7. Disease + Disease - Total
Exposed (E+) a b a+b
Unexposed (E-) c d c+d
Total a+c b+d a+b+c+d
Risk of disease among exposed =
a
a+b
Risk of disease among unexposed =
c
c+d
Relative risk or risk ratio (RR) =
a
a+b
c
c+d
RR = 1: No effect
RR > 1: Harmful effect of exposure
RR < 1: Protective effect of exposure
9. Disease + Disease - Total
Exposed (E+) a b a+b
Unexposed (E-) c d c+d
Total a+c b+d a+b+c+d
Odds of exposure among diseased (cases) =
a
c
Odds of exposure among non-diseased (controls) =
b
d
Odds ratio (OR) =
a
c
b
d
=
ad
bc
OR = 1: No effect
OR > 1: Harmful effect of exposure
OR < 1: Protective effect of exposure
Risk of disease among exposed & risk of disease among non-exposed not meaningful
because cases (Disease +) & controls (Disease -) are selected separately
11. Disease + Disease - Total
Exposed (E+) a b a+b
Unexposed (E-) c d c+d
Total a+c b+d a+b+c+d
Odds of exposure among diseased =
a
c
Odds of exposure among non-diseased =
b
d
Odds ratio (OR) =
a
c
b
d
=
ad
bc
OR = 1: No effect
OR > 1: Harmful effect of exposure
OR < 1: Protective effect of exposure
20. Case-Control Studies
Nested Case-Control Studies
Matched Case-Control Studies
Odds Ratio (OR)
Number Needed to Harm (NNH)
Patient’s Expected Event Rate (PEER)
21. (in order of decreasing likelihood of being valid)
Systematic reviews are ideal because individual RCTs are
seldom large enough to detect rare adverse events with
precision. Unfortunately, systematic reviews are uncommon.
RCTs are difficult to conduct for most studies of harm.
Cohort studies - exposed and unexposed followed for
development of outcome of interest.
Case-control studies - cases with outcome of interest
compared with controls for “exposure.”
Cross-sectional studies
Case reports
23. Is the principle measure of effect from case-
control studies (cannot calculate event rates)
Used as an estimate of the risk ratio if the risk
of the disease in a population is low
Odds of exposure in the disease group divided
by odds of exposure in non-diseased group
25. Sources of Cases
Cases diagnosed or
identified
In a hospital/clinic
In a disease registry
Mass screening
Through a prior cohort
study
Incidence or prevalent
cases
Sources of Controls
Population controls
Neighborhood or friend
controls
Hospital controls
26. 2 value = 28.7 p-value < 0.05
Odds Ratio = 42.7 (95% CI = 5.9-869.5)
Cases (Reye’s) Controls
Use ASA 26 53
Did not use ASA 1 87
TOTAL 27 140
27.
28. Of the 91,523 women in the Nurses' Health Study who did not have cancer
at baseline and who were followed for 14 years, 2,341 women had
developed breast cancer by 1993.
Several studies have used standard cohort analyses to study precursors to
breast cancer, e.g. use of hormonal contraceptives, which is a covariate
easily measured on all of the women in the cohort.
If, on the other hand, one is interested in the association between gene
expression and breast cancer incidence, it would be very expensive and
possibly wasteful of precious blood specimen to assay all 89,000 women
without breast cancer.
In this situation, one may choose to assay all of the cases, and also, for each
case, select a certain number of women to assay from the risk set of
participants who have not yet failed (i.e. those who have not developed
breast cancer before the particular case in question has developed breast
cancer).
https://en.wikipedia.org/wiki/Nested_case-control_study
29.
30. Study whether exposure to a “risk factor” is
associated with a subsequent “outcome”
Select 2 populations that are similar except for
risk factor of interest
Follow them over time and see how many
have the outcome or disease
Important Point: Start with the risk, then look
for the outcome
31. Prospective cohort studies start at the present
time with exposure, then follow patients
prospectively over time into the future
Retrospective cohort studies start with exposure
that happened some time ago, then look forward
to see how many have the outcome today
Important Point: Like prospective cohort studies,
retrospective cohort studies still start with
exposure or risk and then measure the
outcome
32.
