Transcript 1Specialty_resources_for_the_Clinician_07

Medical Education Development
Series: Specialty Resources and Tools
for the Clinician
Sue McGuinness, PhD, MLS
Karen Heskett, MSI
Alice Witkowski, MLS
February 8, 2007
Learning Outcomes
• Understand basic principles of EvidenceBased Medicine (EBM)
• Be aware of EBM and Drug information
Tools
• Practice Using EBM and Drug Information
Tools
Evidence-Based Medicine
• What Is It?
• Clinical Expertise + conscientious,
judicious use of the best available
evidence in making decisions about
patient care
Expertise
Evidence
Patient values
Evidence-Based Medicine
• Why practice it?
–
–
–
–
Dispel “myths”
Deal with information overload
Aids decision making
Improves patient care
• Why teach it?
– AAMC learning objectives
– Improves critical thinking skills
– Can apply to clerkships, residencies, rounds, PBL
courses, reading groups, journal clubs,
Barriers to practicing EBM
• Time
• Information Resources
• Cost
Analysis of questions asked by family
doctors regarding patient care
Ely JW, et al. BMJ 1999;319:358-361
Physicians’ Information Needs
Ely JW, Osheroff JA, Chambliss ML, Ebell MH, Rosenbaum ME
JAMIA 12 (2): 217-224 MAR-APR 2005
Total Questions
N=1,062
Pursued
N= 585 (55%)
Answered, no difficulty
N=238 (41%)
Answered with difficulty
N=180 (31%)
Not Pursued
N= 477 (45%)
Not Answered
N=167 (28%)
Why? Limited time, Suspect lack of information,
Need better resources.
How to practice EBM (The 5 A’s)
Assess- patient history, labs, meds, patient values
* Ask- build a focused searchable question
* Acquire- the best evidence you can find
Appraise- evaluate the evidence
Apply- to patient care
Categories of Clinical Questions
• Diagnosis – Test A or B to differentiate
between positive an negative
• Therapy – treatment A or B to improve a
condition or avoid an adverse event
• Prognosis – future course of condition or
disease
• Etiology/harm- effects of agents on
function, morbidity, mortality
Hierarchy of Evidence
Systematic Reviews &
Meta-analyses of RCTs
Randomized Controlled
Trials (RCTs)
Retrospective (case-control)
Systematic Reviews of
Observational Studies
Observational Studies
Prospective (cohort)
Case Series
Case Reports
Best Evidence for Clinical
Questions
Diagnosis and Therapy- randomized
controlled trial
outcome
– Eligible patients
randomization
outcome
Prognosis and Etiology/Harmobservational studies
Eligible patients
outcome
Circumstances
No outcome
Ask: Formulating the wellbuilt clinical question
Pauline is a new patient who recently moved to the area to be closer
to her son and his family. She is 67 years old and has a history of
congestive heart failure brought on by several myocardial infarctions.
She has been hospitalized twice within the last 6 months for
worsening of heart failure. At the present time she remains in normal
sinus rhythm. She is extremely diligent about taking her medications
(enalapril, aspirin and simvastatin) and wants desperately to stay out
of the hospital. You think she should also be taking digoxin but you
are not certain if this will help keep her out of the hospital. You
decide to research this question before her next visit.
Introduction to Evidence-Based Medicine [online tutorial] Duke University Medical Center Library, UNC-\
Chapel Hill health Sciences Library, 2004. http://www.hsl.unc.edu/services/tutorials/EBM/index.htm.
Accessed 2/6/07
Ask: Build a focused query
(PICO)
•P: terms describing the patient/problem
•I: terms describing the intervention
•C: list any alternative interventions
•O: terms defining the outcome of interest
Do not need PICO for pre-filtered systematic reviews
or meta-analyses
ASK: Focused query
• Patient / Problem
• Intervention
• Comparison
• Outcome
congestive heart failure,
elderly
digoxin
none, placebo
primary: reduce need for
hospitalization
secondary: reduce mortality
The well-built clinical question
• In elderly patients with congestive
heart failure, is digoxin effective in
reducing the need for hospitalization?
• It is a therapy question:
– The best evidence would be a a systematic
review or meta-analysis. If we couldn’t
find one then look for randomized
controlled trial (RCT).
Acquire:
Know the resources (ask a librarian!)
