Notes on PubMed search strategy, Cochrane review usage, and multi-database literature searching for ankle sprains and ultrasound

PubMed and Library Tools

• Library navigation in this course focuses on Neder/Nether library as the primary hub for finding professional-level content; the left-hand block is less used, and the main box in the page provides links to databases with easy access to commonly needed resources.
• Primary databases discussed: PubMed, Cochrane, CINAHL, and Site Info (with other databases available).
• Rationale: these databases cover the most relevant clinical and evidence-based material for physical therapy topics like ankle sprains.
• Task context: the instructor is using a PECO-based question about ankle sprains and an intervention (ultrasound) to illustrate searching and evidence appraisal.

PubMed: starting point and search strategy

• PECO focus: Population (people with ankle sprains), Exposure/Intervention (ultrasound), Comparison, Outcome (recovery/therapy outcomes) used to frame the search.
• Access PubMed via the Neder library and switch to PubMed for building the search strategy.
• What is a search strategy?
  • A structured combination of keywords and terms using Boolean operators (AND, OR) and grouping via parentheses to connect concepts.
  • The goal is to create a concise, reusable search string to query over time.
• Auto-suggestions: typing a term (e.g., $ankle$) triggers the system to suggest related terms; these suggestions are dynamic and change over time as user behavior evolves.
• Example initial search term: $ankle$; results show suggestions and an initial hit set; the system may surface related concepts like $ankle sprain$ as a separate item.
• Hits and page structure:
  • Example: total hits around $27{,}000$ references for a broad ankle-related search.
  • The system typically returns about $2{,}700$ hits per page, illustrating large initial results when the search is broad.
• First reference (example, older study):
  • Publication date: $2007$ (02/2007).
  • Nature: free access article.
  • Data set: large samples (e.g., >$200{,}000$ patients) and tens of thousands of ankle sprains in the abstract.
  • Significance: demonstrates how older articles can be foundational but may be dated with newer evidence.
• Second reference (example, newer study):
  • Publication date: $2022$.
  • Nature: systematic review.
  • Significance: indicates a much more current synthesis of evidence and larger aggregated datasets through other studies (not necessarily a single large dataset).
  • The overlap: both are systematic reviews, illustrating how reviews build on prior primary studies.
• Key takeaway: broad searches yield many results; systematic reviews are valuable anchors but may require updating.

Refining the search: quotes, nesting, and the advanced search builder

• To target a precise phrase, use quotes around the exact phrase: for example, $"ankle sprain"$.
  • Result: reduces the hit count significantly (illustrated as a ~90% reduction in the example), narrowing to articles specifically mentioning the exact phrase.
• How quotes affect results:
  • Without quotes: the system searches for articles containing ankle and sprain separately, which can return many irrelevant hits.
  • With quotes: the search engine looks for the exact phrase, reducing extraneous results.
• Advanced search builder:
  • Allows step-by-step construction of the query (adding terms, changing operators, and visualizing parentheses).
  • Example flow: add $ankle sprain$ as one block, then add $ultrasound$ as a second block with operator "AND" to require both terms.
  • Parentheses appear around each block to clarify the logical grouping (e.g., (ankle sprain) AND (ultrasound)).
  • You can save the history of searches to reuse and modify later.
• Effects of using AND vs OR:
  • AND narrows the search by requiring both concepts to appear.
  • OR broadens by including either concept.
  • Poor grouping with OR without proper parentheses can yield an explosion of results where one term appears in many articles, diluting relevance.
• Practical workflow example:
  • Start with a broad query: ankle AND ultrasound.
  • Narrow by adding ("ankle sprain") with quotes to force exact phrase matching.
  • Use the builder to add a third term related to the outcome or modality, for example, ("recovery" OR "functional outcome") AND ultrasound.
  • If results become unwieldy, reintroduce parentheses and filters to strike a balance between breadth and relevance.

A focused search path: ankle sprain + ultrasound; then broaden for context

• Broad search: ankle sprain AND ultrasound → hits around a few hundred to a few thousand depending on filters.
• Narrow search by modality or context: add a more specific ultrasound context (e.g., therapeutic ultrasound) or look for diagnostic ultrasound vs therapeutic ultrasound.
• Narrowed example: (ankle sprain) AND (therapeutic ultrasound) → may yield very few hits (even as low as 1 in the example) depending on indexing.
• If results are too narrow to be useful, broaden again by removing strict phrases or parentheses and use the builder to re-balance the query.
• Note on date and article type filters:
  • PubMed allows custom date ranges (e.g., last year, last five years) to limit results to newer evidence.
  • You can filter by article type to pull meta-analyses or systematic reviews when needed, which is useful for literature synthesis.
• Practical takeaway: use a combination of exact phrases and strategic broadening to explore the evidence landscape efficiently.

Multi-database searching and practical tips

• It is essential to search more than one database to capture a wider swath of literature; no single database contains everything.
• Recommended approach: search in 3–4 databases to ensure coverage (e.g., PubMed, CINAHL, Cochrane, and another database like Hall or Netter-accessed search tool if available).
• Example rationale:
  • PubMed provides broad, high-tier journal coverage and strong indexing for medical literature.
  • CINAHL covers nursing and allied health literature, which can include useful clinical studies not indexed in PubMed.
  • Cochrane provides high-quality, pre-appraised systematic reviews that can anchor the evidence base and show synthesis of multiple trials.
• Practical exercise: for your PICO/PECO question, run the same search strategy in PubMed and in a second database (e.g., CINAHL or Hall) to compare results and identify additional relevant studies.
• Homework deliverable: provide a screenshot showing your search strategy and sample results from two search engines, demonstrating a valid, reproducible approach.

