Conducting a systematic search and documenting it thoroughly is the cornerstone of a high-quality literature review or meta-analysis. It ensures transparency, reproducibility, and minimizes bias.

Here's a comprehensive guide:

Guide to Conducting a Systematic Search and Documenting It

Phase 0: Preparation – Laying the Groundwork

Before you even touch a database, clear preparation is crucial.

• Define Your Research Question (PICO/PEO/SPIDER):


• PICO (Quantitative/Intervention studies): Population/Patient, Intervention, Comparison, Outcome.


• PEO (Qualitative/Experience studies): Population/Problem, Exposure, Outcome.


• SPIDER (Qualitative/Synthesis): Sample, Phenomenon of Interest, Design, Evaluation, Research type.


• Why it's important: A well-defined question will guide your search terms and inclusion/exclusion criteria.


• Documentation: State your precise research question clearly.


• Develop Clear Inclusion and Exclusion Criteria:


• Based on your research question, establish upfront what types of studies you will include and exclude. Be specific.


• Examples:


• Study Design: (e.g., RCTs only, qualitative studies, systematic reviews)


• Population: (e.g., adults aged 18-65, patients with type 2 diabetes)


• Intervention/Exposure: (e.g., specific drug, mindfulness-based therapy, exposure to air pollution)


• Outcome: (e.g., reduction in blood pressure, improved quality of life, experiences of caregivers)


• Language: (e.g., English only)


• Publication Date: (e.g., published between 2010-2023)


• Publication Type: (e.g., peer-reviewed journals, dissertations, conference proceedings)


• Why it's important: These criteria will be used during the screening process and must be consistently applied.


• Documentation: List all inclusion and exclusion criteria clearly, with rationales if necessary.


• Create a Study Protocol (and Register it):


• A protocol outlines your entire methodology before you start. It should detail your research question, search strategy, selection process, data extraction, quality assessment, and synthesis methods.


• Registration: For systematic reviews and meta-analyses, consider registering your protocol with a public registry like PROSPERO (for health and social care topics) or Open Science Framework (OSF). This prevents duplication and reduces reporting bias.


• Why it's important: Ensures transparency, reduces bias, and helps keep your project on track.


• Documentation: Include a reference to your registered protocol (if applicable) and/or provide the full protocol as an appendix or supplementary material.

Phase 1: Developing the Search Strategy

This is an iterative process that requires careful thought and testing.

• Identify Key Concepts/Keywords:


• Break down your research question into its core components (P, I, C, O, etc.).


• For each component, brainstorm a comprehensive list of keywords.


• Consider:


• Synonyms (e.g., "mindfulness," "meditation")


• Related terms (e.g., "anxiety," "stress," "depression")


• Broader and narrower terms


• American and British spellings (e.g., "behavior," "behaviour")


• Singular and plural forms (e.g., "child," "children")


• Acronyms and full names (e.g., "PTSD," "Post-Traumatic Stress Disorder")


• Documentation: Create a table or list for each concept with all identified keywords.


• Utilize Controlled Vocabulary (Subject Headings):


• Databases use specific subject headings (controlled vocabulary) to categorize articles consistently, regardless of the author's chosen keywords.


• Examples:


• PubMed/MEDLINE: Medical Subject Headings (MeSH)


• Embase: Emtree


• CINAHL: CINAHL Headings


• PsycINFO: APA Thesaurus of Psychological Index Terms


• How to find them: Use the database's thesaurus or MeSH database (for PubMed/MEDLINE) to identify relevant subject headings. Look at highly relevant articles you've found manually and see what subject headings they are tagged with.


• Why it's important: Captures articles that might not use your chosen keywords in their title/abstract.


• Documentation: List the controlled vocabulary terms used for each concept and which database they were applied to.


• Combine Terms with Boolean Operators:


• AND: Narrows your search; retrieves results containing all specified terms (e.g., "mindfulness AND depression").


