This is an archived version of the Handbook. For the current version, please go to training.cochrane.org/handbook/current or search for this chapter here.

22.3.4  Text of a Cochrane Overview

The target audience for a Cochrane Overview is people who make decisions about health care (e.g. clinicians, informed consumers and policy makers) who already have some basic understanding of the underlying disease or problem and wish to discover the extent to which the potential interventions for the problem have been addressed in The Cochrane Library. The Overview should provide an overview of the findings of relevant Cochrane reviews, and direct the reader to the individual reviews for additional detail.

 

The text of a Cochrane Overview contains a number of fixed headings. Subheadings may be added by the author at any point. Certain specific headings are designated as ‘recommended’. The content of recommended sections should be included in all Overviews, but the use of the actual subheading is not mandatory and should be avoided if they make individual sections needlessly short. Additional subheadings that may or may not be relevant to a particular review are also provided. In the rest of this section, the relevant category (fixed, recommended, optional) is noted for each of the headings described.

 

Background

[fixed, level 1 heading]

This section should address the already-formed body of knowledge that comprises the context of the Cochrane reviews summarized in the Overview. The background helps set the rationale for the Overview. It should specify the research question(s) being addressed by the Overview, including a clear description of the condition of interest, the interventions, comparisons, and the outcomes considered. Furthermore, it should explain why the questions being asked are important. It should be presented in a fashion that is understandable to the users of the health care under investigation, and should be concise (generally around one page when printed). The background section should contain the following components. Although subheadings are not mandatory, they are recommended.

 

Description of the condition

[recommended, level 2 heading]

The review should begin with a brief description of the condition being addressed and its significance. It may include information about the biology, diagnosis, prognosis and public health importance (including prevalence or incidence).

 

Description of the interventions

[recommended, level 2 heading]

This section should mention all of the interventions currently available for the condition, whether or not the interventions have been evaluated in a Cochrane Intervention review. Where reasonable, grouping interventions will simplify the text (e.g. listing non-steroidal anti-inflammatory drugs rather than providing an exhaustive list of all such drugs by name). The possibility of concurrent use of different interventions (e.g. radiation plus chemotherapy) should be addressed, if applicable. The relative status of the various potential interventions in current clinical practice may be mentioned (if feasible).

 

How the interventions might work

[recommended, level 2 heading]

Systematic reviews gather evidence to assess whether the expected effect of an intervention does indeed occur. This section might describe the theoretical reasoning why the interventions under review might have an impact on potential recipients of health care, for example, by relating a drug intervention to the biology of the condition. Authors may refer to a body of empirical evidence such as similar interventions having an impact or identical interventions having an impact on other populations. Authors may also refer to a body of literature that justifies the possibility of effectiveness. References to existing literature should not include any discussion of the results of the systematic reviews contained in the Overview or the studies addressed in those reviews; this material should be covered in the Results section.

 

Why it is important to do this overview

[optional, level 2 heading]

The background helps set the rationale for the Overview, and should explain why the questions being asked are important. It should make clear why this Overview was undertaken, who the target audience is, and what decisions it is intended to help inform.

 

Objectives

[fixed, level 1 heading]

This should begin with a precise statement of the primary aim of the review, including the intervention(s) reviewed and the targeted problem. This might be followed by a series of specific objectives relating to different participant groups, different comparisons of interventions or different outcome measures.

 

Methods

[fixed, level 1 heading]

The Methods section in a protocol should be written in the future tense. The Methods section of the review should describe what was done to obtain the results and conclusions of the current version of the Overview. It should not discuss the methods of the underlying systematic reviews that are being summarized. Comments on the methods of these reviews should be addressed in the section ‘Description of included reviews’. The Methods section should have a number of component subsections.

