A Coomarasamy (Research Fellow in Evidence Based Medicine)

Harold Gee (Consultant Obstetrician)

Mary Publicova (Medical Librarian)

Khalid S. Khan*

Dr. A Coomarasamy

Education Resource Centre

Birmingham Women’s Hospital

Birmingham B15 2TG

Phone: 0956 498 457

E-mail: arricoomar@hotmail.com

The main target audience of general medical journals is practitioners and the ultimate aim is to improve patient care. The authors argue that the information needs of practitioners are poorly met by these journals and propose various schemes to improve the value of journals to practitioners. They further suggest that journals need to move from a passive dissemination mode to focus on active dissemination of clinically relevant information and suggest various strategies for improvement.

Keywords: journalology, dissemination, effective dissemination, implementation, clinical medical journals, general medical journals.

The ultimate purpose of research in healthcare is to improve patient care. However, lack of effective strategies to disseminate the findings of such research has been an age-old problem. Lemon juice was shown to be effective in preventing scurvy in 1601 but it was almost 200 years later that British Navy took this intervention on board ¹. Such delays in implementation of research findings have serious and often deleterious effects on patient care.

The first medical journal, Medicina Curiosa, was published in 1684 in England ². Although it ceased to exist after two issues in the same year, medical journals have proliferated since then. In 1996, over 30,000 clinical journals were in publication with nearly 3,500 journals cited in the Medline ( mary to update the data). One might imagine that these journals would provide an effective medium of communication between scientists who do research and clinicians who use their results in practice, hopefully leading to timely implementation of effective interventions. Unfortunately, this does not seem to have been the case ³.

Liggins ⁴ performed a randomised controlled trial in the early 1970s that showed the effectiveness of antenatal corticosteroids in preterm delivery. Despite several repeated randomised trials throughout the 1970s and 1980s and a systematic review in 1987 ⁵providing conclusive evidence about the effectiveness of ante-natal steroid therapy, practitioners all over the world have been slow in adopting this treatment ⁵. In particular professional bodies have been slow to react to the emerging evidence and the Royal College of Obstetricians and Gynaecologists recommended this treatment in UK in only 199? (mary to check). By denying patients the benefit this treatment for almost 2 decades, we have done them a great disservice.

In this commentary we explore how and why journals are failing to bring useful studies to clinicians’ attention so that a change in practice with an expectation of improved patient care can be achieved.

A typical clinical journal's contents include an editorial, commentaries, papers (basic science and clinical research), reviews (non-systematic and systematic), case-reports and correspondence. Selection of material for publication rests on the editorial team and peer-reviewers. The publication of various types of reports allows communication between researchers, clinicians and the lay public to varying degrees.

In table 1, we analyse the contents of medical journals in terms of the opportunities for effective communication. Our thesis is that communication aimed at practitioners will lead to improved dissemination and implementation of research findings that have value for patients.

Scientists are investigators with no or minimal clinical work researching mostly on basic science subjects and conducting field and bench studies. Academic clinicians include clinical professors and their teams and other clinicians actively involved mainly in clinical research. Practitioners are those whose major commitment is clinical work.

Health research often begins in the laboratory but for it to lead to clinically important changes in practice, it would have to be evaluated in patient centred research first. Scientists test most new ideas in laboratories on animals (bench studies) or on small human population (field studies) ⁶. Many successful bench studies and field studies fail at the next stage of rigorous 'clinical trials'. It is, therefore, unwise to base a change in practice on bench or field studies except in a few circumstances such as when a study result is exceptionally impressive in a cohort of patients who would have otherwise had a uniformly bad outcome ⁷. Such studies are rare and overall the value of bench and field studies to clinicians is minimal. Whilst being possibly being interesting, these studies have the potential to harm if a clinician were to base practice on small studies before the arrival of definitive large studies or systematic reviews. This was the case with flecainide for ventricular arrhythmia prophylaxis. ⁸

Table 1 highlights that systematic reviews and large robustly conducted clinical trials are likely to be the most useful sections of a medical journal aimed at practitioners with view to directly benefiting patients. These form a small part of the general medical journals as illustrated by an article in the ACP Journal Club that showed the yield of articles that met the criteria for scientific merit and direct relevance for clinical practice in internal medicine was less than 1 article per issue and for journals lower than the top 6, the yield decreased to less than 5 articles per year ⁹. Balanced editorials and commentaries may also be useful in dissemination but they rely on existing systematic reviews and trials as supporting evidence. Journals have traditionally not paid enough attention to secondary research.

