Practical Literature Search & Evidence Mapping Playbook

Welcome — what this playbook helps you do

If you need to find the strongest evidence quickly and in a way someone else can reproduce, this playbook gives a compact, practical workflow: define the question, build reproducible searches, look beyond journal databases for gray literature, manage results, screen reliably, and produce a quick evidence map with basic bibliometrics and citation mapping to reveal clusters and gaps.

Core principles

• Design for reproducibility: record exact queries, dates, filters, and exports.
• Reduce bias by searching multiple sources and using transparent inclusion criteria.
• Practical rigor: balance thoroughness against the time you can commit—document decisions.

Quick workflow (high level)

1. Define the question and explicit inclusion/exclusion criteria.
2. Choose databases, controlled vocabularies, and gray literature sources.
3. Construct and test search strings; save exact queries and dates.
4. Export results, deduplicate, and load into a citation manager or screening tool.
5. Screen titles/abstracts (ideally dual independent), extract key data, and create a quick evidence map.
6. Run basic bibliometrics and citation mapping to spot influential works and gaps.
7. Document limitations and make the search reproducible for later updates.

1) Define question & inclusion criteria (practical template)

Write a short question and then a checklist of criteria you can apply quickly during screening.

• Question: (one or two sentences — population/problem, intervention/exposure, comparison if relevant, outcome, study type, date/language limits)
• Inclusion criteria (examples): empirical studies in humans, English or Spanish, published 2010–2025, study designs: RCTs, cohort, case-control, systematic reviews.
• Exclusion criteria (examples): animal-only studies, commentaries, non-empirical modeling papers.

2) Choose sources & controlled vocabularies

Use at least two academic databases plus gray literature sources.

• Academic: PubMed/MEDLINE, Scopus, Web of Science, Embase (as available)
• Broad discovery: Google Scholar, Lens.org, Dimensions
• Gray literature: preprint servers (bioRxiv, medRxiv), clinical trial registries, theses/dissertations, conference proceedings, regulatory agency reports, company white papers
• Controlled vocabularies: MeSH (PubMed), Emtree (Embase). Map your keywords to these when possible.

3) Construct reproducible search strings (examples & tips)

Record the exact syntax, database, date, and any filters. Operator syntax differs between platforms; test and adapt.

Example (PubMed-style using MeSH + keywords):

("diabetes mellitus"[MeSH] OR diabet*) AND ("telemedicine"[MeSH] OR telehealth OR teleconsult*) AND (randomized OR trial)

Example (Scopus-style using proximity and wildcards):

TITLE-ABS-KEY((diabet* W/3 telehealth) OR telemed*) AND (random* OR trial)

Tips:

• Use controlled vocabulary where available and supplement with keywords and synonyms.
• Use truncation (diabet*) to capture variants; use proximity operators to keep concepts close when supported.
• Test sensitivity: run a few known relevant papers to see if your query finds them.
• Save/export the exact query, database name, date run, and number of hits.

4) Exporting, deduplication & managing results

Export in RIS/BibTeX/EndNote formats where possible. Import into a citation manager (Zotero, Mendeley, EndNote) or a screening tool.

• Deduplication: use automated dedupe in your manager, then do a quick manual pass (check title, authors, year).
• Keep a search log: database, query, date, hits, export filename.

5) Screening & extraction — fast reproducible approach

Suggested minimal workflow for pragmatic reviews:

1. Title/abstract screening: two reviewers independently if feasible; use a third for conflicts.
2. Full-text retrieval for included abstracts.
3. Data extraction: create a simple template capturing citation, study type, population, intervention/exposure, outcomes, key results, limitations, and funding/conflict notes.
4. Record reasons for exclusion at full-text stage (needed for transparency).

Screening checklist (short)

• Does title/abstract address the defined population/problem?
• Is study design within the included designs?
• Is outcome of interest reported?
• If unclear, carry to full-text screening.

6) Quick evidence map template (practical fields)

Use a simple spreadsheet with these columns to build an evidence map you can sort and visualize:

• Citation
• Year
• Study design
• Population / setting
• Intervention / exposure
• Outcome(s)
• Key result (brief)
• Sample size
• Funding / declaration
• Notes (limitations, important context)

7) Basic bibliometrics & citation mapping (what to run)

Run a few simple measures to spot influential works and clusters:

• Top-cited papers (by citation count)
• Publication trend over time (yearly counts)
• Top journals and authors
• Keyword co-occurrence to reveal topics and clusters
• Backward & forward citation searching (snowballing) from key papers

Use citation managers and tools (Scopus, Web of Science, Lens) to generate these counts and simple co-occurrence maps. Treat numeric indicators as exploratory, not proof.

8) Reduce bias & improve credibility

• Search multiple sources and include gray literature to reduce publication bias.
• Use at least partial dual screening or random double-checks if resources are limited.
• Predefine inclusion/exclusion and stick to them; log any protocol changes.
• Document search dates and limits so others can update or replicate the work.

9) Reproducibility checklist (minimum)

• Saved search strings for each database
• Date each search was run (and timezone if helpful)
• Filters applied and versions of controlled vocabularies used
• Export filenames and formats
• Screening decisions and extraction template

10) Practical limitations & when to stop

If your goal is a rapid evidence scan, focus on high-yield sources, run citation mapping on seed papers, and be explicit about the scope and limits. For a full systematic review, follow formal protocols (e.g., register a protocol, plan dual independent screening, and consider a methodologist).

Short resources & next steps you can adapt

• Create a reusable search log spreadsheet from the templates above.
• Build a simple extraction form in your citation manager or a spreadsheet using the evidence map fields.
• Use citation network views to prioritize full-text retrieval for central clusters.

Where this playbook can go next (capability ideas): interactive screening forms, a PRISMA flow generator that fills from the search log, saved-query alerts, and an interactive citation-mapping dashboard that stores bibliometric snapshots for future updates.