33. 2 value = 4.56 p-value = 0.03
Relative Risk = 2.7 (95% CI = 1.04-7.03)
GI bleeding (Side effect)
Yes No
Drug A 12 124
Drug B 6 178
TOTAL 18 302
34. Study
design
Advantages Disadvantages
Case-control Cheaper
Valuable for rare condition
Short duration
Baseline risk not measured
No temporal relationship
Recall bias
Nested case-
control
Cheaper than cohort
Baseline risk measured
Temporal relationship
Decrease selection bias
(cases & controls are from
the same cohort)
Controls may not represent
entire cohort due to die or
loss of follow up
Cohort Baseline risk measured
Temporal relationship
Expensive
Not valuable for rare disease
Long duration
35. Number of patients needed to be treated
for one additional patient to be harmed
Deriving NNH from odds ratio
NNH =
CER OR − 1 + 1
CER (OR − 1)(1 − CER)
Note: CER = control event rate (rate of outcome among the unexposed); OR = odds ratio
Deriving NNH from risk ratio
NNH =
1
Absolute risk increase
36. 1. A randomized trial of new drug “Ligatite” reveals that 25% of World Cup skiers
who take the drug for one year have ACL tears whereas 50% of World Cup skiers
who take the placebo for the year have ACL tears. What is the NNT?
NNT = 1/(AR reduction) = 1/(0.50-0.25) = 4
3. An advertisement for a new drug fails to mention that it increases the relative risk
of myocardial infarction by 50 % over 5 years. You read a valid study describing this
finding. What is the NNH?
Unknown without knowing the event rate in the control population.
2. The study of the drug “Ligatite” also notes that 20% of athletes taking the drug
develop clinical depression whereas 10% of athletes taking the placebo develop
depression. What is the NNH?
NNH = 1/(AR increase) = 1/(0.20-0.10) = 10
Adapted from University of Washington EBM Course Slides
37. Converting OR to NNH
Calculator available at:
http://www.cebm.utoronto.ca/practise/ca/statscal/orToNnt.htm
For OR greater than 1
1.1 1.25 1.5 1.75 2 2.25 2.5
PEER
0.05 212 86 44 30 23 18 16
0.1 113 46 24 16 13 10 9
0.2 64 27 14 10 8 7 6
0.3 50 21 11 8 7 6 5
0.4 44 19 10 8 6 5 5
0.5 42 18 10 8 6 6 5
0.7 51 23 13 10 9 8 7
0.9 121 55 33 25 22 19 18
Adapted from University of Washington EBM Course Slides
38. หญิงไทยอายุ 25 ปี กินยาคุมกาเนิด Nordette (levonorgestrel)
มีเพื่อนมาแนะนาให้เปลี่ยนเป็น Mercilon (desogestrel) จึงมา
ถามแพทย์ว่ายาคุมตัวใหม่จะมีความเสี่ยงต่อการเกิดการตีบตันของ
เส้นเลือดดาหรือไม่
P: In young women taking contraceptives
I: does desogestrel
C: when compared to levonorgestrel
O: increase risk of venous thromboembolism
41. 1. Did investigators assemble clearly defined groups of
patients similar in all important ways other than exposure?
2. Were exposures and outcomes measured in the same
ways in both groups (objective/blinded)?
3. Was follow-up sufficiently long and complete (5% and
20% rule: if <5% of subjects lost follow-up, little concern,
but if >20% lost follow-up, should be concerned)?
4. Do the results of the harm study fulfill some of the tests
for “causation”?
Adapted from University of Washington EBM Course Slides
42. Is it clear that the exposure preceded the onset of
the outcome?
Is there a dose-response relationship?
Any positive evidence from a dechallenge-
rechallenge study?
Is the association consistent across studies?
Does the association have biological plausability?
Adapted from University of Washington EBM Course Slides
43. 1. What is the magnitude of the treatment effect?
Absolute risk difference = Exposed Event Rate - Unexposed Event Rate
Relative risk difference = (Exposed Event Rate - Unexposed Event Rate)
Unexposed Event Rate
NNH = 1/(absolute risk difference)
2. How precise is this estimate of the treatment
effect?
95% CI - range of values within which we can be 95% sure that
the population value lies.
Adapted from University of Washington EBM Course Slides
44. 1. Is our patient so different from those in the study
that its results cannot apply?
2. What is our patient’s risk of benefit and harm
from agent?
3. What are our patient’s preferences, concerns,
and expectations from this treatment?
4. What alternative treatments are available?
Adapted from University of Washington EBM Course Slides