Richard Smith: What are medical journals for and how well do they fulfil those functions?

richard_smith_2014Last week I gave a talk to the International Society of Medical Publication Professionals entitled “Medical journals: time for something different.” My core argument was that “Medical journals have played an important role in spreading medical knowledge, but they are now beset with problems. Some will transform, most will disappear. New forms of disseminating medical knowledge are appearing and will continue to appear.” As part of the talk I amused myself by reflecting on the functions of medical journals and then scoring them for how well they do with each function.

I promised the audience that I would write a blog on my scores, and this is it. If you read on you’ll see that my scores are mostly low, and the low scores are partially attributable to me giving the scores in the context of what the modern world, particularly information technology, has to offer.

Primary functions:

1. Publishing science: this is what most of the 20 000-30 000 medical journals do. Overall score 2/10.

a. Selecting high quality research most relevant to readers is impossible when a million or more studies are published a year and editors have no more than 10 000 (and usually far fewer to choose among). The overlap between what would be optimal for readers and what journals are sent must be minimal. Worse there is likely to be bias, particularly in some journals, towards work funded by drug companies. 2/10.

b. Quality assuring research is poor as we know that most published findings are false, and it’s not surprising as peer review is fraught with problems and lacks evidence of effectiveness.

c. Organising medical research: This is impossible for journals when a million studies are published a year, and in fact journals disorganise research by breaking studies up into multiple journals, many of them inaccessible to most people. – 2/10.

d. Promoting high quality research: Unfortunately studies published in high impact journals are more likely to be wrong than those published in lesser journals. And some of studies funded by drug companies and published in major journals are more about marketing than clinical practice. Plus clinicians should never change practice on the basis of one study. 3/10.

2. Education: words on paper alone change nothing and are as poor a form of education as lectures (and at least in some lectures you can ask questions). 4/10.

3. Reference: The Times (of London) once hoped to provide readers with a newspaper of reference, recording “everything that mattered.” This is now impossible for journals, even those with the narrowest subjects. 2/10.

4. Informing (news): most journals don’t attempt this, and even for those with websites updated daily it’s hard to keep up with better resourced news gathering operations. 3/10.

5. Providing a forum: This was once an important function for journals with, for example, The BMJ being the only communication that regularly reached most British doctors, allowing them to debate positions on clinical, ethical, and political issues. But social media now allow much better ways to do this. 4/10.

6. Campaigning/reform/”speaking truth to power”: journals can do this well, and there are many examples of leading journals campaigning on major issues—like “babyfarming,” quack medicines, nuclear war, inequalities, and much else. 7/10 (but very few journals have the resources to do this effectively).

7. Investigating/disclosing malpractice: again this is something that journals can do effectively, as The BMJ has done recently. 7/10 (but again few journals have the resources, including legal back up, to do this well).

8. Agenda setting/legitimising: evidence from media studies shows that the media, including journals, are good at putting issues on the agenda—like, for example, prison medicine, climate change, corruption. Although “good at telling people what to think about they are less good at telling them what to think,” which, I think, is a good thing. 9/10 (but again few journals attempt this).

9. Leading: whether journals can lead or are doomed simply to reflect what others are making happen is debated, but I judge that The Lancet has become an important leader in global health—just as 150 years ago it led on reforming British medicine. 5/10 (but again it’s only a handful of journals that attempt this).

10. Entertaining: Richard Asher famously asked “Why are medical journals so dull?” In the “attention economy” journals compete with Hollywood films, Premier League football, video games, and other entertainments much more entertaining than they are. 2/10.

Secondary functions

11. Scoring the performance of academics: this was never intended to be a function of medical journals, but ironically it is the function that keeps them alive: academics are desparate to publish in high impact journals. The journals perform the function badly, however, because there is virtually no correlation between the impact factor of a journal and citations to articles published in the journal—because the impact factor is driven by a few highly cited articles. So it is unscientific to use the impact factor of a journal as a surrogate for the impact/quality of an individual article, but universities, funding organisations, and others continue to do it. I find it extraordinary that universities should outsource to a publishing system the core function of measuring the performance of their academics. 2/10.

12. Promoting drugs for pharma companies: major journals can do this every effectively, which is why they can make large sums from selling reprints of the articles to the companies who funded the research. The implicit message is “the XXX journal supports this study (and so drug).” 7/10. (but for only some journals)

13. Making money: because the main value in the journal (the research) is delivered free to publishers some journals are very profitable and scientific journals overall are more profitable than almost any other business (30% profit margin versus 5% for, say, supermarkets). It’s as if oil companies could sell petrol without having to go to the expense of discovering oil, extracting it from the ground, transporting it, and distilling it. 10/10.

14. Providing employment: Scientific journals provide tens of thousands of often enjoyable and sometimes well paid jobs. 9/10.

Conclusion:

Medical journals are now a poor way to publish science, but they can perform well the functions traditionally associated with the mass media—campaigning, investigating, agenda setting, and even leading. The second (and major) part of my talk was on the problems besetting journals, and my prediction is that science will soon be published in a different way but that a few journals might continue with the functions that journals can perform well.

But I’m not clear who will pay for these functions, and I can’t see journals continuing to make 30% profit margins. Because there are money and jobs to lose and because universities and others continue to misguidedly use publications as a means of judging the performance of academics journals may take some time to fade way.

Richard Smith was the editor of The BMJ until 2004. 

Competing interest: RS spent 25 years working for The BMJ and his pension comes from the BMA, which has The BMJ as a source of income.

  • susanne stevens

    There are 60-70 journals published by BMJ Many ‘user-public-patient’ groups publish journals or newsletters but only on specific topics. Some use on line more than others. But as journals have not yet had their day is there not room for another BMJ journal – The BMJ Public -Patient/User’ journal. Which would need to be edited and controlled by them with contributions by respondents as in TheBMJ . Although The BMJ medical journal does make space for ‘the patients voice’, including through rr’s but it is quite a minimal space.

  • Jacob Barhak

    What about just aggregating data? It may fall in point 1 and partially point 3 , yet considering new technology it may be a new use. For example it may be useful for the future use considering machine learning work like IBM Watson, or even clinicaltrials.gov web site. What is your score for that task?

  • Jacob Barhak

    You provided scores , yet did not provide solutions to improve the scores. In the past you wrote blogs supporting post publication review. Here is one example: http://blogs.bmj.com/bmj/2011/11/11/richard-smith-a-woeful-tale-of-the-uselessness-of-peer-review/
    What are solutions for the rest of the low scores? Also, what about conferences and proceedings? Do those improve scores in some categories? In other words, where should one publish today in your mind?