WEEK 5: COFOUNDING, EFFECT MODIFICATION

Question 1

Threats to validity

Answer 1

chance, bias (selection, information), cofounding

Question 2

what is cofounding

Answer 2

distortion of the actual association due to a mixing of effects between the exposure and an incidental variable(s), known as a cofounder

Question 3

why does cofounding occur

Answer 3

• Because the exposure ground and the unexposed ground are not exchangeable
• They differ by factors other than their exposure status
• Cofounders are particularly a problem for observational studies
• Randomization in experimental studies removes the possibility of cofounding by making the exposed and unexposed ground the same except for the exposure status

Question 4

criteria for defining a cofounding variable

Answer 4

1. Casually associated with the outcome (a true risk factor)
2. Noncausal or casually associated with the exposure
3. Not an intermediate (mediator) in the casual pathway between exposure and outcome

Question 5

5 steps to assessing cofounding

Answer 5

1. conceptualize the relationship
2. think of variables with criteria
3. exclude those on the pathway
4. identify true cofounders
5. deal with cofounders

Question 6

dealing with cofounding at the design stage

Answer 6

• Restriction
  o Limit study inclusion criteria with respect to cofounding factors; study only men or women
• Matching
  o Produce case and control (exposed/non exposed) groups that have similar characteristics
• Randomization
  o Experimental studies

Question 7

dealing with cofounding at the analysis stage

Answer 7

• Standardization
  o Age standardization is in fact “adjustment” for age
• Stratified analysis
• Include cofounding factors ina. Multivariate regression model

Question 8

effect modification

Answer 8

*the 3rd factor, called the effect modifier, modifies the effects of the exposure on the outcome
o not a cause of the exposure and outcome ( that’s cofounding)
o not on the casual pathway (intermediary)

Question 9

cofounding is a problem for________

Answer 9

validity

Question 10

Mediator

Answer 10

• An intermediary variable occurs between an exposure and an outcome
• Do not need to adjust for the effects of intermediaries in assessing exposure – disease (E – D) associations

Question 11

what is interaction

Answer 11

when the incidence rate of disease in the presence of 2 or more risk factors differs from the incidence rate expected to result from their individual factors

Question 12

positive and negative interaction

Answer 12

The effect can be greater than what is expected = positive interaction; synergism
The effect can be less than what is expected = negative interaction; antagonism

Question 13

what is the best way to reduce sampling error

Answer 13

o increase the rise of the population/study sample

Question 14

p value meaning

Answer 14

If the p-value < 5%/ probability that results arose by chance <5% (type I error, ) results are considered statistically significant

Question 15

Interpreting P-Values & Confidence Intervals

Answer 15

If p< 0.05, this does not guarantee the association is real or if p> 0.05, does not mean there is no association; only means that there is not enough evidence to conclude if there is an association

Question 16

what tools can help assess results if association exists or not

Answer 16

hypothesis tests and p values

Question 17

When a significance level of 0.05 is used - what does this mean

Answer 17

there is only a 1 in 20 chance that the null hypothesis will be rejected in error (i.e you will conclude there is an association when there is none, the chance of this happening increases as more tests are done)

Question 18

what does an RR mean for significance

Answer 18

A RR > 2,0 would be considered strong, thus practically significant
An RR < 2.0 wild be dismissed immedisately bc a modest relative risk can lead to high absolute or attributable risk