Skip navigation

Mindstretcher: Abstracts should carry health warnings

We all manage to try and keep up-to-date with the literature by skimming through the abstracts of papers, and although Bandolier seeks to provide help and support for its readers it sometimes has to reveal unpalatable truths that disturb the reader's equanimity.

It is almost 10 years since the proposal for more informative abstracts of clinical articles was published [1]. Yet still many articles do not carry the type of structured abstract recommended by the working group for critical appraisal of medical literature.

An excellent study of "Methodology and overt hidden bias in reports of 196 double-blind trials of non-steroidal anti-inflammatory drugs in rheumatoid arthritis" by Peter Gøetzche published in 1989 [2] reported a number of defects in the trials reviewed. The most worrying statement was that "doubtful or invalid statements were found in 76% of the conclusions or abstracts. Bias consistently favoured the new drug in 81 trials and the control in only one trial."

This message was reinforced by the article about assessing quality of reports [3] which was reviewed in Bandolier 17 . The additional issue of the need to beware relative risk reduction as a way of presenting information was described in detail in Bandolier 21 . Rules for abstract reading should therefore be:
  1. If the abstract is not structured, be very cautious.
  2. If the abstract of a trial does not describe randomisation, beware.
  3. If the results are only expressed in terms of 'p' values or relative risk reductions, watch out.

References:

  1. Ad hoc working group for critical appraisal of medical literature. A proposal for more informative abstracts of clinical articles. Annals of Internal Medicine 1987 106: 598-604.
  2. PC Gøetzche. Methodology and overt and hidden bias in reports of 196 double-blind trials of nonsteroidal anti-inflammatory drugs in rheumatoid arthritis. Controlled Clinical Trials 1989 10: 31-56.
  3. KF Schultz, I Chalmers, DA Grimes, DG Altman. Assessing the quality of randomization from reports of controlled trials published in Obstetrics and Gynecology journals. Journal of the American Medical Association 1994 272: 125-128.



previous or next story in this issue