Cochrane database: what it offers and how to use it

• Cochrane reviews are synthesized, systematic reviews focusing on timely, important clinical questions and updated periodically by professionals in the field.
• How to use Cochrane:
  • Look for current Cochrane reviews relevant to your question (e.g., ankle sprains in older adults, fall prevention, etc.).
  • Read the abstract to get the rationale, search strategy, and synthesis of results.
  • Review the synthesis of results to identify key conclusions and the magnitude of effects.
• Example discussed: preventing falls in older people in care facilities.
  • Headline synthesis topics include: individualized care, staff engagement, exercise, medication optimization, and increased dairy intake.
  • Exercise is highlighted as beneficial, but sustained, ongoing exercise is necessary for lasting effects.
  • Medication optimization showed little difference; vitamin D had limited impact; increasing dairy reduced fall risk and fracture risk.
  • Metrics to indicate evidence strength: reports of the number of trials and participants (e.g., seven trials with ~$3{,}500$ participants; five trials with ~$3{,}000$ participants).
• Ankle sprains in Cochrane:
  • A Cochrane review on ankle sprains exists from $02/2011$, indicating early systematic work in that area; note that it is well over a decade old and may prompt an update by the Cochrane Collaboration.
• Practical takeaway: use Cochrane for high-quality syntheses to inform your understanding and to guide search iterations for newer studies.

Hall, Netter, and multi-database navigation

• The Netter (Nether) library page includes a section that functions similarly to PubMed's advanced search, enabling cross-database querying.
• The idea is to compare and search across different platforms (e.g., PubMed vs. Cochrane vs. CINAHL) to ensure comprehensive coverage.
• The instructor demonstrates switching to Hall (an advanced search interface) to run ankle sprain and therapeutic ultrasound searches in a different system to illustrate consistency and differences among databases.
• Key point: different search engines pull from different journals and indexing; PubMed tends to include higher-tier journals, while some other databases may index different sets of journals.
• Practical implication: always search multiple databases to avoid missing relevant literature.

Homework and exam prep tasks

• Homework assignment:
  • Revisit your PICO/PECO question and construct a search strategy in PubMed and a second search engine (e.g., CINAHL or Hall).
  • Take a screenshot showing both your search strategy and the results obtained to submit to Blackboard.
• Exam prerequisites:
  • You will need a calculator for exams 3, 2–4; you cannot use a phone, even a basic $2 calculator is acceptable.
  • LockDown Browser will be required for exams; ensure it works before test day.
  • If you have issues with LockDown Browser, resolve them before the exam date.

Practical lessons and takeaways

• Build a solid search strategy before diving into databases:
  • Start with key terms related to the population and intervention (e.g., ankle sprain, ultrasound).
  • Use exact phrases with quotes for precision (e.g., $"ankle sprain"$).
  • Use AND to connect concepts and OR to expand within a concept cluster.
  • Use parentheses to group terms properly and to control the order of operations.
• Use the advanced search builder to visualize and refine structure before typing long strings.
• Always consider multiple databases to capture a full range of evidence; PubMed is strong, but other databases add unique records.
• Narrowing and broadening are a balance:
  • Too broad: too many irrelevant hits.
  • Too narrow: risk missing important studies.
  • Iterative adjustment with parentheses and quotes helps tailor results.
• Identify and extract key evidence types from results:
  • Primary studies vs. systematic reviews vs. meta-analyses.
  • Use filters to target study type (e.g., meta-analyses) for specific homework tasks.
• Real-world relevance: the process models how clinicians and researchers stay current with evolving evidence, manage time, and formulate evidence-based interventions.
• Ethical/practical note: access to full texts sometimes depends on licensing; use abstracts to gauge relevance when full text is not freely accessible.
• Reflection on statistical scale and scope:
  • Older articles can have very large data sets (e.g., >$2{,}00000$ patients) but may be outdated.
  • Large, current systematic reviews provide up-to-date synthesis across many studies.
• Final reminder for exams and assignments:
  • Practice with real search tasks, document your strategy and results, and be prepared to justify your search choices.
  • Have a plan to verify the currency and relevance of sources as you advance in your coursework.

Numerical references (for quick review)

• Initial search breadth: large hit counts around $27{,}000$ references; per-page hits around $2{,}700$.
• Example older study: published in $2007$; very old in the context of ongoing clinical guidelines.
• Example newer study: published in $2022$; more current synthesis.
• Cochrane falls review: $02/2011$ (ankle sprains), with later relevance uncertain without updates.
• Falls prevention synthesis: seven trials with $3{,}500$ participants; five trials with $3{,}000$ participants.
• Publication years cited for example articles: $2013$ (therapeutic ultrasound topic in a focused search), $2011$ (Cochrane ankle sprain review).
• Exercise duration note: benefits require sustained engagement beyond short-term programs.

Quick reference tips

• PECO: Population, Exposure/Intervention, Comparison, Outcome (use to frame questions).
• Use PubMed advanced search for structured, reusable queries; leverage the builder to manage parentheses and operators.
• Always cross-check results with a second database to ensure comprehensive coverage.
• Apply filters judiciously (date ranges, article type like meta-analyses) to align with your assignment or exam requirements.
• Consider setting up alerts or RSS feeds for ongoing topic monitoring if needed for long-term projects.
• For exams: verify LockDown Browser compatibility and bring a basic calculator.