Quick answer: The most common literature search strategy mistakes are searching only one database, relying on natural-language terms instead of controlled vocabulary (like MeSH), overly narrow Boolean logic, ignoring synonyms and spelling variants, skipping a validation/pilot search, and failing to document the final search string for reproducibility. Each of these directly increases the risk of missing relevant studies.
A flawed search strategy is invisible until someone else tries to reproduce it — or until a peer reviewer or regulator asks why an obviously relevant study wasn't included. These six mistakes account for the large majority of search strategy problems we see.
Mistake 1: Searching Only One Database
No single database indexes all relevant biomedical literature. A search limited to PubMed alone will systematically miss studies that only appear in Embase, industry conference proceedings, or nursing/allied-health databases like CINAHL — regardless of how well-constructed the search string itself is.
Fix: Search at least two major databases with different indexing systems (commonly PubMed plus Embase), and add a trial registry search when the topic involves interventions.
Mistake 2: Relying on Free-Text Terms Instead of Controlled Vocabulary
Typing a disease or drug name as free text misses studies that were indexed using a different but related MeSH or Emtree term, or that use a synonym in the title/abstract that your search string didn't anticipate.
Fix: Build search strings using the relevant controlled vocabulary (MeSH for PubMed, Emtree for Embase) combined with free-text terms as a supplement, not a replacement.
Mistake 3: Overly Narrow Boolean Logic
Using too many restrictive AND conditions, or failing to account for synonym variation with OR groupings, silently excludes relevant studies before a human reviewer ever sees them.
Fix: Build broad OR groups for each concept (synonyms, spelling variants, acronyms) and combine concept groups with AND — err toward a wider initial pull that gets filtered during screening, rather than an overly narrow search that filters at the search-string level where mistakes are invisible.
Mistake 4: Ignoring Spelling and Terminology Variants
British vs. American spelling (tumour/tumor), older vs. newer drug nomenclature, and abbreviation variants can all cause a search to miss relevant records without any error message or warning.
Fix: Explicitly include known spelling and terminology variants in your OR groups, and consult a subject-matter expert or medical librarian if the field has known naming inconsistencies.
Mistake 5: Skipping a Pilot/Validation Search
Running the full search strategy for the first time as your "real" search means any flaw only becomes visible after screening is already underway — often too late to fix without redoing work.
Fix: Run a small pilot search first, and manually check whether a handful of known, highly relevant "benchmark" studies are captured. If they're missing, the strategy needs revision before the full search is run.
Mistake 6: Not Documenting the Final Search String
Even a well-constructed search strategy is a liability if it isn't documented — you (or a reviewer) can't reproduce or defend a search you can't reconstruct exactly.
Fix: Save and timestamp the exact search string used in each database, along with the date the search was run and the number of results returned — before screening begins, not after.
Frequently Asked Questions
How do I know if my search strategy is too narrow? Run a pilot search and check whether known, highly-cited relevant studies appear in the results. If any are missing, your search is too narrow and needs revision before proceeding.
Should a librarian be involved in search strategy design? For regulatory-grade or publication-bound systematic reviews, yes — many methodological guidelines specifically recommend involving an information specialist or medical librarian in search strategy development.
How often should search strategies be re-run before finalizing a review? Best practice is to re-run the final search close to the review's completion date to capture any newly published studies, since search-to-publication gaps of even a few months can mean missing recent literature.
Can AI tools help avoid these mistakes? Yes — AI-assisted search strategy tools can suggest synonym and MeSH term expansions a human reviewer might miss, though the final strategy should still be validated against known benchmark studies.
Want a second set of eyes on your search strategy before you run it? SkyWeb Service offers a free search strategy review to catch these issues before screening begins.
Comments
Post a Comment