Knowledge Synthesis and Covidence

The Review Process: Step by Step

Developing your Research Question

A well-formulated research question is answerable and clearly defines the scope of the review. you can use a research question formulation framework to help you distil your question into a more answerable format that is focused and searchable on various databases.

The most widely known framework is PICO:

  • Population/Patient/Problem
    • What are the patient characteristics or disease characteristics that are being looked at?
  • Intervention
    • What treatment is being looked at?
  • Comparison
    • What is the current gold standard or alternative treatment option? It is also common for the comparison to be 'usual care'.
  • Outcome
    • What is the measurable goal? For example, improved outcome, reduced adverse effect or sympton.

PICO doesn't fit every possible type of research question, but there are alternatives. The links below provide more information about alternative frameworks that may better fit other question types such as qualitative, policy-focused, or general questions.

Writing the Protocol

For most knowledge synthesis projects, it is recommended to publish your protocol before beginning the review. For systematic reviews, this is a requirement. By planning and considering each step of the review process from conception to publication, you can proactively address potential issues. Even if your review type does not require a protocol, it is still advisable to create one.

A protocol outlines the reason why you're conducting the review and gives a detailed overview of the methods that will be used. PROSPERO is a good option for health science topics. Cochrane and JBI have their own separate processes for reviewing review protocols. An alternative option is the Open Science Framework (OSF).

Example Protocols

Systematic Review:

Xiao Z, Baldwin MM, Meinck F, Obsuth I, Murray AL. The impact of childhood psychological maltreatment on mental health outcomes in adulthood: a protocol for a systematic review and meta-analysis Syst Rev. 2021;10(1):224. Published 2021 Aug 12. doi:10.1186/s13643-021-01777-4.

Scoping Review:

Cassidy CE, Beck AJ, Conway A, et al. Using an integrated knowledge translation or other research partnership approach in trainee-led research: a scoping review protocol. BMJ Open. 2021;11(5):e043756. Published 2021 May 25. doi:10.1136/bmjopen-2020-043756 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8154946/

Methods

Eligibility Criteria

By clearly and explicitly defining what words and phrases are relevant to your research question and what are not helps ensure that all team members are on the same page when it comes to the screening process. After deciding on your research question, come up with the eligibility criteria before you begin developing your search queries as the eligibility criteria will inform your search strategy.

There are many ways to approach systematic review searching. From the beginning, it is important to:

  • Document your entire search process. This will be important when writing your methodology and ensuring that your search strategy is transparent and replicable.
  • Involve a librarian in your research team, if possible. A librarian can help ensure your search strategy is comprehensive.

Initially, you may try conducting a non-systematic basic search in a database relevant to your research question. This can help:

  • Identify existing systematic reviews on the same topic.
  • Find keywords and subject headings to consider in your own search strategy.

As you start developing your search strategy, check:

  • Have you identified all major concepts within your research question?
  • Have you considered correct synonyms/acronyms for each concept?
  • Have you accounted for different spellings of your concepts? For example, regional spelling differences like "colour" and "color".
  • Have you identified appropriate databases to search (at least 3)?
  • Have you identified relevant subject terms/controlled vocabulary for your concepts in each database?
    • Are your subject terms exploded/restricted? Do you need to use subheadings?
  • Have you translated your search syntax to match each database you search? For example, using [tiab] in PubMed vs. :ti,ab in Embase.
  • Have you used Boolean operators (AND/OR/NOT) appropriately?
  • Do you need to use database filters? For example, limiting results to those published between 2010-2024 to fit your research question.

