Editor—Docherty and Smith invite comments on their proposal to structure the discussion section of original scientific papers.1 The proposal, although not new, does not go far enough to address the quality problems of contemporary published research.
The key element that is missing from the proposal is the explicit insistence that “reports of primary research should begin by referring, in the introduction, to the results of a systematically conducted, published review of the results of relevant previously reported studies, explaining why the current study was justified. The discussion section of the report should set the data generated by the new study in the context of all the evidence available at the time of analysis.”2
That proposal, made by Chalmers back in 1991, should be explicitly incorporated in Docherty and Smith’s proposal. The main obstacle to such an approach so far has been the existing tension between the need for publishing sound science and the business survival of most journals, which depend on a steady stream of papers to continue operating. Old habits die hard. Clarke and Chalmers found little evidence that journals adequately implemented CONSORT recommendations that results of randomised controlled trials should be discussed in the light of all available evidence.3 Colleagues and I found editorial change equally difficult to achieve in the field of economic evaluation.4
It is no longer acceptable for authors to publish research that may have an impact on people’s lives without attempting to interpret the results within the available body of systematically collected and evaluated knowledge. Studies that are likely to have a major impact on the way that health care is delivered and financed should be editorially peer reviewed by being compared with their “peer studies” in terms of methods, topic, or results.5
The absence of a systematic review on a topic (which is becoming less likely, as the methods and uses of systematic reviews become accepted) is not an excuse for presenting information from a single study without making any reference to its peer studies. As Docherty and Smith point out, the practice of making such a reference should help minimise bias in discussion sections.
References
- 1.Docherty M, Smith R. The case for structuring the discussion of scientific papers. BMJ. 1999;318:1224–1225. doi: 10.1136/bmj.318.7193.1224. . (8 May.) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Chalmers I. Improving the quality and dissemination of reviews of clinical research. In: Lock SP, editor. The future of medical journals: in commemoration of 150 years of the British Medical Journal. London: BMJ Publishing Group; 1991. [Google Scholar]
- 3.Clarke M, Chalmers I. Discussion sections in reports of controlled trials published in general medical journals. Islands in search of continents? JAMA. 1998;280:280–282. doi: 10.1001/jama.280.3.280. [DOI] [PubMed] [Google Scholar]
- 4.Jefferson TO, Drummond MF, Smith R, Yi Y, Pratt M, Kale R. Evaluating the BMJ guidelines on economic submissions—prospective audit of economic submissions to the BMJ and Lancet. JAMA. 1998;280:275–277. doi: 10.1001/jama.280.3.275. [DOI] [PubMed] [Google Scholar]
- 5.Jefferson TO, Deeks JJ. The use of systematic reviews for peer-reviewing. A population approach. In: Godlee F, Jefferson TO, eds. Peer reviews in biomedical science. London: BMJ Books (in press).