A systematic review search strategy is part of the research method, not a quick database task. The strategy should connect the review question and eligibility criteria with searchable concepts, controlled vocabulary, natural-language terms, appropriate databases and a complete record of what was done.
A search that is too narrow can miss eligible evidence and distort the synthesis. A search that is unnecessarily broad can produce thousands of irrelevant records without improving coverage. The aim is a sensitive, transparent and reproducible strategy that fits the nursing or healthcare question.
What a defensible search needs
- A final or near-final review question and eligibility criteria.
- Essential searchable concepts rather than every word in the question.
- Controlled vocabulary and tested text-word synonyms.
- Correct Boolean, field, phrase, truncation and proximity logic.
- Database selection that matches the topic and review type.
- Translation of the strategy for each database and platform.
- Testing against known relevant studies.
- Complete search dates, result counts and deduplication records.
- PRISMA-S reporting and peer review where feasible.
Start with the review question
Do not build a long search string while the population, intervention, phenomenon, context or study design remains uncertain. The question determines the concepts and the eligibility criteria determine which evidence qualifies.
Choose a framework that fits the review purpose. PICO is useful for many intervention questions, PICo or SPIDER can support qualitative questions and PCC is commonly used for scoping reviews. The nursing research-question guide compares these frameworks.
Suppose the review asks:
Among adults with type 2 diabetes, does nurse-led education improve glycaemic control compared with usual care?
The essential search concepts may be type 2 diabetes and nurse-led education. The comparator and outcome may be important for eligibility and synthesis without needing to appear as mandatory search blocks.
Convert the question into searchable concepts
Create a concept table before opening a database.
| Concept | Controlled vocabulary | Possible text words |
|---|---|---|
| Type 2 diabetes | Diabetes Mellitus, Type 2 | type 2 diabetes, type II diabetes, T2DM |
| Nurse-led education | Patient Education as Topic; Nurses | nurse-led education, nurse delivered education, nurse-led self-management support |
Keep the eligibility criteria beside the table. Not every exclusion criterion needs to become a NOT command. Aggressive exclusions can remove eligible studies that mention both relevant and irrelevant populations.
Find terminology systematically
Authors may describe the same concept using abbreviations, former terminology, regional spelling or different professional language. Candidate terms can come from:
- titles, abstracts and author keywords of known relevant papers;
- database subject headings assigned to those records;
- MeSH, CINAHL Headings, Emtree or another relevant thesaurus;
- clinical guidelines and professional documents; and
- discussion with a librarian, information specialist or subject expert.
Search both the full form and common acronyms where relevant. A strategy using only T2DM or only the full phrase may miss useful records.
Combine subject headings with free text
Controlled vocabulary helps retrieve records indexed under a standard concept even when authors use different terms. MEDLINE uses MeSH, Embase uses Emtree and CINAHL has its own headings. These vocabularies should not be treated as interchangeable.
Free-text terms remain necessary because newly added records may not yet be indexed and emerging terminology may not have a precise subject heading.
A simplified PubMed concept block could be:
"Diabetes Mellitus, Type 2"[Mesh]
OR "type 2 diabetes"[tiab]
OR "type II diabetes"[tiab]
OR T2DM[tiab]The final strategy should be tested against the exact protocol and current database interface.
Use Boolean logic correctly
- OR joins synonyms or alternative expressions.
- AND joins different concepts.
- NOT excludes records and should be used cautiously.
The basic structure is:
(synonym A OR synonym B OR synonym C)
AND
(synonym D OR synonym E OR synonym F)Parentheses preserve the intended logic. Do not use AND between two synonyms merely because both describe the topic; that requires records to contain both expressions and can reduce recall unnecessarily.
Use phrase searching, truncation and proximity carefully
Quotation marks can search exact phrases, but exact matching may exclude useful wording variations. Truncation retrieves words sharing a root. Proximity operators can find terms that occur close together even when wording changes.