Find the best available evidence
Starting with PubMed
Acquire: Finding the best
evidence in MedLine/PubMed:
Therapy Questions
• PICO using MeSH terms where
available
• Limit to meta-analyses OR randomized
controlled trials OR clinical trials
Therapy Question: Acquire
The effect of digoxin on mortality and
morbidity in patients with heart failure. The
Digitalis Investigation Group. New England
Journal of Medicine February 20, 1997;
336(8):525-533.
Therapy Question: Appraise
Are the results valid?
What do the results mean? (are they important?)
Appraise: validity questions for
therapy articles
Were patients randomized?
Was randomization concealed?
Were patients in experimental and control groups similar in
terms of prognostic factors?
Was follow up complete?
Answers: Article: This study methodology appears to be
sound and the results are valid.
Appraise: Part 2- are the results
important?
Outcome
Placebo
Hospitalization 35%
due to CHF
Digoxin
27%
Appraise: Are the results important?
Ways to quantify magnitude of treatment
effect
Relative Risk: Risk of the outcome in experimental group
compared to the risk in the control group. .
RR= outcome (exp) / outcome (con)
= 27/35= .77 = 77%
Relative Risk Reduction: The percent reduction of
risk in the experimental group compared to the control group)
RRR = outcome (con) – outcome (exp) / outcome (con)
= 23%
Appraise…Are the results
important
Absolute Risk Reduction: The difference between
outcome rates in experimental and control groups
ARR= outcome (con) – outcome (exp)
Example: 35% - 27% = 8%
100%
65% avoid the hospital either
way
8% avoid the hospital by taking
digoxin
27% are hospitalized either way
Appraise: Ways to quantify
effect
Number needed to treat (NNT): The number of
patients that need to be treated to observe outcome
in one patient.
Example ARR = 8%
Outcome (death) was avoided 8 times per 100
patients.
How many patients need to be treated in order to
avoid one death? 8/100 = 1/NNT
NNT = 100% / ARR % = 100/8 = 13 patients need
to be treated in order to avoid one hospitalization by
using digoxin
RRR vs. ARR: Some papers
report RRR only. Why is that a
problem?
Calculate RRR, ARR and NNT for the
following three RCTs
% Mortality
RCT 1
RCT 2
RCT 3
Treatment
10
40
60
Control
15
60
90
RRR vs. ARR: RRR alone tells
you nothing
RCT 1: RRR = .33
ARR = .05
NNT = 20
RCT 2: RRR = .33
ARR = .20
NNT = 5
RCT 3: RRR = .33
ARR = .30
NNT = 3
APPLY: Should we prescribe
Digoxin for Pauline?
Diagnosis Questions
• Ask (PICO)
• Acquire (Systematic reviews,
metaanalyses, RCTs)
• Appraise (Validity and Importance)
– Validity questions differ for diagnosis
– Data are different (sensitivity, specificity,
likelihood ratios)
Evidence-Based Medicine: How to practice and Teach EBM, Sackett DL, et al. NY: Churchill, 2000.
Diagnosis: Validity questions
• Key issues for Diagnostic Studies:
– blinding
– identified gold standard test
– patient sample
– each patient gets both tests
Diagnosis: Are the results important
• Sensitivity measures the proportion of
patients with the disease who also test
positive for the disease.
• Specificity measures the proportion of
patients without the disease who also test
negative for the disease.
Search Tip for Diagnosis
• Use the MeSH term “Sensitivity and
Specificity”
• If that doesn’t work, use ONE of the terms
in your search
Need a Break?
The Good News: We have PreFiltered Sources
• Cochrane Library/DARE
• ACP Journal Club
• NG Clearinghouse
• Trip Database (FREE!)
Archie Cochrane 1909-1988
Demonstration: EBM resources
• Search “congestive heart failure” in
Cochrane Library
• Search ACP journal Club
• Search TRIP database
Drug Information Questions
• Now that we’re giving Digoxin to Pauline,
we need information on adverse effects,
interactions with other drugs or foods,
dosing information, and information for
Pauline to read about her new medication.
Drug Information Resources
• Clinical Pharmacology
• Micromedex
• Exercises: work in groups, pick a
question
– Fred: Therapy
– Amelia: Therapy
– Myrtle: Diagnosis
– Any questions you have had in your
practice
• ASK
• ACQUIRE
• Appraise, if you have time. Validity
questions are in your packet