1. Measurements of Risk & Association …
A.Prof. Dr.
Maha A. AL-Nuaimi
Ph.D. Com.Med.

2. Learning objectives
Understand measures of absolute, relative and attributable risks, AR%, Odds Ratio
Estimate absolute, relative and attributable risks, AR%, Odds Ratio, from epidemiological studies.
Interpret Relative risk (RR) & Odds ratio (OR), AR% as a measure of association between exposure and Disease

3. - How can you measure the risk?
Definitions
Risk
Probability of an individual developing a disease or change in health status over a fixed time interval. ( Absolute risk)
- How can you measure the risk?
- How do you know that this risk is high or low?

4. A risk factor
Is any attribute, characteristic or exposure of an individual that increases the likelihood of developing a disease or injury.

5. Association A statistical relationship between two or more variables.

6. How we can estimate the magnitude of the association between exposure & health-related events ??
By Epidemiological Studies (analytic) As :
It indicates the likelihood of developing the disease in the exposed group relative to the non-exposed;
Cohort study.
Case –Control study

7. In case-control studies – Odds Ratio (as the incidence is not known)
In Cohort Study or prospective studies where the risk is estimated directly by:
Relative Risk (RR),
Attributable Risk,
Attributable Risk %,
In case-control studies –
where the risk is estimated indirectly by
Odds Ratio (as the incidence is not known)

8. Relative Risk (Risk Ratio)
Estimates the magnitude of an association between exposure & disease.
It is important in the causation & in the policy of prevention.
Risk Ratio (RR) = incidence among exposed group
incidence among non exposed

9. 2x2 Table to determine the association..
HEALTH EVENT (disease, death, or any outcome) + -
EXPOSURE a b c d

10. Start Outcome 500 children 300 Children (<12 yrs) 1000 200 120
Family smoker
500 children
Exposed
Diseased
300
Children
(<12 yrs)
1000
Not diseased
200
1 year
Diseased
120
Family non-smoker
500 children
Not exposed
Not diseased
380

11. Rate: Incidence rate
Incidence of Resp. Infection among exposed children:
500 * = 60%
Incidence of Resp. Infect. Among non exposed children:
500 * = 24%

12. Cohort Study (cont.) Interpretation of RR
Relative Risk Incidence rate among exposed
Risk Ratio =
Incidence rate in non exposed.
60%
24% = ?
RR = 1: means zero (no effect of the exposure)
RR > 1: means (positive effect of the exposure)
RR < 1: means negative effect of exposure (protective)
Interpretation of RR
Children of family Smokers were 2.5 times as likely to develop chest infections as were children of family non-smokers.
Cohort Study (cont.)

13. AR = Inc.exp – Inc. non-exp
Attributable risk (AR), Risk difference) Excess Risk, (Incidence Difference:
Excess risk of the disease in those exposed
compared with those none exposed.
AR = Inc.exp – Inc. non-exp

14. Base-line risk
Incidence
A R
I. among
exposed
I. among non exposed

15. Attributable Risk: interpretation
60% % = %
interpretation
36% of Resp. Infection among <12 children can be attributed to family smoking.

16. Attributable Risk % (AR%): (population attributable risk)
The proportion prevented in Pop.
The percentage of cases which could be prevented if we remove the exposure to that factor.
AR% = Ie – Io * 100 Ie
AR% = 36% / 60% = 60% interpretation?

17. AR% (population attributable risk)
The value of AR% in any pop. depends on ?
Example:
AR% = 92% (for lung cancer with smoking)
AR% = 38% (for lung CHD with smoking)
Upon which the cessation of smoking will have a greater impact??

18. Rate Ratio
A rate ratio compares two groups in terms of , person-time rates, or mortality rates.
Rate ratio = rate for group of primary interest
rate for comparison group
The interpretation of the value of a rate ratio is similar to that of the risk ratio.

19. Lung Cancer 166 7 166 / 7 = 23.7 Annual Death Rates / 100,000 persons
Exposure Category
Lung Cancer
166 7
166 / 7 = 23.7
Heavy smokers
Nonsmokers
Measures of Risks
Rate ratio:
Doll and Hill study : Mortality of British doctors cited from Mausner, 1985

20. OR (Odds Ratio, Relative Odds)
It is a measure of choice for case-control study.
In (CCS), we cannot calculate the IR → NO RR.
OR used to measure the association between
exposure & disease in (CCs), instead of RR (cohort).

21. Case Control Studies CROSS -RATIO Cases Controls a b Non Exposed c d
Leukemia
Cases
Controls
Benzene Exposed a b
Benzene
Non Exposed c d
a+c
b+d
CROSS -RATIO

22. OR: - Odds refer to single entity“ - Odds of factor among cases divided by odds of factor among control” OR =

23. Odds ratio
a = No. of persons with disease and with exposure of interest b = No. of persons without disease, but with exposure of interest c = No. of persons with disease, but without exposure of interest d = No. of persons without disease & without exposure of interest a + c = total number of persons with disease (“cases”) b + d = total number of persons without disease (“controls”)

24. Example: A study of TB with bad nutrition in a case-control study.
HEALTHY TB
b=40
a=100
Present
d=140
c=80
Absent
b + d
180
a + c
Risk Factor
OR= 100x140 / 40x80 = 4.4
interpretation
TB patients have 4 times history of bad nutrition than healthy.

25. In interpreting the risk estimates, say for:
RR there is a (…) folds/times risk to develop the outcome "?" on being exposed to (!).
For AR (& AR% too); the excess risk to develop the outcome "?" form the exposure to (!) is (…).
For OR, those with the outcome "?" are (…) times more likely to be exposed to (!).

26. Which of the above measures quantifies the strength of the association?

27. Measuring of Risk in epidemiology, does it determine the cause of a disease in a given individual ??