Page 516 Acute Pain Management
P. 516
Of the references listed in the May 2009 retraction notice (Shafer 2009), four were included in
the second edition of Acute Pain Management: Scientific Evidence along with a further two
publications that were not included in this list of retractions.
There are no precedents for how best to manage a problem such as this. However, the editors
responsible for the development of this third edition of Acute Pain Management: Scientific
Evidence decided against including any publications by the individuals affected by these
retractions when listed as first author on the papers. An assessment was made of each of the
meta‐analyses that cited affected articles. This was based upon the other papers included in
these meta‐analyses, other supporting evidence and independent consideration by an expert
in biostatistics. In some cases, although cited, the affected references were not actually
included in the meta‐analysis performed. In other cases, assessment indicated that the
strength of the evidence may be reduced because of the inclusion of affected publications.
Following the consensus that appeared to rapidly emerge among editors of the leading peer‐
reviewed journals in anaesthesiology and pain medicine despite initial concerns about meta‐
analyses that included this work (White et al 2009), the editors of the third edition of Acute
Pain Management: Scientific Evidence felt that indiscriminately omitting all meta‐analyses
purely on the basis of inclusion of one or two of those papers would be to deny inclusion of
some important credible information in the document. Indeed, the purpose of meta‐analysis is
to aggregate results from the literature as a whole, thereby diluting the impact of any one
specific study.
Just prior to finalisation of this third edition of Acute Pain Management: Scientific Evidence, an
article was published in Anesthesiology in December 2009 (Marret et al, 2009) which examined
in detail the effect that excluding data obtained from the retracted articles would have on the
results of 14 systematic reviews (six quantitative and eight qualitative) in which they were
cited. Marret et al (Marret et al, 2009) reanalysed the data after excluding results from affected
articles and concluded that withdrawal of these articles did not alter the conclusions of five
out of the six quantitative reviews (meta‐analyses): the sixth meta‐analysis has not been
included in Acute Pain Management: Scientific Evidence. Thus there was agreement with the
assessments that had already made about the validity of these meta‐analyses which included
the retracted articles. Marret et al (Marret E et al, 2009) concluded that meta‐analyses were
‘vulnerable’ if data from retracted studies made up more than 30% of the total.
A footnote has been added to the relevant sections indicating the systematic reviews
(quantitative and qualitative) that includes affected articles along with a summary of the
effect, if any, on the results obtained. Also, specific note has been made in the text of the third
edition of Acute Pain Management: Scientific Evidence where retraction of the affected papers
involved key messages that were published in the second edition. Should additional
information become available it will be added as needed before publication of this document.
Information that comes to light after publication will be posted as appropriate on the Acute
Pain Management: Scientific Evidence website.
APPENDIX B Cost analyses
The area of acute pain management remains remarkably deficient in research on costs and
cost‐benefit. Where this information was available it was reported. One obvious example is
the costs associated with the adverse effects of treatment. Information to assist clinicians to
better manage both pain and some of the adverse effects of treatment, as well as better
individualise treatment for each patient, may assist in minimising such costs. This is again
noted as an area warranting further research.
468 Acute Pain Management: Scientific Evidence