Resources

Citation Management

Citation managers, also known as reference or bibliographic managers, can help you stay organized during the research process and keep track of the many references you will have. Citation managers allow you to:

  • Gather and organize citations for a variety of resources, including journal articles, reports, books, websites and videos
  • Store PDF copies of journal articles and reports
  • Generate a bibliography or reference list in a specific citation style
  • Insert citations or footnotes while writing
  • Share citations and collaborate with colleagues

Nova Scotia Health Library Services supports Zotero as a citation manager. Please visit our Citation Management guide. to learn more about using Zotero, as well as how to:

  • Cite sources
  • Create bibliographies
  • Avoid plagiarism and copyright infringement

Grey Literature

Grey literature refers to literature that is not 'formally' published, or that is produced by an organization whose main output is not publishing. For example, Nova Scotia Health policies are considered grey literature, as Nova Scotia Health is a health organization and not an academic or commercial publishing house. Other examples of grey literature include:

  • Academic theses and dissertations
  • Government reports
  • Conference proceedings
  • Clinical trial registries
  • White papers

Including grey literature in your review helps eliminate publication bias. This is because publications are more likely to get published if they show positive results/outcomes rather than negative results/outcomes. Of course, even with that in mind, grey literature is not a replacement for academically published literature.

For a list of potential grey literature databases to search, check our A-Z Databases list on the Library Services website. For more information on how to search grey literature, check out:

Step 3.1: Deduplication

After you have run your searches on all the databases you've chosen and compiled your search results, you will likely have duplicates of several papers. Deduplication is the process of removing these duplicate articles from your compiled search results.

This can be done within Covidence. A walkthrough of Covidence's deduplication process is linked below:

Zotero can also help you deduplicate your results. More information can be found at:

Step 3.2: Title and Abstract Screening

Once you've ensured there's only one copy of each article in your collection of studies, you can begin the screening process. Since you have downloaded all the results from your searches, there will likely be many articles not relevant to your research question. That's why the screening process exists: to filter out the irrelevant studies.

Use the inclusion/exclusion criteria you developed to guide your team in selecting articles. Before starting the title/abstract screening process in earnest, pilot the screening process with a small sample of all the articles to be screened. This ensures everyone on your team interprets the inclusion/exclusion criteria consistently and clearly.

When your team is ready to begin screening, members will read the titles and abstracts of each article to determine if they meet the inclusion/exclusion criteria.

  • If at least two members of your team independently conclude that an article should be included, it moves to the next round.
  • If two team members choose to exclude it, the article is discarded.
  • In case of a disagreement, a third team member will break the tie.

Step 3.3: Full-text Screening

The next step is full-text screening. Two team members review the full text of the remaining articles to ensure that the papers are still relevant and independently determine whether to include or exclude them based on the established criteria. They will look for relevant outcomes reported. If an article is to be excluded, the reasons must be explicitly noted.

All screening steps can be facilitated by Covidence. To learn more, you can access Covidence's in-depth walkthrough at the link below:

Appraisal Tools

There are several tools available to standardize your team's approach to assessing studies for validity and risk of bias. These tools are organized according to the type of study being appraised.

Repositories of Tools for Various Study Types

Tools for Non-randomized Studies or Observational Studies

Tools for Reviews

Tools for Grey Literature

Data Extraction

The next step is to review the selected studies and extract the data relevant to your research question. To ensure consistency across team members and studies, use a standardized data extraction form. Templates are provided in the Data Extraction Forms section below.

As with the screening process, it is a good idea to pilot your data extraction form with a sample of all the papers to be extracted to ensure your team is aligned. Two independent team members should extract the data from each article, with any discrepancies resolved by a third team member.

Creating an effective data-extraction form and managing the data requires careful planning. Below are some materials to help your team determine the best way to extract data for your study:

Data Extraction Forms

At this stage, each full-text article must be read through to extract the data relevant to your research question. This is often done with a data extraction form, table, or spreadsheet.

Example of a Data Extraction Form

Reporting guidelines

When writing your report, it is important to know what information to include. Using an appropriate reporting guideline can help. The specific guideline you need will depend on the type of review you're doing and your subject area. Additionally, check the submission guidelines of your preferred journal to understand their standards.

Reporting your searching

The PRISMA-S (also called PRISMA-Search) is a PRISMA extension to help document your knowledge synthesis literature searches. It is a 16-item checklist that helps to ensure all important information regarding your searching is recorded comprehensively and accurately.

Note: PRISMA-S is a tool to help document your searches. It is not a guide for how you should search.