Co-publication of systematic reviews

Geir Wenberg Jacobsen, Magne Nylenna About the authors

This issue of the Journal of the Norwegian Medical Association includes an article on the follow-up of patients with diabetes. It is a result of a new collaborative model, and is published in tandem with a more comprehensive report

Drawing conclusions on the basis of one single study is always fraught with risk, whether the topic is organisational or patient-oriented measures. Coincidences, errors or other circumstances may give rise to misleading results, even in large-scale, soundly based studies. Moreover, in areas in which many studies have been conducted and published, there is a risk of taking single studies as a basis, which are not representative but which are selected because the findings concur with researchers’ own experience or interests. This selective use of research results, or «cherry picking», is a significant potential source of error in decision-making processes.

In order to ensure an optimal evidence base, all available studies on the topic of a relevant intervention or issue should be retrieved and reviewed. The product of a search process for all studies conducted world-wide, assessment of the validity of their findings and compilation of the results is called a systematic review (1). If the data from various studies can be collected and reanalysed, this is known as a meta-analysis (2). By no means all systematic reviews can be accompanied by a meta-analysis. This is usually because the selected substudies are so heterogenous that the required preconditions are unfulfilled.

Systematic reviews are an important foundation for evidence-based decisions in the health services. Not only do they provide more precise information about the effect of interventions and other factors, but they also assess the reliability of the available documentation (3). One of the principle tasks of the Norwegian Knowledge Centre for the Health Services, which from 2016 forms part of the Norwegian Institute of Public Health, is to prepare systematic reviews to underpin sound decisions that are requisitioned by the health administration, health services and professional and user organisations. Some systematic reviews are included in international publications, but the majority are published in the form of Norwegian-language reports.

However, these reports seldom reach a large audience. The Journal of the Norwegian Medical Association and the Norwegian Knowledge Centre for the Health Services have therefore established a cooperation whereby an article version of the review is published at the same time as the full report. When the work on a report is 4 – 6 months from completion, the authors prepare the manuscript of an article and submit it for consideration in the usual manner. The quality requirements are, of course, the same as for other manuscripts, but a tighter schedule than usual must be set. In many ways this is similar to a form of «fast-track» assessment, something that several international journals offer. As medical editor, Geir Wenberg Jacobsen has chief responsibility for the project, and the first result is presented in this issue of the Journal (4). A small group of professional colleagues were contacted in advance and had expressed their willingness to contribute to a rapid assessment of the manuscript. The time from the first submission to acceptance, including several rounds between the authors and the editorial committee, was 89 days, or less than one-third of the median processing time for original articles in the Journal of the Norwegian Medical Association (5).

The article by Larun and colleagues (4) is based on a requisition from the Norwegian College of General Practice, and has resulted in a separate report (6). They were commissioned to clarify the benefit of form-based follow-up of patients with diabetes in general practice. Norwegian Quality Improvement of Primary Health Care Laboratories (NOKLUS) has prepared a follow-up form for use in general practice (7) which is recommended by the Norwegian Directorate of Health. It is striking that the effect of using forms for primary endpoints such as mortality, cardiac infarction and stroke was only investigated in one of the seven studies included in the review, and the use of forms had no effect on these. Although the review indicates a tendency towards an effect for some of the secondary endpoints, it is worth noting that the statistical significance is more evident than the clinical relevance. The authors thus also communicate a clear recommendation to the general practitioner community that «more research» is needed.

We hope that this is the start of a long and fruitful collaboration on the publication of systematic review articles, which may perhaps be emulated by other research groups who mainly publish reports. One of the challenges will be to find topics of sufficient relevance for the readers of this journal, and secondly, a willingness to cooperate among those who requisition summaries, as well as motivated and dedicated authors, well-prepared peer reviewers, a tight but realistic schedule, and a staff on both sides of the table who see the benefit of this mode of publishing.

There has been a debate in academic circles as to whether publications that present systematic knowledge reviews are sufficiently «high-level» to be able to constitute independent subworks of a PhD dissertation, for example. This discussion should now be over. Preparation of a systematic review – with or without meta-analyses – not only represents original research, it is also important and necessary research, as the present article shows. It is for that reason that the Journal of the Norwegian Medical Association has classified it as an original article.

Anbefalte artikler