis with binary outcomes

被引:93
作者
Chen, Bo [1 ]
Benedetti, Andrea [1 ,2 ]
机构
[1] McGill Univ, Dept Epidemiol Biostat & Occupat Hlth, Purvis Hall,1020 Pine Ave West, Montreal, PQ, Canada
[2] McGill Univ, Resp Epidemiol & Clin Res Unit, 2155 Guy St 4th Floor,Off 412,24105, Montreal, PQ, Canada
基金
加拿大健康研究院;
关键词
Individual participant datameta-analysis (IPD-MA); Heterogeneity; Two-stage and one-stage approaches; I-2; LINEAR MIXED MODELS; INTRACLASS CORRELATION; PATIENT DATA;
D O I
10.1186/s13643-017-0630-4
中图分类号
R5 [内科学];
学科分类号
1002 ; 100201 ;
摘要
Background: In meta-analyses (MA), effect estimates that are pooled together will often be heterogeneous. Determining how substantial heterogeneity is is an important aspect of MA. Method: We consider how best to quantify heterogeneity in the context of individual participant data meta-analysis (IPD-MA) of binary data. Both two-and one-stage approaches are evaluated via simulation study. We consider conventional I-2 and R-2 statistics estimated via a two-stage approach and R-2 estimated via a one-stage approach. We propose a simulation-based intraclass correlation coefficient (ICC) adapted from Goldstein et al. to estimate the I-2, from the one-stage approach. Results: Results show that when there is no effect modification, the estimated I-2 from the two-stage model is underestimated, while in the one-stage model, it is overestimated. In the presence of effect modification, the estimated I-2 from the one-stage model has better performance than that from the two-stage model when the prevalence of the outcome is high. The I-2 from the two-stage model is less sensitive to the strength of effect modification when the number of studies is large and prevalence is low. Conclusions: The simulation-based I-2 based on a one-stage approach has better performance than the conventional I-2 based on a two-stage approach when there is strong effect modification with high prevalence.
引用
收藏
页数:9
相关论文
共 29 条
[1]   The exact distribution of Cochran's heterogeneity statistic in one-way random effects meta-analysis [J].
Biggerstaff, Brad J. ;
Jackson, Dan .
STATISTICS IN MEDICINE, 2008, 27 (29) :6093-6110
[2]  
Brewlow N.D. C., 1993, Journal of the American Statistical Association, V88, P421
[3]   Variance partitioning in multilevel logistic models that exhibit overdispersion [J].
Browne, WJ ;
Subramanian, SV ;
Jones, K ;
Goldstein, H .
JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES A-STATISTICS IN SOCIETY, 2005, 168 :599-613
[4]   Individual Participant Data Meta-Analysis for a Binary Outcome: One-Stage or Two-Stage? [J].
Debray, Thomas P. A. ;
Moons, Karel G. M. ;
Abo-Zaid, Ghada Mohammed Abdallah ;
Koffijberg, Hendrik ;
Riley, Richard David .
PLOS ONE, 2013, 8 (04)
[5]   METAANALYSIS IN CLINICAL-TRIALS [J].
DERSIMONIAN, R ;
LAIRD, N .
CONTROLLED CLINICAL TRIALS, 1986, 7 (03) :177-188
[6]  
Diggle PJ., 2002, ANAL LONGITUDINAL DA, V2nd ed., P2002
[7]   Factors explaining the heterogeneity of effects of patient decision aids on knowledge of outcome probabilities: a systematic review sub-analysis. [J].
Gentles S.J. ;
Stacey D. ;
Bennett C. ;
Alshurafa M. ;
Walter S.D. .
Systematic Reviews, 2 (1) :95
[8]  
Goldstein H., 2002, Understanding Statistics, V1, P223, DOI 10.1207/S15328031US0104_02
[9]   Meta-analysis of continuous outcome data from individual patients [J].
Higgins, JPT ;
Whitehead, A ;
Turner, RM ;
Omar, RZ ;
Thompson, SG .
STATISTICS IN MEDICINE, 2001, 20 (15) :2219-2241
[10]   Quantifying heterogeneity in a meta-analysis [J].
Higgins, JPT ;
Thompson, SG .
STATISTICS IN MEDICINE, 2002, 21 (11) :1539-1558