• OR: Broadens your search; retrieves results containing any of the specified terms (e.g., "mindfulness OR meditation"). Use OR within a concept.


• NOT: Excludes terms (use with caution, as it can inadvertently exclude relevant articles) (e.g., "cancer NOT breast").


• Parentheses: Use to group terms logically (e.g., (mindfulness OR meditation) AND (depression OR anxiety)).


• Documentation: Show how you combined terms with Boolean operators in your full search strings.


• Employ Truncation, Wildcards, and Proximity Operators:


• **Truncation (* or $)**: Finds variations of a word stem (e.g., depress* finds depression, depressive, depressed).


• Wildcards (? or #): Replaces a single character within a word (e.g., wom?n finds woman, women).


• Proximity Operators (Near or ADJ): Specifies how close terms should be to each other (e.g., mindfulness N3 therapy finds mindfulness within 3 words of therapy). Syntax varies by database.


• Why it's important: Increases search sensitivity and captures more relevant articles.


• Documentation: Specify which operators were used and their meaning if database-specific.


• Develop a Search String for at least One Major Database (Pilot Search):


• Start by building a detailed search string for one key database (e.g., PubMed or Embase), as it often has the most sophisticated controlled vocabulary.


• Test and Refine: Run your pilot search.


• Too many results? Add more specific terms, use AND, add limits, or refine subject headings.


• Too few results? Add more synonyms, broader terms, use OR, or remove limits.


• Review the titles and abstracts of the first few pages of results to see if they are relevant and if you're missing any obvious keywords or subject headings.


• Documentation: Record each version of your pilot search string and the number of results, along with notes on what you changed and why.

Phase 2: Executing the Search

Once your core strategy is robust, you'll execute it across multiple platforms.

• Select Appropriate Databases:


• Choose databases relevant to your field and research question.


• General/Major Databases: PubMed/MEDLINE, Embase, Web of Science (Core Collection), Scopus, Cochrane Library.


• Specialized Databases (Examples):


• Health: CINAHL (nursing), PsycINFO (psychology), LILACS, Global Health, SportDiscus.


• Social Sciences/Education: ERIC, Sociological Abstracts, Education Source.


• Business: Business Source Premier.


• Why it's important: Maximizes coverage and ensures you don't miss key studies.


• Documentation: List all databases searched and provide a rationale for their selection.


• Translate Search Strategy for Each Database:


• Crucial Step: You cannot simply copy-paste your search string from one database to another.


• Adapt Syntax: Each database has unique syntax for subject headings, field codes (e.g., [tiab] for title/abstract in PubMed, exp for exploded MeSH), truncation, and proximity operators.


• Run Searches: Execute your adapted search strategy in each chosen database.


• Apply Limits: If appropriate, apply language, date, or study type limits within the database interface.


• Documentation: This is critical. For each database, document the exact, full search string used, including all lines, limits, and the date the search was conducted.


• Save/Export Results:


• Export all search results from each database in a compatible format (e.g., RIS, CSV, plain text).


• Documentation: Record the total number of hits for each database search.

Phase 3: Managing Results

After collecting all your search results, you need to manage them efficiently.

• Import into a Reference Management Software:


• Use tools like EndNote, Zotero, Mendeley, Covidence, Rayyan, or EPPI-Reviewer 4.


• Why it's important: These tools help organize, deduplicate, and facilitate the screening process.


• Documentation: Note which software you used.


• Deduplicate Records:


• After importing all results, use your reference manager's deduplication function (and/or manual review) to remove duplicate entries.


• Why it's important: Prevents repeated screening of the same article and inflates your numbers.


• Documentation: Record the number of duplicates removed.

Phase 4: Expanding the Search (Beyond Databases)

Systematic reviews often go beyond traditional database searches to ensure comprehensive coverage.

• Hand-searching Reference Lists:


• Examine the reference lists of included articles (and highly relevant systematic reviews/guidelines) for additional relevant studies.


• Documentation: Document which key articles' reference lists were searched and any new studies identified.


