How to build a search strategy for a systematic review (PICO and Boolean)
Short answer
Frame the question with PICO (or PICOTT), turn the key concepts into blocks of controlled vocabulary plus free-text synonyms joined with OR, combine the blocks with AND, then translate the strategy to each database's syntax. Test it against studies you already know are relevant, have it peer reviewed, and report the full strategy for every source you searched.
Step by step
- 1
Write the question as PICO or PICOTT
Define the Population, Intervention (or exposure), Comparator and Outcomes, and, if useful, the Time frame and Type of study (PICOTT). The eligibility criteria come from this frame, so write it before you search. A precise question gives you a precise search; a vague one gives you thousands of irrelevant records.
- 2
Choose which concepts go into the search
Searches usually combine only two or three concepts, typically the population and the intervention, plus a study-design filter when appropriate. The Cochrane Handbook advises caution with outcome and comparator terms, because they are often poorly reported in titles and abstracts and adding them can miss eligible studies.
- 3
Collect controlled vocabulary and free-text synonyms
For each concept, find the subject headings of each database: MeSH in PubMed/MEDLINE, Emtree in Embase, DeCS in LILACS and the BVS portal. Add free-text synonyms, spelling variants, abbreviations and drug or brand names, and use truncation and phrase searching where the database supports them. Reading the indexing of a few known relevant studies is a good way to find missing terms.
- 4
Combine the blocks with Boolean operators
Join the synonyms of one concept with OR, then join the concept blocks with AND. Use NOT sparingly, since it can silently exclude relevant records. For randomized trials in MEDLINE, consider a validated filter such as the Cochrane Highly Sensitive Search Strategy instead of inventing your own.
#1 "Diabetes Mellitus, Type 2"[Mesh] OR "type 2 diabetes"[tiab] OR T2DM[tiab] #2 Metformin[Mesh] OR metformin[tiab] #3 #1 AND #2 - 5
Translate the strategy to each database
Syntax, field tags and vocabularies differ between PubMed, Embase, the Cochrane Library (CENTRAL), LILACS, SciELO, Scopus and Web of Science. Translate the logic, not the literal string, and re-map subject headings per database. For LILACS and SciELO, include Portuguese and Spanish terms, because many records are indexed in those languages.
- 6
Test, peer review and document
Check that the strategy retrieves a set of known relevant studies; if it misses some, find out why and adjust. Ask an information specialist or a colleague to review it, for example with the PRESS checklist. Record the date, database, platform and full strategy for each source, as required by PRISMA 2020 and PRISMA-S.
Beyond bibliographic databases
A comprehensive search usually also covers trial registries (such as ClinicalTrials.gov and the WHO International Clinical Trials Registry Platform), reference lists of included studies and, depending on the topic, grey literature. Avoid language restrictions unless you have a methodological reason for them, and report any you apply.
How Prolytiq does this step
In Prolytiq you start a project by writing your PICOTT question. The search-strategy step then drafts Boolean queries for nine databases from it, including PubMed, Embase, CENTRAL, LILACS and SciELO, which you review and edit before running them in each database. You then import the exported results (RIS or CSV) into the same project for deduplication and screening.
See the workflow →Frequently asked questions
What is the difference between PICO and PICOTT?
PICOTT extends PICO with Time (the follow-up or time frame of interest) and Type of study (the designs you will include). The extra elements are often used to set eligibility criteria rather than as search concepts.
Which databases should a systematic review search?
It depends on the topic, but the Cochrane Handbook expects at least MEDLINE, Embase and CENTRAL for reviews of interventions, plus subject-specific and regional databases where relevant, such as LILACS for Latin American and Caribbean literature, and trial registries.
Should I include outcome terms in my search?
Usually not. Outcomes are often not mentioned in titles and abstracts, so outcome terms can make the search miss eligible studies. Apply outcome criteria during screening instead.
Methodological sources
- Cochrane Handbook for Systematic Reviews of Interventions, chapter 4: Searching for and selecting studies
- PRISMA 2020 statement (Page et al., BMJ 2021)
- PRISMA-S: an extension to the PRISMA statement for reporting literature searches (Rethlefsen et al., 2021)
- PRESS 2015 Guideline Statement: Peer Review of Electronic Search Strategies (McGowan et al., 2016)
Run your next systematic review in Prolytiq
From research question to publishable chart, with every decision logged and exportable for peer review.