Medical journals need and use editorials for the same reasons that newspapers do. They provide a focus; they give direction; they stimulate interest. Editorials are part of the personality of the journal that will make readers like or loathe it, and either reaction is better than indifference. One of the more pernicious effects of evidence-based medicine has been encouragement of the belief that evidence is the sole basis for clinical decisions. This belief is coupled to denigration of the expert. In some circles, this has made the word expert more an insult than an accolade. There is a tendency for people who truly are experts in their fields to follow this politically correct line and to deny their own expertise.
For all sorts of reasons, these responses must be resisted. Medical evidence is too often too uncertain for treatments to be driven simply by statistical number-crunching. It is my opinion that too much weight is given to the results of meta-analysis, but I accept the danger of experts picking their evidence to suit their prejudged opinions . The basis of good medical advice comes from experts able to mix – to temper? – meta-analysis with experience. To accept this is to acknowledge that medical evidence – whether from a single clinical trial or from a synthesis of trials – is of value only when put into clinical context at the bedside, in the clinic, or in the operating room.
Medical evidence from clinical trials is becoming more controlled. Groups are sitting down and devising structures for single trials and for systematic reviews. The authority in medicine seems to be moving from clinicians or groups of clinicians deciding best treatments by a variety of implicit mechanisms (the tacit knowledge discussed, among others, by Eraut ), to groups who may not include clinicians at all laying down conditions for considering the evidence for treatments. For some time no reputable journal has published a clinical trial without a statement of research ethics approval; it may not be long before journals require proof that the conduct of a trial has followed the current consensus statement for the carrying out of clinical trials. But there will still be a need for experts experienced in whatever disease is under scrutiny . What better place for their thoughts than a journal editorial?
The European Journal of Anaesthesiology already publishes editorials. When asked to consider this one, I scanned the editorials for the previous year. They are as mixed a dozen as one could buy in a mixed wine offer in a supermarket, as well as having resonance with this discussion so far, and being relevant to the purpose of editorials. The topics included the Cochrane collaboration for reviewing and summarizing the results of clinical trials, and the CONSORT and QUORUM statements that present algorithms for clinical trials and for meta-analysis. Three political editorials asked who should provide anaesthesia, the future of the specialty, and the status of intensive care medicine. Two housekeeping editorials discussed how procedures should be coded, and clinical governance (which is best thought of in this context as an attempt to ensure clinical safety and quality). There was an editorial about the economics of anaesthesia, which could come under either of these headings. There were only three editorials directly discussing clinical practice, and one of the topics – a new carbon dioxide absorbent – cannot rely on ‘evidence’ in the same way that evidence is relied on for using one medical treatment rather than another. The only two properly clinical topics were the prophylaxis of thromboembolism, and opioid detoxification under anaesthesia.
All of these topics seem relevant and interesting. They provide information but give plenty of opportunity for disagreement. The political topics were aired in other journals too, and it is not the last time they will receive attention. The clinical topics were a long-running debate, and a controversial innovation. If I am critical, I wonder if thromboembolism is not better suited to a full review, with editorial comment upon it. Jefferson  believes that there is now no place for what he terms ‘desk drawer reviews’, i.e. reviewers just pulling together whatever they have to hand and basing a review upon it. That may be fair comment for a subject with as large a research base as thromboembolism, but Jefferson’s title was ‘What are the benefits of editorials and nonsystematic reviews?’, and he seems to want to remove opinion altogether.
I just do not think the world is like that. Even if we ignore simple clinical topics and the difficulties of gathering and interpreting evidence, there are many topics for which force of argument is more important than force of evidence. This brings me to the remaining editorial of the 12 months’ worth of EJA editorials, which was about medical reading habits . The methods and results sections of medical papers tend to be structured and somewhat formulaic, which (ignoring the tendency of medical writers to write impenetrable and tortuous English) should make them reasonably easy to understand. The editorialists suggested that structuring discussions would make them more easily understood as well. This subject has been aired before, in articles cited in the editorial [6,7]. I agree with Skelton and Edwards  that formally structuring discussions (and by implication editorials) is a step too far, but the whole purpose of an editorial is to introduce a topic, summarize important evidence or arguments, and reach a conclusion (or conclude there is not one to reach). There should be no need to impose structure on an editorial because – the equivalent of the literary essay – without a structure it cannot be an editorial.