• Citation Searching:


• Use tools like Web of Science or Scopus to identify articles that have cited your key included studies.


• Documentation: Note if citation searching was performed and any results.


• Grey Literature Search:


• Search for unpublished or non-commercial literature. This helps reduce publication bias.


• Sources:


• Government reports, policy documents


• Theses and dissertations (e.g., ProQuest Dissertations & Theses)


• Conference proceedings (e.g., Web of Science Conference Proceedings Citation Index)


• Clinical trial registries (e.g., ClinicalTrials.gov, WHO ICTRP)


• Professional association websites


• Specialized grey literature repositories (e.g., OpenGrey)


• Documentation: List all grey literature sources searched and how they were searched (keywords, specific pages).


• Contacting Experts:


• Reach out to researchers in the field to inquire about ongoing or unpublished studies.


• Documentation: Note if experts were contacted and any studies identified.

Phase 5: Comprehensive Documentation (The Reporting Phase)

Your documentation should be so thorough that another researcher could replicate your exact search.

• Search Strategy Table/Appendix:


• Present your full search strategy for EACH database. This is typically done in an appendix or supplementary material.


• For each database, include:


• Database name


• Date of search


• The complete, line-by-line search string, showing all terms, Boolean operators, truncation, field codes, and any applied filters/limits.


• Number of results retrieved.


• Example (simplified for PubMed):

Database: PubMed (via NCBI)
        Date: YYYY-MM-DD
        #1 "mindfulness"[MeSH Terms] OR mindfulness[tiab] OR "mindful based intervention"[tiab]
        #2 "Depression"[MeSH Terms] OR depression[tiab] OR depressive[tiab] OR "major depressive disorder"[tiab]
        #3 #1 AND #2
        #4 #3 Filters: English, Humans, 2010-2023

• Search Log/Record:


• Maintain a detailed log during the search process.


• Fields to include:


• Date of search


• Database name


• Search terms/strategy used (brief overview or link to full strategy)


• Number of results found


• Number of results exported


• Notes (e.g., "adjusted terms," "error occurred," "librarian consulted")


• PRISMA Flowchart:


• The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement includes a flow diagram that visually represents the flow of information through the different phases of a systematic review.


• It documents:


• Number of records identified through database searching.


• Number of additional records identified through other sources (e.g., grey literature, reference lists).


• Number of duplicates removed.


• Number of records screened (titles/abstracts).


• Number of records excluded during title/abstract screening (with reasons if possible).


• Number of full-text articles assessed for eligibility.


• Number of full-text articles excluded (with detailed reasons).


• Number of studies included in the qualitative synthesis (literature review).


• Number of studies included in the quantitative synthesis (meta-analysis).


• Why it's important: Provides a clear, standardized overview of your search and selection process.


• Documentation: Create and include the PRISMA flowchart in your article.


• Rationale for Database Selection:


• Explain why you chose specific databases for your review.


• Date Range and Language Limits:


• Clearly state any date limitations (e.g., "studies published from January 2010 to December 2023") and language restrictions (e.g., "English-language articles only").


• Software Used:


• List any software used for search management, deduplication, and screening (e.g., EndNote, Rayyan).


• Review Team:


• Mention how many reviewers were involved in the screening process (e.g., "Two independent reviewers screened titles and abstracts").

Phase 6: Best Practices and Tips

• Consult a Research Librarian: Librarians are experts in search strategies and database specificities. They can save you immense time and improve the quality of your search.


• Work in a Team: Having at least two reviewers for screening reduces bias and errors.


• Be Flexible and Iterative: The search strategy isn't set in stone from day one. You'll likely refine it as you learn more about the literature.


• Version Control: Keep track of different versions of your search strategies.


• Transparency is Key: Your documentation should allow someone else to exactly replicate your search.

By following these detailed steps for conducting and documenting your systematic search, you will build a strong, credible foundation for your literature review or meta-analysis. Good luck!