The syntax differs by platform. Ovid, EBSCOhost, Embase.com and PubMed do not use identical commands. Always check the platform guidance and test what the operator actually retrieves.
Select databases that fit the nursing question
Database selection should follow the discipline, population, intervention or phenomenon and study designs likely to be relevant.
Depending on the question, nursing and health reviews may use MEDLINE/PubMed, CINAHL, Embase, PsycINFO, CENTRAL, Scopus or other specialist sources. One database is rarely sufficient for a full systematic review.
Cochrane intervention-review guidance has its own expectations for core sources, while other review types should justify a source set appropriate to their objectives (Lefebvre et al., 2025).
Supplementary searching may include trial registers, citation searching, reference lists, conference proceedings, theses and selected grey-literature sources when justified by the protocol.
Build the main strategy one concept at a time
Develop the strategy first in the database most central to the topic. Test each concept block separately before combining them.
A simplified PubMed example is:
(
"Diabetes Mellitus, Type 2"[Mesh]
OR "type 2 diabetes"[tiab]
OR "type II diabetes"[tiab]
OR T2DM[tiab]
)
AND
(
"Patient Education as Topic"[Mesh]
OR "nurse-led education"[tiab]
OR "nurse led education"[tiab]
OR "nurse-delivered education"[tiab]
OR "nurse led self-management"[tiab]
)This demonstrates structure only. A submitted strategy needs terms, headings and syntax tested for the specific review.
Test against known relevant studies
Identify several papers that should be retrieved if they meet the review criteria. If the strategy misses one, investigate why.
The study may use a missing synonym, an unexpected subject heading or wording excluded by a restrictive concept block. Correct the search logic rather than inserting the article title into the query.
Refine the search transparently
Review a sample of retrieved records. Irrelevant papers can reveal ambiguous search terms, while relevant records can identify missing vocabulary.
Ask:
- Are known eligible studies retrieved?
- Does every concept block add useful discrimination?
- Are important synonyms or headings missing?
- Is exact phrase searching too restrictive?
- Does truncation produce unrelated terms?
- Are language, date or design limits justified?
A large result count does not automatically mean the search is poor. Systematic searches often accept lower precision to protect sensitivity.
Translate—not copy—the strategy
Preserve the concepts when moving to another database, but rebuild the search using that database’s subject headings, fields, truncation and proximity syntax.
A MeSH heading may need an Emtree equivalent in Embase. PubMed’s [tiab] field tag will not work in CINAHL. Platform information should therefore be reported as well as the database name.
Use peer review where possible
The PRESS guideline provides a structured approach to reviewing electronic search strategies, including question translation, Boolean/proximity logic, headings, text words, syntax, spelling, limits and filters (McGowan et al., 2016).
An experienced information specialist can identify errors before screening begins. This is particularly useful for high-stakes systematic reviews and complicated searches.
Document the search from the beginning
| Record | What to save |
|---|---|
| Source | Database, register, website or citation source |
| Platform | PubMed, Ovid, EBSCOhost, ProQuest or another interface |
| Date | Exact date the search was run |
| Strategy | Complete query exactly as executed |
| Limits | Language, date or study-type restrictions with rationale |
| Results | Number retrieved before deduplication |
| Export | File type and destination |
| Deduplication | Software and procedure |
| Update | Date and result of any later search |
Create the search log prospectively. Retrospective reconstruction is difficult and can introduce inconsistencies.
Report with PRISMA-S
PRISMA-S is a 16-item extension designed to improve reporting of systematic-review literature searches (Rethlefsen et al., 2021). It asks authors to name databases and platforms, describe other sources, provide full strategies, report limits and dates, and explain record management.
Use PRISMA-S alongside the main reporting guideline appropriate to the review; it does not replace the review methodology itself.
Plan search updates
Reviews can become outdated during screening and writing. Record the final search date and follow the relevant programme, journal or review-method requirements for updating the search before submission or publication.
