OLD_Allied Health Resources: Search Strategies for Databases

Several standard techniques are useful when searching the literature. The average keyword search will only give you a fraction of relevant citations available. As you develop more advanced searching skills this percentage of relevant citation retrievals will increase. In his article on “Searching for information on outcomes: do you need to be comprehensive?” (1) Brettle points out that the goal is to balance precision with recall. Recall pertains to the number of publications of potential interest retrieved in the search. Precision pertains to the relevance of the publications identified in the search.

Typically, as recall increases, the number of irrelevant citations increases as well, thus lowering precision. And as the precision increases, limiting a search to only the most relevant citations, the recall decreases, retrieving fewer items that are of potential interest. A balanced search seeks to retrieve a good working set of relevant articles while limiting the number of irrelevant articles. This module identifies a few commonly used techniques for developing a variety of effective search strategies.

Next Learn More About: 

• Boolean Operators
• Keyword Searching
• Controlled Vocabulary
• Attributes of a Clinical Question
• Working with Limits

Keyword searching can be defined as using words that you think authors would have used when writing about your topic. For some databases and point of care tools keyword searching is the only search option. This section describes ways to enhance keyword searching for better search results.

Truncation: Allows you to find all terms that begin with a given text string. When using truncation in PubMed, the asterisk (*) sign is the truncation representation. Example: diagnos*
Words retrieved might be diagnosis, diagnose, diagnostic, diagnostics

For Ovid users use either the question mark (?) to replace one letter or the asterisk (*) sign for one or more letters. Example: an?mia will retrieve the American or British spellings: anemia or anaemia 

Synonyms: Using words that authors might have used discussing a topic. 
Example: computer assisted instruction OR computer tutorials
Use OR between each term so that either word choice will be present in the selected article(s).

Phrase: Searching with quotation marks. Usually automatic term mapping will be turned off so your search retrieval will be the exact wording entered into the search box. 
Example: “kidney allograft”

Many databases have a set of index terms or controlled vocabulary to define conditions. MEDLINE in PubMed or Ovid uses Medical Subject Headings (MeSH). The Cumulative Index of Nursing and Allied Health Literature uses CINAHL Headings. For example if you type in heart attack both CINAHL and MeSH would return the medical term myocardial infarction. 

These terms describe the content found in the bibliographic record. They are arranged in a hierarchical structure with very broad terms at the top working down to more specific headings or narrower terms at the bottom of the hierarchical list. Searching using controlled vocabulary terms yield more relevant results.

Medical Subject Headings from Ovid are shown here. Myocardial Infarction is listed as a narrower term for the Subject Heading Heart Diseases. Narrower terms are listed below and indented to the right of the broader Subject Heading. Anterior Wall Myocardial Infaction is a more specific heading then Myocardial Infarction.

In evidence-based practice, the PICO Model has been developed as a technique for formulating a patient-specific clinical question. Before initiating this model begin by asking yourself to formulate the clinical question by defining the type of problem. Oxman has identified four “clinical questions”(2).

• Therapy
  • Efficacy and safety of treatment
  • Is the treatment effective and not harmful?
  • Articles about therapy are typically randomized, doubled-blinded studies as they would provide the strongest level of evidence
• Diagnosis/Screening
  • Precision and accuracy of diagnostic tests
  • Preventing false positive and not missing existing conditions
  • Articles about diagnosis are typically controlled trials
• Prognosis
  • Power of predictive markers
  • Determine who has increased risk for condition
  • Long Term Outcome for patient
  • Articles about prognosis are typically cohort or case-control studies. They identify the presence of and determine the increased risk associated with a factor of contention.
• Etiology (Harm/Benefits)
  • Assess treatment benefit to patient
  • Risks of treatment
  • Articles about etiology are typically cohort or case-control studies meaning exposed and unexposed groups of patients. 

A full description of the PICO Model is discussed in the Evidence Based Practice section.

Searching with limits can be a powerful method of eliminating less useful results. Each database has a specific set of options to narrow the search set. Common limits are date of article publication, language, age/gender of patient population, type of article (clinical trial, meta-analysis, etc.) and humans.

Ovid MEDLINE Ovid lists common limits within the initial searching page. “Edit Limits” allow users to customize these common limits if they have an Ovid account. Use "Additional Limits" to see a complete list of available limits

PubMed MEDLINE PubMed lists all available limits on the left hand column after a search is completed.  More options and more filters can be displayed.  Choose the More... at the bottom of a filter to add additional filter options.  Chose the 'Show additional filters' to view more filter options such as, language, age, etc.

EBSCO CINAHL Similar to Ovid, you may choose initial limits before running a search or return to the limits after searching, by choosing the "Edit" feature by each search.  CINAHL also features limits to the left hand side of your results display.

After searching, you will find a small set of limits to the right of the search results. To see the full set of limits use the "Edit" feature found to the right of your search terms in the "Search History"

Web of Science You may Refine Results from the results page, search within the results, refine to specific subject areas, document types, authors, source titles, publication years, and languages. Analyze Results extracts data from a selected field based on subject searched.

Office Hours:

By appointment. To schedule a meeting, please use the "Book an Appointment" button above to view my calendar and select the day and time that is most convenient for you.

Email: lsmith24@augusta.edu