EvidenceBased Medical Information

1
Evidence-Based Medical Information

• Lyn Dennison, M.L.
• Kathy Davies, M.L.I.S.
• Peter Shipman, M.L.I.S.
• Robert B. Greenblatt MD Library

2
Objectives

• Know how to find Greenblatt Library resources
• Recognize what makes good evidence and
  recognize the levels of evidence
• Recognize components and types of clinical
  questions
• Create a search strategy that yields the best
  evidence
• Identify and use relevant and valid resources
  with a minimum of work (work smarter, not harder)
  
• graded library component

3
Greenblatt Library Resources

• Books first floor
• Print journals second floor
• Currentlast few issues
• alpha order near back windows
• Bound1980-
• alpha order on left side of main aisle
• Boundpre-1980)
• alpha order on right side of main aisle
• Computer Lab first floor
• Group study rooms first and second floor

4
Greenblatt Library Resources

• Electronic journals
• About 75 of our journals are electronic
• Some are both print and electronic
• Databases
• Provide citations or links to journal articles
• Provide summary information from research
• www.mcg.edu/library/eresources/
• or
• click Library Resources from MCG Home Page

5
Electronic Resources x

• GIL_at_MCG
• online catalog
• what we own and where it is (print)
• GALILEO
• ePublications
• links and passwords to electronic journals
• some overlap with online catalog, but more
  comprehensive
• Subscribed Resources
• databases and indexes to find articles by subject
  (Ovid, Web of Knowledge)
• links to journal articles (ScienceDirect, Wiley)
• filtered information (InfoPOEMS)

6
Evidence Based Medicine (EBM)

• EBM encourages health professionals to seek out,
  appraise, and apply the best evidence from the
  medical literature to improve the quality of
  clinical judgments
  (McMaster)

Question Seek evidence Appraise
evidence Implement into practice Evaluate results
7
Five steps of EBM

• convert information need into answerable
  questions.
• track down the best evidence to answer the
  question (with maximum efficiency).
• critically appraise the evidence for its validity
  and usefulness.
• integrate appraisal results with clinical
  expertise and patient values.
• evaluate outcomes.

8
How do we recognize good evidence?

• Control group or comparison group
• Randomization
• Follow-up (gt80)
• Blinding
• Population (size, traits)
• Length of study
• Well defined outcome criteria
• Applicable to patients

9
Levels of Evidence (Oxford)

• Quality is labeled 1 5 (high low)

1 5
Outcomes Anecdotes
Medical Literature
10
Levels of Evidence (LOE)Level 4-5 (modified
from School of Health and Related Research.
University of Sheffield. UK)
11
Levels of Evidence (LOE)Level 2-3 (modified
from School of Health and Related Research.
University of Sheffield. UK)
12
Levels of Evidence (LOE)Level 1 (modified from
School of Health and Related Research. University
of Sheffield. UK)
13
Components of a Clinical Question

• Should be
• Focused
• Answerable
• Relevant
• Should have PICO components
• Patient or Problem
• Intervention
• Comparative intervention
• Outcome

14
Ask a Clinical Question

• In the elderly with osteoarthritis, how effective
  are NSAIDs versus Tylenol in relieving pain?
• Patient elderly with osteoarthritis
• Intervention NSAIDs
• Comparison Tylenol
• Outcome reduced pain

15
Ask a Clinical Question

• In patients who have had a stroke, are botulinum
  toxin A injections more effective than placebo in
  improving wrist and finger spasticity?
• P stroke patients
• I botulinum toxin A injections
• C placebo
• O spasticity improvement

16
Types of Clinical Questions

• Big Four
• Therapy
• Does it do more good than harm?
• Diagnosis
• How well does it confirm or exclude a diagnosis?
• Prognosis
• How well does it predict clinical course over
  time?
• Etiology
• How well does it identify a cause for a disease?

17
Types of Clinical Questions

• Therapy
• In coronary patients with heart failure, should
  diuretics alone or diuretics with ACE inhibitors
  be used?
• Diagnosis
• In patients with heart failure, what bedside test
  is effective for determining left ventricular
  function?
• Prognosis
• In an elderly woman with coronary disease, how
  much does the appearance of heart failure shorten
  her life?
• Etiology
• In an elderly woman with coronary heart disease,
  can ibuprofen precipitate heart failure?

18
Searching Journal Literature

• Ovidcommercial vendor that provides multiple
  databases with links to each other and to
  electronic articles
• MEDLINE
• CINAHL (nursing and allied health)
• EBM databases
• Others (PsycINFO, HAPI, SPORTDiscus)
• PubMedfree and provided by the National Library
  of Medicine
• MEDLINE
• Others, particularly genetic databases
• will be required for library assignment

19
Search the medical literature

• Using Ovid MEDLINE
• Create a clinical question
• Identify each concept in the question
• Enter 2-3 of the most important concepts
  separately
• Use auto-mapping to match concepts to official
  subject headings (MeSH)
• Use auto-exploding to include the specific with
  the general
• Use subheadings to narrow topic (use sparingly!)
• Combine concepts using AND, OR
• Limit results (language, age, gender)
• Apply EBM filters for type of question
  (therapy, diagnosis, prognosis, etiology)
• graded library component

20
Identify major concepts

• In the elderly with osteoarthritis, how effective
  are NSAIDS versus Tylenol in relieving pain?
• Major (bold) concepts should include the disease
  and intervention
• Other concepts (italics) such as age, gender, and
  outcomes may be used later to narrow, if necessary

21
Enter and Combine

• Enter most important concepts separately
• 1/ Osteoarthritis
• 2/ NSAIDS
• 3/ Tylenol
• Combine concepts using AND, OR
• AND creates a smaller set with all concepts
  included
• 1 and 2 and 3
• OR creates a larger set with either concept
  included
• Tylenol or acetaminophen
• Limit results (language, age, gender)
• limit 4 to aged

22
Filter for Best Evidence

• Ovid Filters for Big Four plus others
• Ovid Limit - Clinical Queries
• Sensitivity broad strategy that includes
  articles that may be only peripherally related or
  even unrelated
• Specificity specific strategy that includes
  articles that are very relevant to the topic, but
  could exclude other relevant articles
• Optimized balance between the two

23
Ovid MEDLINE search x

• In the elderly with osteoarthritis, how effective
  are NSAIDs versus Tylenol in relieving pain?
• Is gingko useful for improving memory in normal
  adults?
• Are patients with iron deficiencies more likely
  to develop gastrointestinal cancers?

Source Ovid MEDLINE (1996-April 2005)
24
Two Approaches to Finding Best Evidence

• British/Canadian approach
• Focus on methodology, comprehensiveness,
  exhaustiveness, statistics
• Just-in-time information for patient problems
• Cycle of searching, appraising, applying, and
  evaluating
• Teaches how to read the medical literature

• American approach
• Information Mastery
• Focus on efficiency and workflow
• Just-in-case information to keep up to date (as
  well as just-in-time)
• Wants the bottom line of the evidence already
  filtered
• Teaches what and when to read the medical
  literature

25
Resources

• Evidence Based Reviews
• Meta-analysis
• Cochrane Database of Systematic Reviews
• Database of Abstracts of Reviews of Effectiveness
• Controlled Trials
• randomized controlled clinical trials
• Cochrane Central Register of Controlled Trials

• Information Mastery
• Article Reviews
• ACP Journal Club
• Evidence Based Medicine journals
• also called Best Evidence in Ovid
• Filtered evidence
• InfoPOEMs

26
More resources

• Greenblatt Library web links
• Clinical (practice) guidelines
• Evidence-based medicine
• PDAs
• Search engines