 

Criteria for considering reviews for inclusion

[fixed, level 2 heading]

The Overview research question should guide selection of reviews for inclusion, including a clear description of the participants (condition or health problem), the interventions, comparison groups and outcomes of interest. In general, Overviews should include all Cochrane reviews that address one or more of the interventions available for the condition or health problem that is the topic of the Overview. However, in some cases the authors of the Overview may wish to restrict this focus in some way. For example, Overview authors may wish to restrict their scope to certain types of interventions (e.g. all drug therapies, excluding non-drug therapies). Restrictions would be particularly appropriate if the existing Cochrane reviews address varied clinical populations (e.g. groups that differ by age, ethnicity, sex, stage of disease or types of co-morbidity). In making decisions to lump or split, it will be helpful to keep in mind the perspective of the decision maker reading the overview and to focus on the information that would be required to make an individual decision. For example, Cochrane Intervention reviews addressing prevention of a given condition should probably not be grouped in a single Overview with Intervention reviews addressing treatment of the same condition – since prevention decisions and treatment decisions are made for different populations. If such considerations are involved in the selection of reviews for inclusion in the Overview, they should be clearly spelled out in this section.

 

If non-Cochrane systematic reviews are included, this section should specify the criteria that will be used to determine whether non-Cochrane reviews are systematic reviews, and the criteria that will be used to determine which systematic reviews will be included when there are two or more reviews that address the same question.

 

Search methods for identification of reviews

[fixed, level 2 heading]

This should address the methods used in the Overview to find Cochrane reviews or other systematic reviews. The search involved will be much simpler than the search strategies within a Cochrane Intervention review, because the basic search for underlying articles will have already been performed. If only Cochrane reviews are to be included in the overview, the search can be performed within the Cochrane Database of Systematic Reviews without the need to search other databases. If systematic reviews from other sources are included, this section should clearly outline the databases searched (e.g. Database of Abstracts of Reviews of Effects (Petticrew 1999)) and the search strategies and retrieval methods used.

 

Data collection and analysis

[fixed, level 2 heading]

This section should present a brief description of the methods used in the Overview. The following issues should be addressed:

 

Selection of reviews

[recommended, level 3 heading]

The method used to apply the selection criteria to reviews identified in the search and whether the criteria are applied independently by more than one review author should be stated, along with how any disagreements are resolved.

 

Data extraction and management

[recommended, level 3 heading]

The method used to extract or obtain data from the included reviews (for example, using a data collection form) should be described in this section. Whether data are extracted independently by more than one author should be stated, along with how any disagreements are resolved. If relevant, methods for processing data in preparation for analysis should be clearly described. Authors should also describe what, if anything, is done to collect data that are missing from the included reviews.

 

Assessment of methodological quality of included reviews

[recommended, level 3 heading]

Two different quality assessments must be addressed by the Overview authors in each Overview: the methodological quality of the reviews summarized in the Overview, and the quality of the evidence in these reviews, as described below.

 

The methods used in performing both types of assessment should be described in this section. For both assessments it is recommended that more than one review author should apply the criteria independently. This should be stated, along with how any disagreements are resolved. The tools used (e.g. GRADE) should be described or referenced, with an indication of how these assessments are incorporated into the interpretation of the results of the Overview.

 

Quality of included reviews

[recommended, level 4 heading]

The methods used to assess the methodological quality of the reviews included in the Overview should be described. There has been limited research on the assessment of quality, or risk of bias, in systematic reviews, and we are unable to recommend a specific instrument for reaching judgements about the quality of included reviews. However, some questionnaires and checklists are available (Oxman 1994, Shea 2006).

 

Quality of evidence in included reviews

[recommended, level 4 heading]

Cochrane Intervention reviews that use excellent methods may summarize evidence with important limitations, because of potential biases within and across the included studies, conflicting results across individual studies, sparse evidence or a lack of relevance (directness) to the review question (see Chapter 12, Section 12.2). The methods used in the Overview to determine the quality of the evidence in support of each of the Overview’s conclusions should be summarized. Ideally, the information on which to base such assessments should be available in the ‘Characteristics of included studies’, ‘Risk of bias’ and ‘Summary of findings’ tables provided in the included reviews. It is now recommended that assessments of the risk of bias should be reported in a standardized way in Cochrane reviews (see Chapter 8) and that the GRADE approach should be used to assess the quality of evidence across studies for each important outcome for both Cochrane Intervention reviews and Overviews of Cochrane reviews (see Chapter 11, Section 11.5 and Chapter 12, Section 12.2).

