Measures of Association

Outline
Define association
Define and calculate measures of absolute, relative and attributable risk in order to quantify the magnitude of associations
Define P-value
Define Confidence Interval
Is there Association?
Is a certain exposure (E) associated with a particular disease/outcome (D) in a specific population?
Is iron- deficiency anaemia associated with (poor) academic performance in primary school students in Georgetown?
Iron deficiency anaemia may:
Seem to cause poor academic performance (positive association)
Seem to protect against poor academic performance (negative association)
Seem to have no relationship with academic performance (no association)
Reflects or mask the effect of another factor on academic performance (mediation or confounding)

Association is not causation

Estimating Risk

Regardless of which study design is used in Epidemiology , the objective is to determine whether there is an excess risk
(incidence), or perhaps a reduced risk, of a certain disease in association with a certain exposure or characteristic.Risk is the
probability of an event( such as developing a disease) occurring,Recall: Incidence is a measure of risk,Incidence of a disease in a
particular population is called absolute risk
Absolute Risk
Absolute risk indicates the magnitude of the risk in a group of people with a certain exposure but because it does not take into
consideration the risk of disease in non exposed individuals, it does not indicate whether the exposure is associated with an
increased risk of the disease.Does not stipulate an explicit comparison between two groups, but a comparison is implied.
To address the question of association, we must use approaches that involve explicit comparisons

Excess Risk
To determine whether an exposure is associated with an outcome, we use data from case-control
or cohort studies to calculate the excess risk of the disease (outcome) in people with the
exposure compared to the risk of disease in people with no exposure. Excess risk may be
expressed as: A ratio/ A difference
Excess risk-example
Hypothetical investigation of a food borne disease outbreak.

Excess risk- example: difference

Excess riskexample: ratio

Relative Risk
Note: Both case-control and cohort studies are designed to determine whether there is an association between exposure to a factor
and development of a disease.Relative Risk is the probability of an event (developing a disease) occurring in exposed people
compared to the probability of the event occurring in non-exposed people; or the ratio of the two possibilities

Relative Risk Calculations in a Cohort Study

Associations in case-control studies

In order to calculate a relative risk, we must have values for the incidence of the disease in the exposed and the incidence in
the non exposed, as can be obtained from a cohort study.In a case-control study, we do not know the incidence in the exposed
population or the incidence in the non-exposed population because we start with diseased people (case) and non-diseased
people (controls).Hence, in a case-control study we cannot calculate the relative risk directly.We can calculate an odds ratio;
may approximate RR in certain situations. In case-control studies, we can work out the proportion of cases who were
exposed and the proportion of the controls who were exposed.

An alternative approach is that instead of a proportion, we could calculate the odds

Calculating Odds

The odds of an event is the ratio of the number of ways the event can occur to the number of ways the event cannot occur.
Scenario 1:
Epi Beauty(a horse) has a 60% proabability of winning a race (P). The probability of losing is 40%, i.e., 1-P
The odds of winning are (Probability that Epi Beauty will win the race/ Probability that Epi beauty will lose the risk)
=P/(1-P) = 60%/(100%-60%) = 60%/40% = 1.5
There is a 30% probability that a disease will occur in a certain population. What are the odds of this disease occurring?
Odds ratio in a Cohort Study (RelativE Odds)

Odds ratio in a Cohort Study (Relative Odds)

Just as the ratio of the incidence in the exposed to the incidence in the non-exposed can be used to measure an association of
exposure and disease, we can also look at the ratio of the odds that the disease will develop in an exposed person to the odds
that it will develop in an non-exposed person. Either measure of association is valid in a cohort study.
Odds ratio in a Cohort Study (Relative Odds)
In a cohort study, to answer the question of whether there is an association between the exposure and the disease, we can
either use relative risk (RR) or we can use the odds ratio( relative odds).The odds ratio in a cohort study is defined as the
ratio of the odds of development of disease in exposed persons to the odds of development of disease in non-exposed
persons.

Using a standard 2x2 table, it can be calculated as follows:

RR and OR in cohort studies: exercise

RR and OR in cohort studies: exercise

Odds ratio in case-control studies

We cannot calculate the relative risk directly to determine whether there is an association between exposure and outcome.Case
control study asks a different question from cohort studies: i.e., what are the odds that a case was exposed?