We will first describe what we mean by dissemination, the spectrum of views around which range from a limited focus on getting the word out to an all-encompassing focus on seeing the evidence all the way through to implementation. It is important to distinguish between ‘dissemination’ and ‘effective dissemination’: The former term is often used to indicate merely the successful distribution of information whereas the latter term goes much further than distribution, i.e. into implementation. One view might be that medical journals can only distribute information so their role is likely to be limited in effective dissemination. This is because the assumption that simply making research information available (dissemination alone) can lead to application in practice and policy is largely discredited ¹⁰

Dissemination activities seek to strengthen awareness and enhance the impact of research findings amongst relevant target audiences. They help in appraisal of current practice in the light of alternatives and facilitate readiness for change on the part of health professionals and organisations. Getting the findings of research into practice often needs to involve a range of techniques to increase awareness, develop and disseminate guidance, promote them and then maintain the adoption of guidance at the local level. In our view, many of these goals can be achieved by medical journals - but so far the focus of journals seems to have been on scientific research at the exclusion of concern for supporting change.

Some of the reasons why the paper journals are ineffective in dissemination and implementation are highlighted in table 2. The existence of too many journals which has come to be known as the ‘Balkanisation of the medical literature’ ¹¹ results in scattering and wrong targeting of the medical information as shown by one study which found only 15% of the original articles on the risk of oral contraceptive on breast cancer were on specialist obstetrics and gynaecology journals¹²

Supposing the busy practitioner has gone through the laborious task of gathering the scattered articles and selected the clinical studies, discarding the bench and field studies, can he or she now be sure of the methodological quality of these papers? Most people rely on peer review to ensure methodological quality and may not even read the methods section. However, a study by Haynes showed over 90% of articles to be of poor methodological quality and thus should not be the basis of clinical practice ⁹.

Most of the (non-systematic) review articles are narrative where the author’s opinion may be based on a biased selection of primary papers, culminating in conclusions that do not reflect the actual balance of evidence ³. This is a significant disservice from paper journals as the main purpose of review articles is to communicate to practitioners and reviews with incorrect conclusions may end up influencing practice in a way that may be harmful to the patients.

Systematic reviews represent a better balance of the available evidence but both the primary papers and systematic reviews suffer in the hands of journals in the form of publication bias ¹³. A recent study by Sutton et al indicated that publication bias may be present in up to 50% of meta-analyses in the cochrane database of systematic reviews ¹⁴. Studies with less dramatic results may be published in journals with small circulation it has been known that bilingual German researchers are likely to submit studies with impressive results to English journals and modest ones to German journals ⁷. Although articles do get published in these instances, there is a subtle form of publication bias in operation here as most of the practitioners may not become aware of these studies especially if they are not cited in the electronic bibliographic databases.

Authors speculate and exaggerate their results ¹⁵. Many studies are methodologically flawed ⁹. Journals have a duty to educate their readers in being able to select the good ones from bad ones, yet paper journals have traditionally neglected the role of educating the readers in critical appraisal of literature so that effectiveness and safety in considering studies could be assessed. It is rare for a product to be sold without some sort of user’s guide but this is what journals tend to have done in the past. Recently JAMA has been publishing user guides and BMJ has focussed on the subject of getting evidence into practice ¹⁶.