 

Data synthesis

[recommended, level 3 heading]

Many Overviews will simply extract data from the underlying systematic reviews and reformat them in tables or figures. However, in some cases Overviews may include indirect comparisons based on formal statistical analyses, especially if there is no evidence on direct comparisons (Glenny 2005). Statistical methods for undertaking indirect comparisons, and for simultaneous meta-analyses of multiple interventions, are highly relevant to Overviews, and are discussed in Chapter 16 (Section 16.6). Evidence from indirect comparisons may be less reliable than evidence from direct (head to head) comparisons. If no included reviews have investigated direct comparisons, but studies of direct comparisons are known or believed to have been performed, then authors of Overviews should not attempt indirect comparisons. Authors who wish to undertake indirect comparisons or multiple-treatments meta-analyses should seek appropriate statistical and methodological support.

 

When more qualitative or narrative approaches are used, review authors should state what, if any, methods are used to standardize reporting of results across included reviews, including converting summary statistics and any standardization for different control group risks. Authors should be cautious when comparing absolute effects across reviews if there are differences in control group risks see Chapter 11, Section 11.5.5).

 

Results

[fixed, level 1 heading]

Description of included reviews

[fixed, level 2 heading]

The description of included reviews should be concise, but provide sufficient detail to allow the reader to get an idea of the characteristics of participants included in the summarized reviews: the dose, duration, or other characteristics of the interventions. If there are important differences between these component reviews (e.g. differences in the review criteria for inclusion or exclusion of studies, different comparators, or the use of different outcome measures) these should be clearly noted. In addition, any discrepancies between the objectives and eligibility criteria of the included reviews and the objectives of the Overview should be noted. For example, the review authors may have omitted analyses of a specific subgroup or of a key outcome that was of particular interest to the Overview authors. If some reviews have been updated more recently than others, this should also be noted. Much of the material in this section can be summarized in a ‘Characteristics of included reviews’ table (see Section 22.3.6 for details).

 

Methodological quality of included reviews

[recommended, level 2 heading]

Quality of included reviews

[recommended, level 3 heading]

The general quality of the systematic reviews included in the Overview should be summarized, including any variability across reviews and any important flaws in individual reviews. The criteria that were used to assess review quality should be described or referenced under ‘Methods’ and not here. If it is felt to be important to provide details on how each included review was rated against each criterion, this should be reported in an Additional table and not described in detail in the text.

 

Quality of evidence in included reviews

[recommended, level 3 heading]

The general quality of the evidence in the included reviews should be summarized, for example using GRADE for the most important outcomes (see also Chapter 13, Section 13.2).

 

Effect of interventions

[fixed, level 2 heading]

The main findings on the effects of the interventions studied in the included reviews should be summarized here. The section should be organized around clinically meaningful categories rather than simply listing the findings of each included review in turn. These categories could include things such as types of interventions (drug treatments, surgical interventions, behavioural interventions, etc); stages of disease (pre-symptomatic, early disease, advanced disease); participant characteristics (age, sex, ethnicity); or types of outcomes (survival, functional status, adverse effects). Subheadings are encouraged if they make reading easier. The findings of individual reviews, and any statistical summary of these, should be included in summary tables or figures.

 

Note should be made in this section of any outcomes that the Overview authors consider important but for which the review authors could not find evidence (either because no studies were found or because the studies identified did not report on the important outcome). In addition, this section should include a narrative summary of important results that can not easily be summarized using numerical data, and will not likely be included in the results tables of the Overview.

