why bother with evidence-based practice?
DESCRIPTION
An introduction to evidence-based medicine (EBM) with short section in history and why EBM? Then a brief overview of the 4 steps of EBM. These slides have been used for starting a 1-day workshops in EBMTRANSCRIPT
Why bother with Evidence-Based Practice?
Professor Paul GlasziouBond Universitywww.crebp.net.au/
www.testingtreatments.orgFree pdf: Ch 5, 6, 7
Introductory Lecture: Objectives
What is Evidence-Based Medicine?
The steps of doing EBM1. Formulate Clinical Questions2. Search for Evidence3. Appraisal of research4. Apply to clinical problem
What is Evidence-Based Medicine?
“Evidence-based medicine is the integration of best research evidence with clinical expertise and patient values”
- Dave Sackett
Patient Concerns
Clinical Expertise
Best ResearchSackett DL, Rosenberg WMC, Gray JAM, Haynes RB, Richardson WS: Evidence based medicine: what it is and what it isn’t. BMJ 1996;312:71-2.
“EBM” - birth of a term
Update of CMAJ series on how to read a paper
JAMA User guides 1991authors seek a new term
Clinical epidemiology?
Scientific medicine?
Evidence-based medicine!
The need for EvidenceVertebroplasty for osteoporotic fractures
Gray et al Spine 2008. Nation-wide and State-specific primary vertebroplasty rates per 100,000 Part B fee-for-service
Steady rise in use of vertebroplastyfor a decade
2009: two large trials publishshowing no effect over placebo
Slides courtesy R Buchbinder
Unnecessary Procedures
http://www.cbsnews.com/8301-18563_162-5217954.html
Please click on the black screen to view the video
Some popular treatment innovations that did not work
Vertebroplasty for osteoporotic fractures (NEJM, 2010)
Arthoscopic knee lavage (Moseley, NEJM, 2002) Blood glucose monitoring for non-insulin
dependent diabetes (DiGEM trial, BMJ 2007) Tight control of diabetes (ACCORD, NEJM, 2010) Prostate cancer screening (Djulbegovic, BMJ
2010) Ovarian cancer screening (JAMA, 2011)
Most medical innovations don’t workan analysis of 136 trials in myeloma
New TreatmentBetter
New TreatmentWorse
How can we find the research that will improve the care of our patients?
0
500000
1000000
1500000
2000000
2500000
Biomedical MEDLINE Trials Diagnostic?
Med
ical
Art
icle
s p
er Y
ear
5,000?per day
1,500 per day
95 per day
Medic
al A
rtic
les
Per
Year
19 of 20
What should you believe?
Keeping up to dateWhat is your JASPA* score?* (Journal Associated Score of Personal Angst)
J: Are you ambivalent about renewing your JOURNAL subscriptions?A: Do you feel ANGER towards prolific authors?S: Do you ever use journals to help you SLEEP?P: Are you surrounded by PILES of PERIODICALS?A: Do you feel ANXIOUS when journals arrive?
*Modified from: BMJ 1995;311:1666-1668
0 (?liar) 1-3 (normal range) >3 (sick; at risk for polythenia gravis and related conditions)
Coping with the overload: three possible things you might try
A. Read an evidence-based abstraction journal (and cancel other journals)
B. Keep a logbook of your own clinical questions
C. Run a case-discussion journal club with your practice
Part 2: The 4 steps of “pull” EBM
1. Formulate an answerable question2. Track down the best evidence 3. Critically appraise the evidence4. Individualise, based clinical expertise and patient concerns
Step 1Formulate an answerable clinical questionStructure of researchable
questions – PICO-T
Population/Patients
Intervention
Comparison
Outcome
Time
Formulating answerable clinical questionsStructure of researchable
questions – PICO-T
Population/Patients
Outcome
What are your clinical questions?
A 35 year old man says his brother
recently died of a ruptured
cerebral aneurysm.
He is worried about whether he
might have one and what the
chances are that it would rupture.
-> PICO Table
Risk Factors
Cause(s)
SymptomsSigns, Tests
Prognosis
Treatment Effect
Past Current Future
Types of question: stroke
Frequency
Cohort Study SurveyInception Cohort Study
Treatments
Randomised Trial
CT Scan
Cross Sectional Study
What are the … outcomes (PO?)
Outcomes ?
Patients
Qualitative Research
The “best” evidence depends on the type of questionLevel Treatment Prognosis Diagnosis
I
II Randomised trial
Inception Cohort
Cross sectional
III
Level Treatment Prognosis Diagnosis
I Systematic Review of …
Systematic Review of …
Systematic Review of …
II Randomised trial
Inception Cohort
Cross sectional
III
The “best” evidence depends on the type of question
What do you do about conflictingclaims?
2. Searching: finding good answers?
Impact of searching on correctness of answers to clinical questions
Right to Right
Wrong to Right
Right to Wrong
Wrong to Wrong
McKibbon(GP or IM)
28% 13% 11% 48%
Impact of searching on correctness of answers to clinical questions
Right to Right
Wrong to Right
Right to Wrong
Wrong to Wrong
McKibbon(GP or IM)
28% 13% 11% 48%
Quick Clinical(GPs)
21% 32% 7% 40%
Hersh(Med students)
20% 31% 12% 36%
Hersh(Nursing)
18% 17% 14% 52%
3. Rapid Critical Appraisal
It’s peer-reviewed, therefore it must be OK?
Is the PICO a POEM?
Patient
Oriented
Evidence that
Matters
Critical Appraisal Steps
Did you find good quality studies?
Two steps1.What is the PICO (Question)2.Is the potential bias low?• “RAMMbo” (Valid Study?)• “FAITH” (Valid Review?)
Use the RAMMbo to check validity
Was the Study valid?1. Representativeness
Who did the subjects represent?2. Allocation
Was the assignment to treatments randomised?
3. Maintainence Were the groups treated equally?
4. Measurements blinded OR objective
Were patients and clinicians “blinded” to treatment? OR
Were measurements objective & standardised?
Modified from: User Guide. JAMA, 1993
Fundamental Equation of Error
Measure = Truth + Bias + Random Error
Use good study
design
Use large
numbers
Researcher
Critically AppraiseDesign
ConfidenceIntervals
andP-values
Reader
Two methods of assessing the role of chance
P-values (Hypothesis Testing) use statistical test to examine the ‘null’
hypothesis associated with “p values” - if p<0.05 then
result is statistically significant Confidence Intervals (Estimation)
estimates the range of values that is likely to include the true value
Relationship between p-values and confidence intervals - if the value corresponding to ‘no effect’ (RR of 1 or treatment difference of 0) falls outside the CI then the result is statistically significant
Step 4: Applying to the individual
What do the results mean on average?
What do they mean for this individual?
Applying research requires both “Whether to” and “How to”
“Whether to” Evidence quality Individual applicability
“How to” What & where? How long & how often?
BMJ 2003; 327 : 135