Lecture Four-Measurements Of Association And Study Design

Question 1

Multi factorial disease?

Answer 1

There are necessary and contributing causes.
Necessary cause = often pathogen, plus factors that contribute to the likelihood of becoming ill.
A; is the necessary cause (the pathogen) which may not necessarily result in disease
B, C, D, E; are contributing causes (risk factors)
A + B + C + D + E together = sufficient cause (SC)

Question 2

What is the Web of Causation?

Answer 2

Describes a complex web of interacting factors involving Host, Agent, and Environment that is responsible for many diseases.

Question 3

Do Epidemiological studies find Association, or Causation, or both?

Answer 3

Association only.

Causation can be proven in experimental studies.

Question 4

What is a Measure of Association?

Answer 4

How likely it is that there is an association between a particular risk factor and the likelihood of disease.
The objective is to compare the frequency of disease in individuals exposed to a risk factor to the frequency in individuals not exposed to that risk factor.
It is a comparison of risk, measured by the strength of association, and the strength of the potential impact.

Question 5

What is the Strength of Association?

Answer 5

The magnitude of association between a risk factor and disease.
It can be determined by calculating;
Risk Ratio (Relative Risk)
Odds Ratio

Question 6

Measures of of the potential impact?

Measures of effect in the population?

Answer 6

Measures of of the potential impact?
Attributable Risk (= risk difference)
Attributable Fraction
Measures of effect in the population?
Population Attributable Risk
Population Attributable Fraction

Question 7

Mono factorial disease?

Answer 7

A single cause will always lead to disease (poisoning, burns)

Question 8

Risk Ratio/ RELATIVE RISK? (RR)

Answer 8

Is the RATIO of incidence risk in the exposed group, to incidence in an unexposed group.
Performed in cohort and cross-sectional studies, values range from 0>infinity. Assume there is never a zero chance of contracting the disease, but it may be a minute value.

Question 9

How is Relative Risk (RR) interpreted?

Answer 9

The disease is (RR) times as likely to occur among those exposed to the suspected risk factor as among those with no such exposure.
RR 1 = positive association with disease (and may or may not be causal)

Question 10

Formula for Relative Risk (RR)?

Answer 10

RR = (a / (a + b)) / (c / (c + d))
RR = P (D+IE+) / P (D+IE-)

Question 11

Odds Ratio (OR)?

Answer 11

The odds of disease in exposed group divided by odds of disease in unexposed group.

Question 12

How is Odds Ratio (OR) interpreted?

Answer 12

The interpretation of the result is that the odds of disease among exposed animals being OR times the odds of disease among non-exposed.

Question 13

Odds Ratio (OR) formula?

Answer 13

OR = (a / b) / (c / d) = ad/cb
OR = Odds (D+IE+) / Odds (D+IE-)

Question 14

Attributable Risk (AR) (= risk difference) ?

Answer 14

It is the difference in disease frequency between the exposed and unexposed group.
It is a measure of additional risk, or additional disease, present when the risk factor (exposure), and is calculated by subtracting the risk in the non-exposed from the risk in exposed animals.
AR = IR (exposed) - IR (unexposed)
= a / (a + b) - c / (c + d)

Question 15

How is Attributable Risk (AR) interpreted?

Answer 15

The risk, over and above the baseline risk, of developing the disease is increased by the attributable risk for those individuals exposed to the risk factor.

Question 16

Attributable Fraction (AF)?

Answer 16

The proportion of the total risk in exposed animals that is due to the risk factor. Even if the additional risk is large, it may be negligible if the initial risk was small.
AF = AR / (a / (a + b))
= {[a / (a + b)] - [c / (c + d)]} / [a / (a + b)]
= (RR - 1) / RR
= (OR - 1) / OR (estimated AF)

Question 17

Measures of Effect in the Population?

Answer 17

A relatively weak risk factor that is quite common may be a more important determinant of disease in the population as a whole than a strong risk factor that is infrequently encountered.
It is useful to determine how much a risk factor contributes to the overall rates of disease in groups or animals rather than individuals.

Question 18

Population Attributable Risk?

Answer 18

Similar to attributable risk (AR), but reflects the total population. It is the excess frequency of a disease in a population (both exposed and unexposed individuals) that is associated with the occurrence of a risk factor.
It is the overall risk of disease in population that is due to exposure.
PAR = IR (total population) - IR (unexposed)
= [(a + c) / n] - [c / (c + d)]

Question 19

Population Attributable Fraction?

Answer 19

The proportion of disease in the whole population (consisting of both exposed and unexposed individuals) that is attributable to the risk factor/exposure.
PAF = PAR / IR (total)
= PAR / [(a + c) / n]

Question 20

Confounding?

Answer 20

A third variable that distorts the observed relationship between the exposure and outcome.
It is a confusion and should be controlled for if possible in order to obtain a more accurate estimate of the true association between the exposure and disease under study.
Age and breed are common causes of confounding.

Question 21

List methods to control for confounding?

Answer 21

1. Randomisation
2. Restriction
3. Matching
4. Stratification
5. Multi variable Analysis

Question 22

Discuss methods to control for confounding?

Answer 22

1. Randomisation
   Subjects or groups of subjects are randomly assigned to a hypothesised preventative or therapeutic intervention.
   With sufficient sample size, it assures that both known and unknown confounders are controlled. If the sample size is not large enough it may not be enough to control for confounding.
2. Restriction
   Study participation is restricted to individuals who fall within a specified category or categories of the confounder.
   It is straightforward, convenient an inexpensive, but a sufficiently narrow restriction range may severely reduce the number of suitable animals.
3. Matching
   All levels of the confounding factor are allowable for study inclusion, but subjects are selected in a way that potential cofounders are distributed equally among the study groups.
   It allows for a greater analytic efficiency, with an adequate number of cases and controls at each level of the confounder. It can be difficult, time consuming, and expensive to find comparison subjects with the right set of characteristics on each matching variable.
4. Stratification
   Evaluation of the exposure/disease association within homogenous categories or strata of the confounding variable.
5. Multi variable Analysis
   A technique that takes into account a number of variables simultaneously. It is a mathematical model to describe the association between exposure and disease, as well as other variables that may confound or modify the effect of the exposure.