Multiple choice questions in EBP
Samir Haffar M.D.
Which one of the following studies has the highest
risk of bias?
A- Case report/series
B- Cross-sectional study
C- Case- control study
D- Cohort study
E- RCT
Evidence pyramid
Increase in Decrease in
evidence level bias risk
McGovern D, Summerskill W, Valori R, Levi M. Key topics in EBM.
BIOS Scientific Publishers, 1st Edition, Oxford, 2001.
What is the best design you choose to study the
prevalence of a disease?
A- Ecologic study
B- Cross sectional study
C- Case- control study
D- Cohort study
E- RCT
Cross-sectional study design
prevalence study
At one point of time
eg.: prevalence of coronary heart disease in smokers
What is the best trial design to study the incidence of
a disease?
A- Ecologic study
B- Cross-sectional study
C- Case-control study
D- Cohort study
E- RCT
Cohort study
investigate etiology or outcome of disease
Prospectively over a period of time (years or decades)
Can be retrospective if clear point of 1st exposure
2 groups well matched to avoid confounding factors
Which of the following studies is considered a gold
standard for analytical epidemiology?
A- Ecologic study
B- Cross-sectional study
C- Case-control study
D- Cohort study
E- RCT
The cohort study is the gold-standard of
analytical epidemiology
Types of clinical studies
Observational Interventional
Descriptive Analytic
Case report/series Cross-sectional study RCT
Ecological study Case-control study
Cohort study
RCT: randomized controlled trial
Time is key
Cross-sectional
Exposure Case-control Outcome
Cohort
You want to assess the efficacy of a new anti-epileptic
drug versus an old drug? What is the best design you
choose for this purpose?
A- Ecologic study
B- Cross sectional study
C- Case-control study
D- Cohort study
E- RCT
Basics of RCT – 3
RCTs are regarded as
• Quantitative studies (quantified outcomes)
• Most rigorous method of hypothesis testing
• Experimental studies versus observational studies
• Gold standard to evaluate effectiveness of interventions
Jadad AR, Enkin MW. Randomized control trials.
Blackwell Publishing, 2nd ed, 2007.
Basic structure of a RCT
Parallel trial is the most frequently used design
Akobeng AK. Arch Dis Child 2005 ; 90 : 840 – 844.
Question type & study design
Question Study Design
Prevalence Cross-sectional study
Diagnosis Cross-sectional study
Etiology & risk factors Cohort or case-control
Incidence & prognosis Cohort study
Intervention RCT
In each case, SR of all available studies better than individual study
You read in a paper that a p value is 0.01. Is this
result clinically significant?
A- Yes
B- No
C- Cannot tell
Probability value (p value)
• p > 0.05 Statistically insignificant
• p < 0.05 Statistically significant
statistically doesn't clinically
significant mean significant
An open label randomized controlled trial means:
A- Everyone participating in the trial is aware of
assigned treatment
B- Patients are ignorant of assigned treatment
C- Investigators are ignorant of assigned treatment
D- Patients, investigators and data evaluators are
ignorant of assigned treatment
Blinding or masking
Depending on blinding extent, RCTs classified as
• Open label Everyone aware
• Single-blind Only patients or investigators ignorant
• Double-blind Patients & investigators ignorant
• Triple-blind Patients, investigators & data evaluators
ignorant
A critical appraisal of a RCT takes into consideration
one of the followings:
A- Randomization
B- Blinding
C- Precision of the estimate (CI)
D- Benefice versus harm
E- All of the above
Internal & external validity of a RCT
Attia J & Page J. Evid Based Med 2001 ; 6 : 68 - 69.
Appraising a RCT (checklist) – 1
Are the results valid?
At start of trial Were the patients randomized?
Was the randomization concealed?
Similar prognostic factors in 2 groups?
During trial Was trial blinded & to what extent?
At end of trial Was follow-up complete?
Was ITT principle applied?
Was the trial stopped early?
Guyatt G, et al. User’s guide to the medical literature.
Essentials of evidence based clinical practice. Mc Graw Hill, 2nd ed, 2008.
Appraising a RCT (checklist) – 2
What are the results?
8- How large was the treatment effect?
9- How precise was estimate of treatment effect (CI)?
How can I apply the results to patient care?
10- Were the study patients similar to my patient?
11- Were all patient-important outcomes considered?
12- Are the likely treatment benefits worth harm & cost?
Guyatt G, et al. User’s guide to the medical literature.
Essentials of evidence based clinical practice. Mc Graw Hill, 2nd ed, 2008.
External validity
Applicability of results to your patients
Issues needed to consider before deciding to
incorporate research evidence into clinical practice
• Similarity of study population to your population
• Benefit versus harm
• Patients preferences
• Availability
• Costs
* Guyatt G, et al. User’s guide to the medical literature.
Essentials of evidence based clinical practice. Mc Graw Hill, 2nd edition, 2008.
The problem of applying trial results
The Trial
patients
The trial
report
The actual
patients
Benefit versus harm
“All that glisters is not gold”
W. Shakespeare
In “The Merchant of Venice”
Furberg BD & Furberg CD. Evaluating clinical research.
Springer Science & Business Media – 1st Edition – New York – 2007.
The receiver operating characteristic is used to report:
A- Incidence of a disease
B- Prevalence of a disease
C- Prognosis of a disease
D- Diagnostic test with 2 results (yes/no)
E- Diagnostic test with more than 2 results
Accuracy of tests & number of results
• Dichotomous test (only 2 results)
Sensibility & Specificity
PPV & NPV
with 95% CI
Likelihod ratio + & –
Diagnostic OR
• Multilevel test (> 2 results)
Receiver Operating Characteristic (ROC)
Make continuous test dichotomous: fixed cut-off value
Which of the followings is used to know the cut-off
values of a diagnostic accuracy test (disease positive
versus disease negative):
A- Positive predictive value
B- Negative predictive value
C- Likelihood ratio
D- Receiver operating characteristic
Useful properties of ROC curve
AUC provides an overall measure of a test’s accuracy
Accuracy of binary diagnostic test for a cut-point value
Determination of cut-off point to distinguish D + & D –
Comparison of different tests for dg of a target disorder
Area under the ROC curve in IDA
If we select 2 patients at random one with IDA & one without
Probability is 0.91 that patient with IDA will have abnormal ferritin
Accuracy of diagnostic test using AUC of ROC
Value Accuracy
0.90 - 1.00 Excellent
0.80 - 0.90 Good
0.70 - 0.80 Fair
0.60 - 0.70 Poor
The higher AUC the better the overall performance of the test
Pines JM & Everett WW. Evidence-Based emergency care: diagnostic testing & clinical decision
rules. Blackwell’s publishing – West Sussex – UK – 2008.
Accuracy of binary dg test for a cut-point value
Determination of cut-off point
to distinguish D + & D –
Cut-off point discriminates between
subjects with or without disease
Indicated by the point on curve that
is far away from chance diagonal
Peat JK. Health science research. Allen & Unwin, Australia, 1st edition, 2001.
Comparing different tests for target disorder
Diagnosis of IDA AUC of the ROC
Seurm ferritin 0.91
Transferrin saturation 0.79
MCV 0.78
RCP 0.72
* RCP: Red Cell Protoporphyrin
Guyatt GH et al. J Gen Intern Med 1992 ; 7 : 145 – 153.
Thank You
