Skip to Main Content

Evidence-Based Medicine: Searching Techniques

This is a libguide for praticing evidence-based medicine.

Translating PICO into Search

You have broken down the clinical question into its PICO components and corresponding search terms. Next, you can combine search terms into a variety of search strategies, adapted to each resource you use. The databases for top EBM resources are highly selective and relatively small, so you can keep searches simple. One or two search terms for the population or problem and for your intervention or exposure will find most relevant resources. For example, if you are interested in the diagnosis for schizophrenia, simply searching the term “schizophrenia” in summaries (e.g. in UpToDate) will usually suffice. However, searching the large primary studies databases such as PubMed usually require more specific and structured searches.To find the evidence you need in large databases, your search terms should closely relate to the components of your PICO question.


Clinical question: In patients with suspected schizophrenia, is magnetic resonance imaging (MRI) compared with computed tomography (CT scan) more accurate in diagnosing schizophrenia?

PICO Components

Potential Search Terms

Examples of Search Strategies (PubMed)

Examples of Search Strategies (Ovid MEDLINE)


patients with suspected schizophrenia



exp Schizophrenia/ OR



magnetic resonance imaging (MRI)

magnetic resonance imaging (MRI)

Magnetic Resonance Imaging OR MRI  OR NMR

exp Magnetic Resonance Imaging/ OR

"Magnetic Resonance".ti,kw.


(MRI OR NMR).ti.


computed tomography (CT scan)

computed tomography (CT scan)

"Tomography, X-Ray Computed"[Mesh] OR “computed tomography” OR CT

exp Tomography, X-Ray Computed/


("comput* tomography" OR CT).ti,kw.


accurate diagnosis



exp Diagnosis/



Note: ti,kw means searching the terms in title and keyword fields.

Boolean operators AND or OR are usually used when combining search terms. OR finds studies containing either of the specified words/phrases and broadens your search. AND finds studies containing both specified words/phrases and narrows your search. The AND operator is useful when attempting to link different concepts together from your PICO question. Below is the Ovid MEDLINE search strategy for the example clinical question:

          #     Searches     
          1     exp Schizophrenia/     
          2     Schizophreni*.ti,kw.     
          3     1 or 2     
          4     exp Magnetic Resonance Imaging/     
          5     "Magnetic Resonance".ti,kw.     
          6     (MRI or NMR).ti.     
          7     4 or 5 or 6     
          8     exp Tomography, X-Ray Computed/     
          9     ("comput* tomography" or CT).ti,kw.     
         10     8 or 9     
         11     exp Diagnosis/
         12     diagnos*.ti,kw.     
         13     11 or 12     
         14     3 and 7 and 10 and 13


Broad vs. Narrow Searches

If you initially find little evidence, you can broaden your search (that is, increase its sensitivity) by adding synonyms for each concept or using truncated terms (e.g. schizophreni* will retrieve schizophrenia, schizophrenic, variant spellings, and other similar terms with different endings). Conversely, if your initial search retrieved too many citations to be screened, you can narrow your search (that is, increase its specificity) by linking more PICO components with “AND” or by adding limits (publication date, age group, etc.) and methodological filters (e.g., clinical trials). More sophisticated approaches include entering PICO components sequentially according to their importance to obtain a manageable number of articles in large databases, such as PubMed.

Being very specific in searches will ensure that the answer you get is applicable to the patient. However, it is a double-edged sword. We may fail to find any studies that meet all PICO elements. The solution is to start with specific PICO components but be ready to remove specifications to find a relevant article. For example, we may be ready to remove the "gender," "age group" and other limits when we think it is likely that those specifications won't make any difference in clinical outcomes.


Searching the Medical Literature Is Sometimes Fruitless

The medical literature will not be helpful when no feasible study design or measurement tools exist that investigators could use to solve a clinical issue. Your search also will not be successful if no one has conducted and published the necessary study. Before embarking on a search, carefully consider whether the yield is likely to be worth the time.

Get Help

Because of the complexity of medical databases, some searches simply require the help of an information specialist. Collaboration with librarians who are knowledgeable about evidence-based practice is essential to get efficient answers to clinical questions. With librarians, you may talk about your clinical questions, formulate PICO questions, understand evidence-based information resources, learn about structure and searching in different databases and discuss search keywords, subject headings and strategies. Ask the librarians for literature search help.