Ch. 3 M&F: Finding Relevant Evidence to Answer Clinical Questions

Useful Databases for Pre-Appraised Evidence

*Cochrane Database of Systematic Reviews* BMJ Clinical Evidence Database of Reviews of Effects (DARE) National Guideline Clearinghouse (NGC) Physician's Information and Education Resources (PIER) American College of Physicians Journal Club (ACP)

Cochrane Databases

+ CDSR—FT systematic review + CENTRAL—citation and abstract of clinical trials Note: Only two of the five Cochrane databases are described here Search Strategies: KW, SH (i.e., MeSH if you know the heading)

*MEDLINE*

+ Journal article citation and abstract of primary studies, reviews, and synopses + FT (varies on PubMed; other vendors with FT subscription) Search Strategies: KW, KP, TI, SH, clinical queries

*CINAHL*

+ Journal article citation and abstract of primary studies, reviews, and synopses + FT (with FT subscription) Search Strategies: KW, KP, TI SH (i.e., CINAHL headings)

EMBASE

+ Journal article citation and abstract of primary studies, reviews, and synopses + conference coverage Search Strategies: KW, KP, TI, SH

PsycINFO

++ Journal article citation and abstract of primary studies, reviews, and synopses + FT (with FT subscription) Search Strategies: KW, KP, TI, SH

National Guideline Clearinghouse

+Clinical practice guidelines + syntheses of selected guidelines + FT Search strategies: KW, category

Trip

+Journal article citation and abstract of primary studies, reviews, and synopses + guidelines + linkout to FT when available Search Strategies: KW, KP, TI, PICOT search builder, proximity

Final Tips to an Efficient Search

-Begin with PICOT question to generate keywords -Use subject headings when available -If search results are sparse, expand it using the explode option (if not automatic) -Use available search engine mechanisms to focus the search so that the topic of interest is the main point of the article -Establish inclusion/exclusion criteria before searching so that the studies that answer the question are easily identifiable. Apply these criteria after search strategy is complete.

PubMed

-Database produced and maintained by the National Center for Biotechnology Information (NCBI) at the National Library of Medicine -Contains >19 million citations from >5,600 journals of biomedical sciences, nursing, dentistry, and pharmacy -Free and committed to providing access to all evidence -Provides free online access to the MEDLINE® database -Includes automatic term mapping that uses the keywords entered to map them to appropriate MeSH® terms

Licensed Databases vs. Web-based Search Engines

-Licensed databases such as PubMed® list the journals indexed, which allows users to know which journals they are searching -Internet search engines such as Google and Google Scholar search the Internet, but there is no transparency as to what information is included in the search -Internet search engines include the grey literature, which include unpublished drug trials, reports, or conference proceedings. Because there is no peer review of this evidence, it should be appraised to ensure that the information is reliable. -Combining a licensed database with an Internet search engine may yield the best search results

Using Reference Management Software Systems (RMS) in Searches

-Often referred to as citation managers -Used to save, search, sort, share, and continuously add, delete, and organize promising citations -Web-based proprietary examples include *RefWorks* and Endnote® -Open-source options include Mendeley (http://www.mendeley.com) and Zotero (http://www.zotero.org)

Combing Searches

-Placing several concepts from the PICOT question in one search allows a simultaneous search, but it cannot be determined which concept has the most available evidence -Running multiple single-word searches allows the number of "hits" to be seen for each. Then decisions can be made to possibly use Boolean operators. -Using the Boolean operator "AND" is useful when narrowing a search to combine two search results. BOTH terms need to be present or an article will not be included in the results. -Using "OR" will expand a search to include either one or both terms in the results

Using Limits

-Using the "limit" function pares down a large results list -Options for limiting the results vary by database -Limiting to RCTs or meta-analysis first can help determine the highest level of evidence that is available -Limiting the search may result in missing relevant evidence (e.g., limiting the search to "full-text only" eliminates all publications that the database does not subscribe to in full text)

Questions That Need to Be Asked About Databases Used for Searching Evidence

1. Is the evidence current? 2. Which search strategies are supported (e.g., are keywords, title searches, and subject heading searches all supported)? 3. How frequently is the database updated?

