Relation between treatment benefit and underlying risk in meta-analysis

BMJ 1996; 313 doi: https://doi.org/10.1136/bmj.313.7071.1550a (Published 14 December 1996) Cite this as: BMJ 1996;313:1550

Standard of “label invariance” should not be abandoned

  1. Stephen Senn
  1. Professor of pharmaceutical and health statistics University College London, London WC1E 6BT

    EDITOR,—Stephen J Sharp and colleagues' discussion of the dangers of attempting to compare treatment benefit with underlying risk in meta-analyses of placebo controlled trials of a presumed active treatment is excellent in two respects.1 Firstly, the authors clearly show a point that I have made—that when the difference in mean observed outcome between groups is related to the mean outcome in the placebo group then a spurious correlation is induced.2 Secondly, they make the important point that it …

    View Full Text

    Log in

    Log in through your institution


    * For online subscription