SEARCH

SEARCH BY CITATION

Keywords:

  • Meta-analysis;
  • Meta-regression;
  • Non-ignorable selection;
  • Reporting bias

Summary.  It is perhaps underappreciated that ruptured abdominal aortic aneurysm is a significant cause of mortality in the UK. The only curative treatment is an emergency operation and quantifying the success of this presents many difficulties. In particular, there is empirical evidence of reporting bias, suggesting that studies failing to report operating theatre mortality may be those where death in theatre is more common. We suggest a procedure for correcting for this bias and re-examine a recent meta-analysis of the available data. This casts considerable doubt on some conclusions from naïve analyses that do not take into account the potential bias. Perhaps most importantly, our procedure indicates a modest improvement in operating theatre mortality over the last 50 years, which is a trend that is not evident from the usual naïve analyses.