Databases Useful for Finding Individual Research Reports

MEDLINE® CINAHL PsycINFO®

"Sort by Pub Date"

Search results appear in the order in which they were added to the database. To find the most recently published article, use the "Sort by Pub Date" option.

Answering Clinical Questions

Searching for evidence that has *already been appraised for validity and reliability* decreases the amount of time needed to determine whether the information is reliable

Pre-appraised literature can include:

Systematic reviews and meta-analyses Meta-syntheses Integrative reviews Synopses/critiques of single studies

Is the following statement true or false? Subject headings searching (also known as controlled vocabulary searches) may yield fewer hits than a keyword search, but these hits are more likely to be relevant to the clinical question.

True Rationale: Controlled vocabulary systems exist to increase the relevance of search results while limiting the number of less relevant hits.

Which of the following sources of evidence would be the best evidence to use to suggest a clinical practice change? a. A well-designed randomized controlled trial (RCT) b. A systematic review that encompasses multiple studies c. Expert opinion of experienced and educated nurses d. A case study that addresses a similar clinical situation

b. A systematic review that encompasses multiple studies Rationale: Systematic reviews are pre-appraised evidence, which are considered to be higher on the hierarchy of evidence than expert opinion, individual RCTs, or case studies.

Which of the following online evidence sources is most likely to provide pre-appraised evidence? a. CINAHL b. MEDLINE c. PubMed d. The Cochrane Library

d. The Cochrane Library Rationale: The Cochrane Databases include the Cochrane Database of Systematic Reviews (CDSR), which is a collection of systematic reviews that synthesize RCTs from multiple peer-reviewed sources. CINAHL, MEDLINE, and PubMed may contain evidence at a synthesis level, but most of the sources are individual research articles that must be appraised by the user.

Subject heading searching

uses a standardized set of preselected terms for the search -Also referred to as controlled vocabulary, subject terms, thesaurus, descriptors, or taxonomies -Major strengths: Searches can be broadened without considering every synonym for the chosen keyword; studies selected only if at least 25% relevant to the topic, thus decreasing the number of irrelevant hits -Major weaknesses: Newly developed technologies, phrases, and acronyms may not yet be linked in the database and thus be missed

Automatic term mapping

uses a three-step process to search keywords entered in the search box and to map them to appropriate MeSH® terms: Step 1—MeSH® term: Looks for a match between keywords entered and a list of MeSH® terms. If a match is found, MeSH® term plus the keyword is used to run the search. Step 2—Journal title: If no MeSH® term match, keywords are compared with a list of journal titles. If a match is found, the journal title is used to run the search. Step 3—Author name: If no match is found in steps 1 and 2, words in the search box are then compared with a list of author names. If there is a match, author's name is used to run the search. If no match is found after step 3, search engine will drop the keyword farthest to the right in the search string and will repeat the three-step process If a match is found, then automatic term mapping will use the match (MeSH® term, journal title, or author name) plus the keyword as part of the search and return to process the term that was previously dropped to begin a separate search

Title Searching

uses keywords generated from the "P," "I," and "O" components of the PICOT question to search article titles with the same keywords -As with the use of keyword searches, all appropriate common terms, synonyms, acronyms, phrases, coined phrases, and brand names need to be used -Major strengths: Increases the chance of the article found being relevant to the PICOT question and is highly effective in finding relevant articles -Major weaknesses: Misses studies that do not contain the keywords in the title

Keyword Searching

uses words generated from each component of the PICOT question -All appropriate keywords, including common terms, synonyms, acronyms, phrases, coined phrases, and brand names, need to be used -Major strengths: Provides a quick snapshot of how helpful a database will be in finding relevant evidence -Major weaknesses: May miss studies that do not exactly match the authors' keyword choices; may find many studies irrelevant to the PICOT question