We are currently performing maintenance and will be offline for approximately 4hrs. Follow us on Twitter @Covidence for any updates.

How to write the methods section of a systematic review

Covidence breaks down how to write a methods section

The methods section of your systematic review describes what you did, how you did it, and why. Readers need this information to interpret the results and conclusions of the review. Often, a lot of information needs to be distilled into just a few paragraphs. This can be a challenging task, but good preparation and the right tools will help you to set off in the right direction 🗺️🧭.

 

Systematic reviews are so-called because they are conducted in a way that is rigorous and replicable. So it’s important that these methods are reported in a way that is thorough, clear, and easy to navigate for the reader – whether that’s a patient, a healthcare worker, or a researcher. 

 

Like most things in a systematic review, the methods should be planned upfront and ideally described in detail in a project plan or protocol. Reviews of healthcare interventions follow the PRISMA guidelines for the minimum set of items to report in the methods section. But what else should be included? It’s a good idea to consider what readers will want to know about the review methods and whether the journal you’re planning to submit the work to has expectations on the reporting of methods. Finding out in advance will help you to plan what to include.

Figure 1: The PRISMA checklist is a minimum set of items for reporting systematic reviews of healthcare interventions

Describe what happened

While the research plan sets out what you intend to do, the methods section is a write-up of what actually happened. It’s not a simple case of rewriting the plan in the past tense – you will also need to discuss and justify deviations from the plan and describe the handling of issues that were unforeseen at the time the plan was written. For this reason, it is useful to make detailed notes before, during, and after the review is completed. Relying on memory alone risks losing valuable information and trawling through emails when the deadline is looming can be frustrating and time consuming! 

Keep it brief

The methods section should be succinct but include all the noteworthy information. This can be a difficult balance to achieve. A useful strategy is to aim for a brief description that signposts the reader to a separate section or sections of supporting information. This could include datasets, a flowchart to show what happened to the excluded studies, a collection of search strategies, and tables containing detailed information about the studies.This separation keeps the review short and simple while enabling the reader to drill down to the detail as needed. And if the methods follow a well-known or standard process, it might suffice to say so and give a reference, rather than describe the process at length. 

Follow a structure

A clear structure provides focus. Use of descriptive headings keeps the writing on track and helps the reader get to key information quickly. What should the structure of the methods section look like? As always, a lot depends on the type of review but it will certainly contain information relating to the following areas:

  1. Selection criteria ⭕
  2. Search 🕵️
  3. Data collection and analysis 👩‍💻
  4. Study quality and risk of bias ⚖️

Let’s look at each of these in turn.

1. Selection criteria ⭕

The criteria for including and excluding studies are listed here. This includes detail about the types of studies, the types of participants, the types of interventions and the types of outcomes and how they were measured. 

2. Search 🕵🏾‍♀️

Comprehensive reporting of the search is important because this means it can be evaluated and replicated. The search strategies are included in the review, along with details of the databases searched. It’s also important to list any restrictions on the search (for example, language), describe how resources other than electronic databases were searched (for example,  non-indexed journals), and give the date that the searches were run. The PRISMA-S extension provides guidance on reporting literature searches. 

Systematic reviewer pro-tip:

 Copy and paste the search strategy to avoid introducing typos

3. Data collection and analysis 👩‍💻

This section describes:

  • how studies were selected for inclusion in the review
  • how study data were extracted from the study reports
  • how study data were combined for analysis and synthesis

To describe how studies were selected for inclusion, review teams outline the screening process. Covidence uses reviewers’ decision data to automatically populate a PRISMA flow diagram for this purpose. Covidence can also calculate Cohen’s kappa to enable review teams to report the level of agreement among individual reviewers during screening.

To describe how study data were extracted from the study reports, reviewers outline the form that was used, any pilot-testing that was done, and the items that were extracted from the included studies. An important piece of information to include here is the process used to resolve conflict among the reviewers. Covidence’s data extraction tool saves reviewers’ comments and notes in the system as they work. This keeps the information in one place for easy retrieval ⚡.

To describe how study data were combined for analysis and synthesis, reviewers outline the type of synthesis (narrative or quantitative, for example), the methods for grouping data, the challenges that came up, and how these were dealt with. If the review includes a meta-analysis, it will detail how this was performed and how the treatment effects were measured.

4. Study quality and risk of bias ⚖️

Because the results of systematic reviews can be affected by many types of bias, reviewers make every effort to minimise it and to show the reader that the methods they used were appropriate. This section describes the methods used to assess study quality and an assessment of the risk of bias across a range of domains. 

Steps to assess the risk of bias in studies include looking at how study participants were assigned to treatment groups and whether patients and/or study assessors were blinded to the treatment given. Reviewers also report their assessment of the risk of bias due to missing outcome data, whether that is due to participant drop-out or non-reporting of the outcomes by the study authors.

Covidence’s default template for assessing study quality is Cochrane’s risk of bias tool but it is also possible to start from scratch and build a tool with a set of custom domains if you prefer.

Conclusion

Careful planning, clear writing, and a structured approach are key to a good methods section. A methodologist will be able to refer review teams to examples of good methods reporting in the literature. Covidence helps reviewers to screen references, extract data and complete risk of bias tables quickly and efficiently. Sign up for a free trial today!

Share...

Share on facebook
Facebook
Share on twitter
Twitter
Share on linkedin
LinkedIn

Perhaps you'd also like...

Better systematic review management

By using our site you consent to our use of cookies to measure and improve our site’s performance. Please see our Privacy Policy for more information.