 

Authors should avoid making inferences in this section. A common mistake to avoid (both in describing the results and in drawing conclusions) is the confusion of 'no evidence of an effect' with 'evidence of no effect'. When there is inconclusive evidence, it is wrong to claim that the Overview shows that an intervention has ‘no effect’ or is ‘no different’ from the control intervention. In this situation it is more appropriate to report the data, with a confidence interval, as being compatible with either a reduction or an increase in the outcome.

 

Discussion

[fixed, level 1 heading]

Summary of main results

[recommended, level 2 heading]

Provide a concise summary here of the main findings, the balance between important benefits and important harms and highlight any outstanding uncertainties.

 

Overall completeness and applicability of evidence

[recommended, level 2 heading]

Are the reviews included sufficient to address all of the objectives of the Overview? If not, what gaps are present? Have all relevant types of participants, interventions and outcomes been investigated? Describe the relevance of the evidence to the Overview question. This should lead to an overall judgement of the external validity of the Overview. Comments on how the results of the Overview fit into the context of current practice might be included here, although authors should bear in mind that current practice might vary internationally and between populations.

 

Quality of the evidence

[recommended, level 2 heading]

Do the reviews included in the Overview allow a robust conclusion regarding the objective(s) addressed in the Overview? The discussion might include whether all relevant studies were identified in the original review, whether all relevant data could be obtained, or whether the methods used (for example, searching, study selection, data collection and analysis) could have introduced bias. This may vary for different interventions, outcomes or clinical subgroups. If so, the discussion should clearly identify the quality of evidence for each of the key areas of interest.

 

Potential biases in the overview process

[recommended, level 2 heading]

State the strengths and limitations of the Overview with regard to preventing bias. These may be factors within, or outside, the control of the Overview authors. The discussion might include whether all relevant reviews were identified and included in the Overview, whether all relevant data could be obtained, or whether the methods used (for example, searching, study selection, data collection and analysis) could have introduced bias.

 

Agreements and disagreements with other studies or reviews

[recommended, level 2 heading]

Comments on how the included reviews fit into the context of other evidence might be included here, stating clearly whether the other evidence was systematically reviewed.

 

Authors’ conclusions

[fixed, level 1 heading]

This section should present the conclusions of the authors of the overview, not simply restate the varying conclusions of the authors of the included/underlying reviews. The primary purpose of this section should be to present information rather than to offer advice. Conclusions of the authors are divided into two sections as follows.

 

Implications for practice

[fixed, level 2 heading]

The implications for practice should be as practical and unambiguous as possible. They should not go beyond the evidence that was reviewed and should be justifiable by the data presented in the review. ‘No evidence of effect’ should not be confused with ‘evidence of no effect’.

 

Implications for research

[recommended, level 2 heading]

This section should address the key clinical issues that remain unresolved after review of the evidence presented in the included/underlying reviews. If there are important potential interventions for the condition under consideration that have not been addressed in a Cochrane Intervention review, this gap should be clearly noted in this section. In addition to providing an agenda for future research, this section can be useful to clinical decision makers by clearly indicating the remaining areas of uncertainty.

 

Acknowledgements

[fixed, level 1 heading]

This section should be used to acknowledge any people or organizations that the authors wish to acknowledge, including people who are not listed among the authors: see Chapter 4 (Section 4.5).

 

Contributions of authors

[fixed, level 1 heading]

The contributions of the current co-authors should be described in this section: see Chapter 4 (Section 4.5).

 

Declarations of interest

[fixed, level 1 heading]

Authors should report any present or past affiliations or other involvement in any organization or entity with an interest in the review that might lead to a real or perceived conflict of interest: see Chapter 4 (Section 4.5). Authors must state if they have been involved in a study included in a component review, or in authoring a systematic review included in the Overview.

 

Differences between protocol and review

[fixed, level 1 heading]

It is sometimes necessary to use different methods from those described in the original protocol: see Chapter 4 (Section 4.5).

 

Published notes

[fixed, level 1 heading]

See Chapter 4 (Section 4.5).