Commentary
Combining heterogenous studies using the random-effects model is a mistake and leads to inconclusive meta-analyses

https://doi.org/10.1016/j.jclinepi.2010.01.009Get rights and content

First page preview

First page preview
Click to open first page preview

References (28)

  • A.P. Verhagen et al.

    The influence of methodologic quality on the conclusion of a landmark meta-analysis on thrombolytic therapy

    Int J Technol Assess Health Care

    (2002)
  • S. Greenland et al.

    On the bias produced by quality scores in meta-analysis, and a hierarchical view of proposed solutions

    Biostatistics

    (2001)
  • G.F. Borm et al.

    The evidence provided by a single trial is less reliable than its statistical analysis suggests

    J Clin Epidemiol

    (2009)
  • S.A. Doi

    The influence of methodologic quality on the conclusion of a landmark meta-analysis on thrombolytic therapy

    Int J Technol Assess Health Care

    (2009)
  • Cited by (77)

    • Nonphysician Out-of-Hospital Rapid Sequence Intubation Success and Adverse Events: A Systematic Review and Meta-Analysis

      2017, Annals of Emergency Medicine
      Citation Excerpt :

      Meta-analyses of heterogeneous studies are often performed with the random-effects model26; however, our analysis was completed with the quality-effects model described by Doi et al27 and Doi and Thalib.28 The quality-effects model adjusts for study-level risk of bias and has advantages over the random-effects model, given that the latter model estimate does not allow direct interpretation.29 Also, the random-effects estimator suffers from faulty error estimation so that CIs generated are too narrow,30 and the random-effects model also exacerbates estimation of publication bias.31

    • A logician's approach to meta-analysis with unexplained heterogeneity

      2017, Journal of Biomedical Informatics
      Citation Excerpt :

      However, this random-effects approach is sometimes criticised. As advocated in [31], large heterogeneity in no way indicates that smaller studies should be more trusted nor that there is a fault in bigger studies. The approach proposed later in this paper preserves the weights of big studies and thus it should not be susceptible to such a criticism.

    View all citing articles on Scopus
    View full text