Evaluating Causation Flashcards
Definitive groups in case-control studies
Cases
Controls
1st column in table
Exposed
Non exposed
Total
2nd column
Cases
A
C
A+Cd
3rd column
Controls
B
D
B+D
4th column
A+B
C+D
A+B+C+D
1st row
Cases, Controls
2nd row
Exposed, A, B, A+B
3rd Row
Non exposed, C, D, C+D
4th row
Total, A+C, B+D, A+B+C+D
Case exposure rate
A/A+C
Control exposure rate
B/B+D
Odds of exposure in cases
A/C
Odds of exposure in controls
B/D
What is the odds ratio?
Odds of being exposed if you are a case rather than a control
Calculation for odds ratio?
AD/BC
At what point does odds ratio = relative risk?
If disease is rare i.e. prevalence <5%
Values for odds ratio
0 to infinity
How does one calculate CI from odds ratio
Using log transformations
What can we calculate from cohort studies?
Risk rates
Why can risk not be calculated from case-control studies?
Risk = incidence, which needs a prospective observation
1st column in cohort studies
Exposed cohort
Non exposed
Total
2nd column in cohort studies
Disease develops
A
C
A+C
3rd column in cohort studies
No disease
B
D
B+D
4th column in cohort studies
A+B
C+D
A+B+C+D
1st row in cohort studies
Disease develops, No disease
2nd row in cohort studies
Exposed cohort, A, B, A+B
3rd row in cohort studies
Non exposed, C, D, C+D
4th row in cohort studies
Total, A+C, B+D, A+B+C+D
Absolute risk of disease in exposed
A/A+B
Absolute risk of disease in non-exposed
C/C+D
Odds of finding disease among exposed
A/B
Odds of finding disease in non-exposed
C/D
Odds ratio for cohort studies
AD/BC
I.e. same as that for case-control studies
What is the risk difference?
Difference between two absolute risks (exposed and non-exposed)
Formula for risk difference
Absolute risk in exposed group - absolute risk in non-exposed group
Another name for risk difference?
Absolute risk reduction
Attributable risk
Calculation of Relative risk
[A/(A+B)] / [C(C+D)]
What is the relative risk?
Ratio between the probability of developing the disease if exposed to the probability of developing the disease if not exposed
Another name for relative risk
Risk ratio
Calculation of population attributable risk (PAR)?
Absolute risk reduction x proportion exposed
What does PAR tell us?
Number of cases that could be eliminated by eradicating the exposure
What is the rare disease assumption?
When the probability of an event is low (<5% prevalence), odds ratio will = relative risk
What happens to odds ratio as prevalence increases?
Odds ratio increases
When are odds ratio used?
Case control studies
Cohort studies
Analysis of dichotomous outcomes in clinical trials
Range of values of probabilities
0-1
Range of values of absolute risk
0-1
0 = no one in group has disease 1 = everyone in group has disease
Range of values of attributable risk
- 1 to +1
- 1 = risk greater in non-exposed group
Range of values of relative risk
0 to infinity
Meaning of value of relative risk
<1 = risk in exposed is less than risk in non-exposed
1 = risks are equal in both groups
Range of values for odds ratio
0 to infinity
