When is it appropriate to conduct a(nother) systematic review?

Abstract

Despite some concerns regarding their validity 1, systematic reviews and meta-analyses remain important in bridging the gap between scientific evidence and clinical practice. The Centre for Evidence-based Medicine recommends that the highest level of evidence (Level 1) should be based on systematic review of studies with homogeneity 2. This applies to research dealing with therapy, prevention, etiology, harm, prognosis, diagnosis, differential diagnosis, and symptom prevalence, as well as economic and decision analyses. Given the enormous amount of published information, it is impossible for medical practitioners to read all that is published, to synthesise the evidence, to come to a meaningful interpretation of results and apply that in clinical practice. Systematic reviews, such as Cochrane reviews, respond to this challenge by identifying, appraising and synthesizing research-based evidence and presenting it in a form that is easy to access and understand. Many systematic reviews contain meta-analysis, which is the use of statistical methods to summarize the results of independent studies. By combining information from all relevant studies, meta-analyses can provide more precise estimates of the effects of healthcare than those derived from the individual studies included within a review. They also facilitate investigations of the consistency of evidence across studies, and the exploration of differences across studies 3. Systematic reviews often exist in published literature, and new studies related to the topic of the systematic review are being published all the time. Health policy makers, professional organisations and authorities issuing clinical guidelines could also solicit systematic reviews. Journals are generally interested in publishing systematic reviews because they present the highest level of evidence and are attractive for readers. Furthermore, they are cited more often than other articles and may potentially help to increase a journal′s impact factor. We have recently noticed a trend in the increasing number of systematic reviews being submitted to our journal. However, occasionally we see that independent teams have been working on the same topic. This may lead to duplication of the reviews, and potentially wasted efforts. An international prospective register for proposed systematic review (PROSPERO) exists 4 and supports the aims of the REduce research Waste And Reward Diligence group (REWARD, www.researchwaste.net). As with any research, systematic reviews should be performed carefully, using correct methodology, and results should be appropriately interpreted. A badly performed systematic review can do a lot of harm. It is also important to realize that systematic reviews that are not maintained or updated become out of date and misleading. For this reason, the policy of the Cochrane collaboration is that the review should be reviewed every 2 years. In general, a new review is useful and prudent when significant new information is available. Does it mean that every time a new study is published, a new systematic review is needed or justified? There are several publications about the characteristics of systematic review/meta-analysis and factors related to them being cited in medical journals. Uthman et al. 5 studied the bibliometrics of the100 most cited systematic reviews and meta-analyses. They reported a statistically highly significant negative correlation between average citations per year and year since publication. Helfer et al. 6 showed that a significant proportion of recently published meta-analyses either did not identify a previous systematic review/meta-analysis on the same topic (34%) or did not describe (64%) or discuss (80%) it. Along with Royle et al. 7, they reported that previous systematic reviews were more likely to be cited, described, or discussed by a recent meta-analysis if the previous article was published in a journal with a high impact factor, if results were different, and if meta-analytic methods were used. Helfer et al. 6 propose that journals should make the discussion of related meta-analyses mandatory. Currently, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement recommends that authors of systematic reviews explain in the introduction how their work adds to what is already known, and explain whether it is a new review or an update 8. There may be several reasons why researchers may wish to perform a new systematic review on a particular topic. Newer studies with bigger numbers or better methods may have been published since the publication of the last systematic review/meta-analysis and the researchers may wish to re-confirm the findings. Sometimes, the conclusions of the previous review may change with the inclusion of data from newer studies. Researchers may disagree with the inclusion/exclusion criteria, quality assessment, analysis and interpretation of data presented in a previous review, and want to offer an alternative or different interpretation. Sometimes, they are just looking for another publication. Therefore, we are vigilant and carefully scrutinize all the submissions, including systematic reviews, submitted to our journal. We strongly recommend all the researchers to prospectively register their systematic review protocols in a publically accessible database, such as PROSPERO. We believe there is no specific time limit for the validity of a published systematic review. However, the rationale for repeating a systematic review/meta-analysis should be stated explicitly. The manuscript should include what is already known from previous reviews and what this current review adds. In accordance with the Cochrane reviews, a newer systematic review should include an obligatory section of ‘agreements and disagreements with other studies or reviews’. This would make systematic reviews more useful, transparent, and a valuable source of information for clinicians, researchers, policymakers and patients.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 5 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.