Education and debate

At the frontier of biomedical publication: Chicago 2005

BMJ 2005; 331 doi: 10.1136/bmj.331.7520.838 (Published 6 October 2005)
Cite this as: BMJ 2005;331:838

Access to the full text of this article requires a subscription or payment. Please log in or subscribe below.

  1. Kristina Fister, Roger Robinson editorial registrar1 (kfister@bmj.com)
  1. BMJ, London WC1H 9JR

    Last month the fifth congress on peer review and biomedical publication was held in Chicago. The presentations highlighted that we still have plenty of room to improve the quality of published research

    Introduction

    Evidence started to matter in biomedical publishing soon after it came to matter in medicine—relatively recently. The first international congress on peer review and biomedical publication was held in Chicago in 1989. At the time of the third congress, in 1997, only 146 original scientific articles had been published on peer review, of which 22 were prospective studies and 11 randomised controlled trials.1 Since then, the body of evidence has been growing, with about 200 abstracts indexed in Medline a year.2 We now have plenty of evidence to support the contention that peer review is “expensive, slow, subjective and biased, open to abuse, patchy at detecting important methodological defects, and almost useless at detecting fraud or misconduct.”3 The evidence on how to improve the process is scarce. What did the fifth congress add?

    Industry funding

    Some of the presented research looked into what happens when the pharmaceutical industry sponsors meta-analyses—the top of the hierarchy of evidence. Yank and colleagues analysed the agreement between results and conclusions in 71 meta-analyses of antihypertensive drugs published between 1966 and 2002.4 In about a third, authors disclosed financial ties with the pharmaceutical industry. Meta-analyses sponsored by industry were five times more likely than those funded by other sources to report conclusions favouring the study drug when such conclusions were not supported by the results. Meta-analyses funded by academic institutions showed no disagreement between the results and conclusions. Richard Smith, former editor of the BMJ, said: “It's a marvellous study and very disturbing.” This indicates an embarrassing editorial failure, commented Yank. But she refused to be drawn on the identity of …

    Access to the full text of this article requires a subscription or payment

    Subscribe

    Article access

    Article access for 1 day

    Purchase this article for £20 $30 €32*

    The PDF version can be downloaded as your personal record

    * Prices do not include VAT

    Rapid responses

    THIS WEEK'S POLL

    BMJ most popular