Peer reviewing is intended to act as an intellectual quality control to reject methodologically unsound studies, avoid publication bias, curb the exaggeration of results and improve presentation. Apart from the deductive evidence that peer reviewing is likely to be ineffective given the proportion of poor quality studies in clinical journals, their is published evidence for publication bias, nationality bias, language bias and for the inability to detect defective studies. Peer reviewing is widely accepted to be ineffective and even corrupt ^17-19 Another disadvantage is the delay it incurs from submission to publication of about 12-16 months which can unnecessarily delay clinically useful information reaching clinicians ²⁰. Some journals have now started using fast track procedures ²¹

We have earlier indicated that effective dissemination should aim to achieve change which is not easy. The reasons for opposition to change are many, not least the comfort with status quo, fear of loss of control, low level of motivation and poor morale, low pressure to change, lack of ownership, perceived non-feasibility of change and doubt about ability to respond to new demands. Research by the Cochrane Effective Practice and Organisation of Care Group has evaluated the effectiveness of different dissemination and implementation interventions ^22;23. Most interventions were found to be effective under some circumstances, although none were effective under all circumstances ²².

Activities associated with improved adoption of research evidence and change clinical practice include academic detailing ²⁴ reminder systems ²⁵ feeding back of cost information ²⁶, ongoing feedback ²⁷, multifaceted approaches ²⁸and interventions based on assessment of potential barriers . Further research is needed to better inform which strategies to use for a particular topic, in a particular setting, for a particular audience ^22;29 Table 3 highlights the various implementation strategies and their potential implications for medical journals.

Clinical medical journals, in addition to the usual role of keeping the members informed of news and defending their interests, need to focus on rigorous well-conducted clinical trials as opposed to bench and field studies which are best left to the biomedical journals. Systematic reviews and other review articles with a balanced view of evidence should take prominence. The articles should be readable and aim to benefit patients directly. To that end, authors should be encouraged to express their results in clinically meaningful ways.

As case studies are essentially anecdotal evidence, they should not generally be the basis of clinical decisions. They should not therefore be published in their current format advocating a novel intervention or management strategy which may be of interest to a clinical researcher but not to the practitioner. One exception is perhaps when they are reporting an adverse effect to a new treatment . Fiona Godlee ³⁰ describes a better use of case reports where the application of process of evidence based medicine to a usual clinical problem is highlighted. This would be a tool which the paper journals can use to educate their readers of critical appraisal, evidence based medicine and achieving implementation.

Much of the progress, however, is also likely to come from the electronic media and journal publishers should not shy away from this but actively promote it. Easy and quick accessibility to the large data bases where information is linked in a useful manner will clearly improve usability. The currency could be maintained with regular updating by the authors, especially in the case of systematic reviews, and from comments and letters from the readers. These comments and letters could be directly linked to the article rather than appearing months later in the paper edition. Different forms of the article could be produced aimed at different audiences ³¹. A clinician may benefit from a succinct, pre-digested summary with information on implementation strategies that may work whilst a researcher may appreciate a fuller version of the same article with such information as the background, detailed methodology, raw data, statistical analysis and details of the peer review for that article. Having that document translated into other languages will improve its use world-wide and the electronic version is more likely to be beneficial and affordable to the practitioners in the developing world. It is also likely that the time delay from submission to publication will be improved with the electronic format.

Clinical medical journals need to move from a passive dissemination mode to focus on active dissemination. Researchers should be encouraged to consider how and by whom the findings will be used and provide information on implications for implementation such as possible strategies that may work, cost-effectiveness, side-effects and potential barriers to implementation. Recent times have seen an increase in evaluative research and systematic reviews and this needs to be reflected in the clinical medical journals with the addition of pre-appraised evidence summaries. This has been aided by publication of evidence-based journals. However, general medical journals can also become proactive by producing supplements like the ACP journal club or by having critically appraised summaries as a section in the journal.

Although implementation is essentially a task at the local level, journals are in a unique position to advance the cause (see table 4)

Summary of the role of medical journals in effective dissemination

Of the several hundred medical journals that flourished in the 17the century, none made it to the following century. Unless the current clinical medical journals adjust their focus to the needs of their main users i.e. practitioners, they may find that they lose their struggle for survival, especially in the current climate of evidence based practice where journals that summarise evidence begin to proliferate and the electronic media offers a real alternative to paper journals.