LibGuides: PBL Research Guide: Article Databases

Quick Start Guide to Searching PubMed for EBM

Things to remember about searching PubMed for EBM:

The goal of the search is to capture the body of evidence relevant to the PICO, not a single article.
A narrower search is not always better. A broader search is often more strategic for identifying all the studies on your topic.
Choose the best available evidence. High levels of evidence do not exist for all PICOs so don’t be discouraged if you don’t find any systematic reviews or RCTs!

Start here:

Choose initial search terms, removing any words that are implied. For example, keywords patients with, decreased, increased, outcome should almost never be included in a search. Search each term separately then combine with AND.

Next consider how to adjust your search:

Considering search terms for each PICO concept separately, improve your search by using one or more search techniques.

Summary of Search Techniques
Search Technique	MeSH	Combine synonyms with OR	Filter for Age, Sex	Broader Term	Eliminate PICO element
When to try it:	A keyword search for a concept has other meanings or applications. A keyword search is returning results that are unrelated to your search concept. A MeSH term matches the search concept.	Search is returning too few results. Multiple common ways to refer to a search concept No MeSH term for the search concept or the MeSH term does not exactly correlate to the search concept.	Age, sex of the patient is important for relevance of articles Age, sex not implied by other search terms Search returns too many results	Search is returning too few results A broader term is still clinically relevant	Search is returning too few results One or more of your PICO elements is implied by others The term you eliminate isn't overly important to the relevance of your search results.

Identify the best evidence in your search results:

The results of your search should be a reasonable attempt to capture the body of evidence related to the PICO so that you can compare studies and choose the best evidence.

The question domain helps you identify study designs appropriate to the question. Not sure which study types to look for? Use the domain and study type table to identify appropriate study types for each question domain.

Domain Examples with Suggested Study Types
Question Domain	Patient or Problem	Intervention	Comparison	Outcome	Suggested Study Types
Diagnosis	In middle aged men with DVT	is Homans’ sign	when compared to an ultrasound	as accurate in diagnosing DVT?	Guideline > Systematic reviews > cohort study > case-control study
Therapy	In adults over 65 with history of stroke	are SSRIs	compared to routine care	more effective at improving functional independence?	Guideline > Systematic reviews > randomized controlled trials > cohort study
Etiology/ Harm	In young women with BRCA1 gene mutation	do oral contraceptives	versus no oral contraceptives	lead to increased risk of early-onset breast cancer?	Guideline > Systematic reviews > randomized controlled trials > cohort study
Prognosis	Are adults	with uncontrolled hypertension	versus those with well-controlled hypertension	more likely to develop Alzheimer's disease?	Guideline > Systematic reviews > cohort study > case series
Economic Analysis	For healthy middle-aged adults	is a standard colonoscopy	when compared to a virtual colonoscopy	more cost effective?	Guideline > Systematic reviews > costs and cost analysis > audit or outcomes

Using article type filters in PubMed can help you systematically comb through your search results so you’re checking for appropriate study designs first.

Tips for appraising articles and applying them to a PBL case:

To what extent do the subjects of the study match the patient?
To what extent does the context of the study correspond to the clinical scenario in the PBL case?
What are strengths and limitations of the study's methods respective to its study design?

To quickly get information needed to assess answers to those questions:

Tips for Quickly Screening An Article
Needed Information	Where to Find It
1. Identify the publication date.	The article citation with the publication date is located at the top of the record in PubMed.
2. Determine if the paper is original research. I.e., a clinical study or a systematic review	The type of paper may be identified in the PubMed record or in the title. A structured abstract usually indicates original research.
3. Determine the aims of the study.	The last sentence of before the methods section in the full text typically states the research question.
4. Determine the study design.	The methods section describes the design.
5. Consider whether the study participants are relevant to the patient.	The beginning of the methods section typically describes how participants are recruited, with a table breaking down demographic characteristics.

Some PBL PICOs are tricky. Feel free to email your group's librarian if you get stuck!

Article Databases

Cochrane Library
A collection of Cochrane databases including systematic reviews, reviews of effectiveness, register of controlled trials in addition to methodology, technology and economic evaluations.
PubMed
A database of journal citations and abstracts in the fields of medicine, nursing, dentistry, veterinary medicine, healthcare systems, and pre-clinical sciences. PubMed focuses on North American peer-reviewed literature. Full-text coverage of journals licensed or owned by Tufts are available via full-text links.
Web of Science
Scholarly literature in the sciences, social sciences, arts, and humanities.