To return to the worst excesses of evidence-based medicine, there is a temptation to try to remove human influence from medical research. Meta-analysis and megatrials involve tens or hundreds of medical and other people; sometimes there seem almost as many investigators as subjects. Perhaps there is loss of ownership of the research as well. The end-product can then be guidelines drawn up by yet more faceless committees. The occasional gross bias of commercial interest has led to the demand for admission of conflicts of interest from writers of reviews and editorials. Imposed disinterest may become lack of interest: I agree with Eger  that there are risks to removing the passion from research. It is up to readers to beware, and most doctors have a healthy enough scepticism that we should allow passion to remain.
There are dangers in editorials. Boba  complained that more than 10% of one journal was editorial comment, often with reference lists longer than the papers being commented on, and asked why the commentators did not submit their own articles and get them peer-reviewed.
Certainly, a common comment of referees is that discussion sections of submitted papers are too long, and it seems dishonest then to take the authors’ glory for an editorial. Feinstein, writing about epidemiology , complained that ‘too much attention has been given to the interpretive editorials and not enough to the basic scientific quality of the [original work]’. Feinstein was concerned about some fairly technical aspects of epidemiological method but he, like Boba, sees a problem, and we must not forget that a clinical editorial depends on the data and not the other way round.
There is the question whether editorials should be peer reviewed, which one editor (see comment to ) answered in the way I suspect most editors would: the constraint of time. Waiting for peer review of editorial comment on a topical subject could make the subject old news by the time of publication.
All sorts of other questions can be asked about editorials, and in the end it depends on the editor. He or she can do what they like with editorials, and that is their appeal both for the editor and for the readers. For the New England Journal of Medicine, Kassirer and Angell  see no reason why they should publish both sides of every story in editorials. They are clear that an editorial is the view of its author, and wonder if a journal as a corporate body can have ‘an opinion’. With the current increase in litigation, it is important that editorials do not attain a spurious legal authority, and there have been some recent forceful editorials in anaesthesia journals that would not represent majority opinion (although it has to be said that there is no reason why majority opinion is correct: democracy does not define truth).
There is no strict definition of what makes an editorial. Editorials fall somewhere between a long letter and a review, and the only constant feature is that they rarely contain primary data. The EJA commissions most of its editorials, but we encourage anyone who can muster a succinct argument to do so and submit. I agree that scientific editorials are ‘a precious and scarce element in medical journals’ , although they may now be more common than in 1985, when that editorial was written. That same journal later published editorial advice on how to write an editorial .
What does the editorialist gain? In private correspondence, I have been told that there is never need to cite an editorial, no matter how good. This is a mindset that holds nothing of value unless it can be measured. An editorial should not be cited as a lazy way of citing the original research but, if an editorial puts forward a new idea or interpretation, the idea or interpretation belongs to the editorial writer as much as clinical data belong to the researcher, and should be acknowledged. That it is not, and that assessment committees pay less attention to editorials than to research papers, is part of Morgan’s complaint  that it is difficult to get researchers to write editorials. For the moment, though, it is a fact of life and, unfashionable though it may be, the editorialist’s reward is likely to be (apart perhaps from a small fee) just personal satisfaction.
1 Silverman WA. Crap-trapping. Lancet
1997; 349: 1471–1472.
2 Eraut M. Developing Professional Knowledge and Competence
. London, UK: Falmer, 1994.
3 Walker A. Meta-style and expert review. Lancet
1999; 354: 1834–1835.
4 Jefferson T. What are the benefits of editorials and non-systematic reviews? BMJ
1999; 318: 135–135.
5 Snelgrove H, McAnulty G. A glance at medical reading habits in Italy and England: is the writing on the wall? Eur J Anaesthesiol
2000; 17: 717–719.
6 Docherty M, Smith R. The case for structuring the discussion of scientific papers. BMJ
1999; 318: 1224–1225.
7 Skelton JR, Edwards SJL. The function of the discussion section in academic medical writing. BMJ
2000; 320: 1269–1270.
8 Eger EI. Conflicts of interest: are they a problem for anaesthesia journals? Br J Anaesth
2001; 86: 734–734.
9 Boba A. On editorial comments. Lancet
1997; 350: 1635–1636.
10 Feinstein AR. Scientific news and epidemiological editorials: a reply to the critics. Epidemiology
1990; 1: 170–180.
11 Neel JV. Should editorials also be peer-reviewed? Am J Hum Genet
1988; 43: 981–982.
12 Kassirer JP, Angell M. Controversial Journal editorials. New Engl J Med
1997; 337: 1460–1461.
13 Morgan PP. Scientific editorials – a precious and scarce element in medical journals. Can Med Assoc J
1985; 132: 315–315.
14 Squires BP. Editorials and platform articles: What editors want from authors and peer reviewers. Can Medical Assoc J
1989; 141: 666–667.