Common search-strategy mistakes
- Searching the complete research question as one sentence.
- Making every PICO element mandatory.
- Using keywords without controlled vocabulary where headings are available.
- Copying one search syntax into every database.
- Applying language, date or full-text limits for convenience rather than methodological reasons.
- Using NOT without testing what is excluded.
- Reporting only a keyword list rather than the exact search.
- Failing to record dates, platforms, counts and deduplication.
From searching to synthesis
The search is one part of the evidence-review chain. After deduplication, records move through screening, appraisal, extraction and synthesis according to the approved methodology.
Use the nursing literature review support route for broader synthesis problems or systematic review support when the full review method needs project-specific feedback.
Final search-strategy checklist
- The review question and eligibility criteria are stable enough to guide searching.
- The correct question framework is being used.
- Only essential concepts are made mandatory.
- Synonyms, abbreviations and spelling variants have been tested.
- Controlled vocabulary and free-text terms are combined.
- Boolean logic and parentheses are correct.
- Field codes, phrase searching, truncation and proximity commands work as intended.
- Known relevant studies are retrieved.
- Database selection fits the review question.
- Supplementary sources are justified where used.
- The strategy is translated for each database/platform.
- Limits are minimal and protocol-based.
- Search dates, result counts and deduplication are recorded.
- Reporting follows PRISMA-S and the relevant main review guideline.
Frequently asked questions
How many databases should a systematic review search?
There is no universal number. Selection depends on the discipline, question, study designs and governing methodology. The source set should be justified rather than chosen by convenience.
Should the outcome be included in a PICO search?
Not automatically. Outcome terms may be inconsistently reported and can reduce recall. Test whether they add useful precision without losing known relevant studies.
What is the difference between keywords and MeSH?
Keywords search author language in selected fields. MeSH is controlled vocabulary used to index MEDLINE concepts. Combining controlled vocabulary with free text improves coverage.
Can the same search string be used in every database?
No. Preserve the concepts but translate subject headings, field codes, truncation and proximity syntax for every database and platform.
Is Google Scholar enough for a systematic review?
It should not normally replace planned searches of appropriate bibliographic databases. It may have a supplementary role when justified and reported transparently.
Related Nursing Guides
- How to Conduct Thematic Analysis in a Systematic Review: Nursing Guide
- Mental Health Nursing Systematic Review Topics: 40 Focused Ideas
- How to Conduct a Critical Appraisal in Nursing Research: CASP, JBI and Other Tools
Conclusion
A strong systematic-review search strategy connects the review question to a transparent retrieval method. Define the concepts, identify controlled vocabulary and text words, test Boolean logic, choose suitable sources, translate the strategy carefully and preserve a complete record.
If you already have a draft search strategy and need a methodological review, use nursing systematic review support or the contact page.
References
- Lefebvre, C., Glanville, J., Briscoe, S., et al. (2025). Chapter 4: Searching for and selecting studies. In Cochrane handbook for systematic reviews of interventions (Version 6.5.1). Cochrane. https://www.cochrane.org/authors/handbooks-and-manuals/handbook/current/chapter-04
- McGowan, J., Sampson, M., Salzwedel, D. M., Cogo, E., Foerster, V., & Lefebvre, C. (2016). PRESS Peer Review of Electronic Search Strategies: 2015 Guideline Statement. Journal of Clinical Epidemiology, 75, 40–46. https://doi.org/10.1016/j.jclinepi.2016.01.021
- National Library of Medicine. (n.d.). Medical Subject Headings (MeSH). https://www.ncbi.nlm.nih.gov/mesh/
- Rethlefsen, M. L., Kirtley, S., Waffenschmidt, S., et al. (2021). PRISMA-S: An extension to the PRISMA Statement for Reporting Literature Searches in Systematic Reviews. Systematic Reviews, 10, 39. https://doi.org/10.1186/s13643-020-01542-z