Interpreting the odds ratio

Odds ratio = 1 : If the exposure is not related to the disease
Odds ratio > 1 : If the exposure is positively related to the disease
Odds ratio < 1 : If the exposure is negatively related to the disease
When is the odds ratio a good estimate of the relative risk
When the following three conditions are met:
When the cases studied are representative, with regard to history of exposure, of all people with the disease in the population
from which the cases are drawn
When the controls studied are representative, with regard to history of exposure, of all people without the disease in the
population from which the cases were drawn.
When the disease being studied does not occur frequently.
OR approximates RR when disease not
common
Points to remember

The relative odds (odds ratio) is a useful measure of

association, in and itself, in both case-control and cohort
studies
In a cohort study, the relative risk can be calculated directly
In a case-control study, the relative risk cannot be calculated
directly, so that the relative odds or odds ratio (crossproducts ratio) is used as an estimate of the relative risk when
the risk of the disease is low.

Odds ratio in a case-control study

The odds ratio in a case-control study can be calculated for :One in which the controls were not matched to the cases and One
in which the controls were matched to the cases
Odds ratio in a case-control study
(unmatched): exercise

Odds ratio in a case-control study (matched)

In selecting the study population in case-control studies, controls are often selected by matching each one to a case according to
variables that are known to be related to disease risk, such as sex, age, or race (individual matching or matched pairs).The results
are analysed in terms of case-control pairs rather than for individual subjects.
Odds ratio in a case-control study (matched)
If exposure is dichotomous (a person is either exposed or not), there are only four types of case control pairs that are possible.

A pairs: both the cases and controls were exposed

B pairs: the case was exposed and the control was not
C pair: the case was not exposed and the control was exposed
D pair: Neither the case nor the control was exposed
A and D are numbers of concordant pairs
B and C are numbers of discordant pairs
Calculation of the odds ratio in a case control study (matched)
Calculations of the odds ratio in a matched pair study is based on the discordant pairs (only b and c).the concordant pairs (a and
d) are ignored, because they do not contribute to our knowledge of how cases and controls differ in regard to past history of
exposure.

Odds ratio in matched case-control

study: exercise

Attributable Risk
Relative risk is important as a measure of the strength of the association which is important for causal inferences
(establishing aetiology). Attributable risk answers a different question:
How much of the disease that occurs can be attributed to a certain exposure?
Attributable risk, is defined as the amount or proportion of disease incidence (or disease risk) that can be attributed to a
specific exposure.Example, how much of the lung cancer risk experienced by smokers can be attributed to smoking
Attributable risk is, in many ways, more important than RR for clinical practice and public health because it can indicate
how much of the occurrence of a disease we may be able to prevent if we eliminated the exposure.

calculating attributable risk in the

exposed

Calculating attributable risk in the exposed

Incidence of a disease that is attributable to exposure in the exposed group:

Attributable risk can be expressed as a proportion of the risk in exposed persons due to the exposure:

important note:
If all the incidence of a disease were the result of a single factor, the attributable risk for that factor would be
100%. This hardly happens
calculating attributable risk in the
exposed

Calculating attributable risk in the exposed

If we prefer, it can be expressed as a proportion.
The proportion of the total incidence in the exposed group that is attributable to the exposure:

Calculating attributable risk in the

exposed

Calculating attributable risk in the

populatio

Population attributable risk: calculation

Incidence of the disease in whole populations often not known.PAR usually obtained using information from various
sources on:
The incidence among exposed
The incidence among non-exposed
The proportion of the total population that is exposed
Relative versus attributable

Population attributable risk

P-Value
P-value: The probability of obtaining the observed sample results (or more extreme values) given that the null hypothesis is
true.Null hypothesis is the hypothesis of no effect, no correlation, no association, etc.Example: CMV infected babies have
the same average birth weight as non infected babies.B-thalassaemia does not have any effect on ferritin level.P-value is
used as a tool for deciding whether to reject the null hypothesis.
Before the statistical test is performed, a threshold is chosen, called the significance level of the test, traditionally 5% (0.05)
or 1% (0.01)and denoted as (alpha).Significance level is the probability of rejecting the null hypothesis given that it is true.
It is implying that it is acceptable to have a 5% or 1% probability of incorrectly rejecting the null hypothesis.
Interpreting P-value
The p-value is a number between 0 and 1
A small p-value (typically 0.05) indicates strong evidence against the null hypothesis, so you reject the null hypothesis
A large p-value (>0.05) indicates weak evidence against the null hypothesis, so you fail to reject the null hypothesis
Confidence Intervals
A confidence interval gives an estimated range of values which is likely to include an unknown population parameter, the
estimated range being calculated from a given sample data.There are different confidence intervals : 90%, 95%, 99%, etc.In
the interpretation of a 95% confidence interval, we can say, we are 95% confident that the true value of the parameter is in
our confidence interval.
If a corresponding hypothesis test is performed, the confidence level is the compliment of the respective level of
significance, i.e. a 95% confidence interval reflects a significance level of 0.05.