Building the search
A couple of helpful tips when building your search strategy include:
- Reviewing the methods sections and supplementary materials of published SRs to see what strategies / terminology was used. Use this to inform your search but not to determine or dictate your full strategy.
- Running a preliminary search for your topic (perhaps a similar search that was used when looking for exemplar articles in the Pre-Review Step), sort the results by Relevance or Best Match, and skim through titles and abstracts to identify terms to include in your search strategy.
Using search tools: Boolean Operators, Nesting, Field Tags
Boolean operators are a computer programming language that, in computer science, provide rules or instructions for systems with true or false directional prompts. In library science, Boolean operators provide rules or instructions to databases to make the search more precise or accurate. The basic Boolean operators used in library databases are AND, OR, and NOT.
AND - Narrows the search so it has to include ALL terms.
In this instance, we would be looking for all articles that include both "asthma" and "COVID" in the title or abstract (or perhaps in the full text of the article depending on how the specific database functions).
OR - Expands the search so it includes EITHER term.
In this instance, we would be looking for all articles that include "asthma" or "COVID" in the title or abstract (or perhaps in the full text of the article depending on how the specific database functions). Some articles can have both terms where others may only have one term.
NOT* - Narrows the search so it EXCLUDES the term.
In this instance, we're explicitly looking for articles that include "asthma" studies but DO NOT mention or include the term "COVID".
*In SRs, NOT should be used sparingly. Authors should rely on the citation screening process to exclude irrelevant articles. Consult your librarian to help with determining if you should use NOT in your search strategy.
Consider the following mathematical problem:
(6 + 4) × 2 − (10 ÷ 5) = ?
When practicing NESTING (or organizing terms inside parentheses) to calculate the expression the solution (18) is different than if you used the order of operations (12) or left-to-right (2) techniques.
The same is true for building your search. Databases use nesting to translate what you're looking for in your search strategy, so researchers should use this to their advantage to help answer their questions.
Be sure to combine terms for a single concept inside a parentheses set.
EXAMPLE:
("Neoplasms"[Mesh] OR Tumors OR Neoplasia OR Neoplasm OR Tumor OR Cancer OR “Malignant Neoplasm” OR Malignancy OR Malignancies OR “Benign Neoplasm”)
Combing your search elements is very important in any literature search but in SR especially considering the volume of results and time involved in conducting this type of study. How terms are nested and combined with Boolean operators can drastically impact search results.
Consider this search string example: (plastic OR polymers OR acrylic) AND agriculture
PLASTIC, POLYMERS, and ACRYLIC, which appear nesting or in parentheses, will be searched first & relative to AGRICULTURE, which appears outside the parentheses.
Performing this search without nesting yields 1.5M results in Web of Science. With the nesting, the search yields 35,000 results.
Field tags tell the database which field to search. Most databases have many field options to choose from, some of which are title, abstract, keyword, author, publication, and affiliation for instance. Options vary between databases, but it is a good practice (especially for SR searching) to include using field tags to EACH TERM in your search strategy. Most databases default to predetermined data points (or tags) or sometimes even overwrite your search using embedded algorithms if you don't use field tags.
EXAMPLE:
("Neoplasms"[Mesh] OR Tumors[tw] OR Neoplasia[tw] OR Neoplasm[tw] OR Tumor[tw] OR Cancer[tw] OR “Malignant Neoplasm”[tw] OR Malignancy[tw] OR Malignancies[tw] OR “Benign Neoplasm”[tw])
This example uses field tags from the PubMed database, but you'll see how they differ in other databases in the chart below. Be sure to visit each databases' search tips or help sections to learn how to use their field tags.
| Database | Field Tags | Example |
|---|---|---|
| PubMed |
|
|
| Embase |
|
|
| EBSCO Databases (CINAHL) |
|
|
| Proquest |
|
|
Translating the search to other databases
Because you will be using multiple databases for your SR, it is important to understand that one search strategy does not fit all databases. You'll need to adjust (or translate) your search across databases to retrieve results. ZSR librarians or the following resources are available to help with this process:
- Systematic Review Accelerator: Polyglot - Can translate your PubMed search to other databases
- Cochrane's Database Syntax Guide - Provides tips on syntax translation across multiple databases
- Cornell's Guide to Translating Search Strategies - Provides multiple translation resources in one research guide
Other search methods
Hand searching is a way to supplement literature searches.
A few ways to hand search are to:
- Review the reference lists (backwards citation searching) and citing articles (forward citation searching) of studies included in the review.
- Manually browsing key journals in your field to make sure no relevant articles were missed.
- Check out clinical trial registries, white papers, pharmaceutical or other corporate reports, conference proceedings, theses and dissertations, or professional association guidelines (all of these are otherwise known as grey literature).
Grey literature refers to literature not published in a traditional manner (like in a peer-reviewed academic journal) and often not retrievable through large databases like:
- Theses & dissertations
- Preprint articles
- Government Reports
- Conference proceedings
- Technical papers
- Clinical trials
- Datasets
- White Papers
Systematic review project teams should consider how grey literature can fit into their review; including grey literature can help reduce bias and increase comprehensiveness with the study.
Learn more about searching for grey literature sources on the University of North Carolina Health Sciences Library's Finding Grey Literature research guide.
Documenting the search
The quality of a SR study is only as strong as the literature searches used to identify studies. Documenting your search is not only a best practice relative to reproducibility and transparency, it's also a part of the PRISMA Statement 2020 (see the Step 2: Develop a Protocol page).
PRISMA also provides a statement and checklist for reporting literature search details in PRISMA-S (Search)
Maintaining a research log is the best way to keep track of your search and the details that are expected to be included when publishing the SR (databases searched including names, and platforms, websites, registries, and grey literature sources searched, along with any filters or limits, and dates searched. Also, include citation searching and reaching out to experts in the field if applicable.
The librarian on your project team (who creates and runs the searches) will handle the search documentation. They will also likely have had the search strategy peer-reviewed. If your team does not include a librarian, it is a best practice to consult with one in order to review the strategy and document that the search has been reviewed.
Search strategies can be managed and documented in word processing software (e.g. Microsoft Word or Excel, Google Docs or Sheets). A template to help you with documenting your search strategy is coming soon!
Updating the review
Since medical research is constantly evolving, it is a good practice to commit to updating your SR when new developments occur in your field of study. There are several other reasons to update a review including:
- changes in clinical practice or policy
- to improve precision of effect estimates
- to enable new analyses or subgroup analyses
- changes in the review question or scope
- new methodological advances
- maintaining credibility and trustworthiness
- identification of errors or omissions in the original review
The most efficient and structured way to rerun your search is to use the previously documented strategy and adjust date filters in your databases based on the latest date of the previous search. Keep in mind that terminology can change. Be sure to add any new terms or phrases to your search and document the revisions in your updated review.