From midterm practice exams

Question 1

We are assessing the relation between measles vaccination and mortality. Define the causal effect of interest in terms of potential outcomes.

Answer 1

The causal effect of interest is the incidence of mortality among vaccinated children compared to the incidence of mortality among the same children had they not been vaccinated.

Question 2

Interpret an incidence ratio of 0.72

Answer 2

The incidence rate ratio indicates that the rate of mortality among children who received a measles vaccination was 28% lower than the rate of mortality among children who did not receive a vaccination

Question 3

3 criteria for confounder

Answer 3

i) be associated with outcome
ii) cause the outcome
iii) not be a mediator

Question 4

2 criteria for case-crossover study

Answer 4

i) abrupt effect of exposure on disease

ii) well-defined onset of disease

Question 5

Odds ratio is the probability…

Answer 5

Probability of the event / probability of non-event

Question 6

The same _______ can be selected at the beginning of follow-up and compared to various _____

Answer 6

control group; case series (multiple diseases)

Question 7

What interpretation are you gonna make if you have a LR- of 0.04?

Answer 7

The LR- suggests that individuals who are truly negative for disease are far more likely to test negative than individuals who are truly positive.

Question 8

Non-differential misclassification biases…

Answer 8

the association towards the null

Question 9

Odds ratio from a cumulative-case control study approximates…

Answer 9

The risk ratio and the rate ratio, if the rare disease assumption is met

Question 10

Odds ratio from a density case control study approximates…

Answer 10

The rate ratio

Question 11

Odds ratio from a case cohort study approximates…

Answer 11

The risk ratio

Question 12

Two ways to adjust for confounding bias at the design stage of case control study?

Answer 12

• Matching

- Restriction

Question 13

What type of bias may be introduced by matching?

Answer 13

Selection bias, to the extent that the distribution of the exposure in the controls becomes more similar to that of the cases, and misrepresents the exposure distribution of the source population

Question 14

Lack of allocation concealment can induce 2 biases. Which, and how?

Answer 14

1. Selection bias: if investigators select preferentially certain patients from source population to be included in study base
2. Confounding bias: if investigators broke randomization and diverted sicker patients to one of the two groups

Question 15

Give an interpretation of an incidence rate difference of 793 per 100,000 person year

Answer 15

The incidence rate difference indicates that pesticide exposure is associated with 793 additional cases of depression for every 100,000 person year of follow-up compared to non-exposure

Question 16

What is the problem of measuring a time-varying variable as time-fixed?

Answer 16

May introduce error in measurement of exposure, and thus bias results.

Question 17

What is a stage shift?

Answer 17

When more patients are diagnosed at an earlier stage of the disease

Question 18

Interpret a risk ratio of 0.5

Answer 18

Participants in the treatment arm are 50% less likely to develop outcome than controls

Question 19

Interpret a risk difference of -2.1%

Answer 19

There are 21 fewer diseases occurrences per 1000 participants in the treatment group compared to the control group