Cochrane systematic reviews provide reliable, evidence-based information on health issues.
A systematic review is the result of a rigorous scientific process consisting of several well-defined steps, including a systematic literature search, an evaluation of the quality of each included study and a synthesis, quantified or narrative, of the results obtained. The findings summarize the evidence on the efficacy of a treatment, the risk of adverse events or the accuracy of a diagnostic test, for example. However, sometimes the authors have to acknowledge that there is a lack of rigorous scientific studies.
All Cochrane Systematic Reviews answer a well-defined health question, such as the efficacy and safety of a surgical procedure or drug therapy, by considering all studies conducted on this question over time that meet established quality criteria. See also "How is a systematic review produced?"
The full text is a document, often of several dozen pages, divided into several parts, the main ones being:
- A scientific abstract and a plain language summary, the latter very often translated into several languages;
- The main sections of the review (background, objectives, methods, results, discussion, conclusion);
- Tables describing the included studies (in detail), the excluded studies (with reason for exclusion) and the results of meta-analyses (if applicable);
Graphs, especially forest plots.
The elaboration of a systematic review is a rigorous scientific process consisting of several steps:
- Clearly define the question to be addressed;
- Search and identify all relevant references of clinical trials or other appropriate studies, published or unpublished, that aim to answer the review’s question;
- Assess the quality of each study using standardized tools;
- Extract and organize relevant data from the included publications and other sources of information;
- Prepare an appropriate synthesis of the extracted results. If the data permit, perform a statistical analysis called a meta-analysis, which is used to combine quantified findings from several studies into a single pooled estimate.
One of Cochrane's principles is to avoid redundancy as much as possible. To do this, the Cochrane Review Group in charge verifies that no other Cochrane review with the same question exists yet. If that’s not the case and other criteria are met, the title of the new review can be registered. For more information see the page "Become an author" on cochrane.org.
- Not all relevant studies answering the same specific health question are necessarily published, for example, when their results do not support the efficacy of a new treatment. Sometimes only part of the results are published, for example, only the outcomes with statistically significant differences. Cochrane supports the AllTrials initiative for the publication of all clinical trials and their full methods and results (www.alltrials.net).
- Certain publications do not describe the study methods with sufficient detail to allow for critical review and evaluation.
- Studies are often carried out under "ideal" conditions that do not account for factors compromising the efficacy of a treatment in routine care, such as patients’ co-morbidity or non-compliance with therapy.
The workload and time required to prepare and complete a Cochrane Systematic Review project is frequently underestimated, especially if the contributors are inexperienced.
Systematic reviews are conducted by health professionals or scientists, often with the ad hoc support of patients or patients' relatives. To prepare the review, the authors collaborate closely with the appropriate thematic review group (Cochrane Review Group), which ensures the editorial follow-up during the registration of the title, the writing of the protocol, the implementation of each step of the review and the publication in the Cochrane Library. The authors of Cochrane reviews are not funded by Cochrane but very often by public funds. Funding from the commercial sector is not accepted.
The Cochrane Review Groups are organised in 8 networks and ensure that the rigorous quality standards that have built Cochrane’s reputation are maintained. Both, the protocol and the full Cochrane review are peer-reviewed prior to publication in the Cochrane Library.
Systematic reviews produced by Cochrane are published in the Cochrane Library (www.cochranelibrary.com/). In addition to the Cochrane reviews, this online library also includes the CENTRAL database with references to controlled studies identified in PubMed, EMBASE and through manual searches ("hand-search"), and information about Cochrane.
Under the label "Cochrane Clinical Answers" a selection of systematic reviews of wider interest (in particular, in primary care medicine) are presented in a question-answer format with interactive tables facilitating rapid access to the results.
In Switzerland, all content of the Cochrane Library is freely accessible through a national license.
Cochrane Systematic Reviews are all built according to the same scheme, all the steps of their conduct are well described and all the choices made during the process are outlined. This transparency in the process helps readers to understand what options were taken and why they were chosen.
In addition, Cochrane Systematic Reviews are regularly updated, which is rarely the case for systematic reviews published elsewhere. These updates are performed as needed, for example, when a significant number of new studies have been published. Updating is important to ensure that the latest clinical research is taken into account.
Cochrane Systematic Reviews address questions other than the efficacy and safety of therapy. An important area is the performance and accuracy of diagnostic tests. A diagnostic test is a test, such as a laboratory test, imaging technique or clinical examination, performed on a person with a suspected disease or condition. It is used to confirm or rule out the presence of that disease or condition and should lead to a therapeutic decision (whether and which treatment to undertake). Cochrane has set standards for the development of "diagnostic reviews" and introduced this type of review as a routine process from 2008 on. One of the methods groups (Cochrane Screening and Diagnostic Test Methods Group) monitors the ongoing methods development and supports the author groups conducting this type of reviews.
As time goes by, other types of reviews have been admitted. At present, these reviews are still few in number and are not part of Cochrane’s routine processes. These include reviews of prognostic studies, qualitative evidence syntheses and living systematic reviews.
To assess the efficacy of a medical intervention, the results of randomized clinical trials are central to the analysis. However, they often provide little evidence of safety, especially of serious adverse events. Typically, observational studies include larger, less selected populations with longer follow-up periods. In a systematic review of interventions, it is advisable to include the results of good quality observational studies to gain a more complete picture of the benefits (effectiveness) and risks of an intervention.
Some other relevant questions can only be answered using the results of observational studies. For example, to determine the prevalence of a disease based on estimates made in different countries, a systematic review based on data from population-based studies, e.g. cross-sectional studies, might be conducted.