Authors' Reply


We thank Jørgensen et al.[1] for their interest in our paper.[2] We appreciate the points raised; however, we take issue with some of them. First, the dismissive view on repeating meta-analyses and hence perpetuating earlier errors. Evidence-based practice is not about creating evidence, rather it is about synthesising the best available evidence, which needs to be periodically updated. A case in point is the published meta-analyses by the groups of Jørgensen and Lamont,[3, 4] addressing questions reported in earlier multiple meta-analyses by others. Second, we re-read our paper to verify our alleged own admission of stating that the included studies were heterogeneous with respect to sample size and found no such statement. Variation in sample sizes of the included studies is not a recognised measure in assessing heterogeneity of studies included in a systematic review. Our colleagues seem to confuse grading the quality of the body of evidence for each outcome with the quality of included studies. Rating the quality of evidence refers not to individual studies but to the pooled body of evidence.[5]

We commend Jørgensen et al. for responding to a gap in research identified through systematic reviews (in this case the long-term effect of fetal exposure to antibiotics) by initiating a trial to address the issue. We hope that the end point of ‘poor outcome’ upon which they based power calculation is not a composite outcome; otherwise we will have to continue resorting to what the colleagues referred to as pooling small to medium-sized randomised controlled trials of only moderate quality for best available evidence. We believe that researchers should not lose sight of the fact that science, by its own nature, is cumulative.


  1. Top of page
  2. References