MPH 6011 Final Exam 2023 with complete solution
True or false: A retrospective cohort study is an observational study but a
prospective cohort study is an intervention study.
False
What study designs are both analytic, observational studies?
Retrospective and prospective cohort studies
What is the measure of association if we start with a cohort free of disease and
analysis is by cumulative incidence?
relative risk
What is the ratio of cumulative incidence rate in exposed compared to cumulative
incidence rate in the unexposed?
relative risk
What is the measure of association if we start with a cohort free of disease and
analysis is by incidence rates with person-time in the denominator?
rate ratio
In a cohort study, several design and implementation problems can impact the
strength of the measure of association. Choose which of these problems below is
least likely in a cohort study?
a. information bias
b. absence of temporality
c. loss to follow-up
d. none of the above
b
Among the items listed below, which design feature or bias is NOT similarly
found in both case-control and cohort studies?
a. both make comparisons between exposed and unexposed groups
b. confounding
c. both measure exposure before disease occurs
d. both can have misclassification of exposure
c
Exposure is measured __________ disease occurrence in cohort studies.
a. before
b. after
a
In a cohort study, the association between smoking and coronary heart disease
was investigated by selecting for study a group of 3000 smokers and 5000 non-
smokers who were free of disease at baseline. Both groups were followed for the
development of coronary heart disease. 84 smokers developed coronary heart
disease and 87 non smokers did too. 4913 persons did not develop disease and
did not smoke. The measure of association relating exposed to unexposed would
be:
(84/3000)/(87/5000)
, True or false: Attributable risk tells us how much disease can be prevented if
exposure is removed from the population.
True
In a 20 year cohort study with 30 cases of heart disease in exposed group and
54,308 person years, while the unexposed group had 60 cases and 51,477 person-
years. The measure of association showed 0.47. The best interpretation of this
finding was:
a. exposed group had a greater risk of heart disease than unexposed group
b. unexposed group had reduced risk of heart disease compared to exposed
group
c. exposed group had a reduced risk of heart disease compared to the unexposed
group
c
Which of the following is an advantage of cohort studies?
a. can examine multiple outcomes from a single exposure
b. can match on confounders
c. can avoid recall bias
d. can enroll cases at beginning of study
a
True or false: A case control study can be embedded within a cohort?
True
What are the two examples of case-control design embedded within a cohort?
Nested case control and case-cohort
What is the design feature unique to experimental studies?
Randomization
True or false: The purpose of randomization is to eliminate conscious bias by
provider on patient selection for treatment group.
True
What are the two reasons why randomization is preferred to non-randomized
studies?
1. eliminates conscious bias by provider on patient selection for treatment group
2. averages unknown confounders and makes groups similar
The primary purpose of a double blind study is to:
a. achieve comparability of treated and untreated subjects
b. reduce efforts of sampling error
c. avoid provider bias and participant bias
d. avoid participant bias and increase recruitment
c
Randomized controlled trials do not include which design:
a. randomization
b. blinding
True or false: A retrospective cohort study is an observational study but a
prospective cohort study is an intervention study.
False
What study designs are both analytic, observational studies?
Retrospective and prospective cohort studies
What is the measure of association if we start with a cohort free of disease and
analysis is by cumulative incidence?
relative risk
What is the ratio of cumulative incidence rate in exposed compared to cumulative
incidence rate in the unexposed?
relative risk
What is the measure of association if we start with a cohort free of disease and
analysis is by incidence rates with person-time in the denominator?
rate ratio
In a cohort study, several design and implementation problems can impact the
strength of the measure of association. Choose which of these problems below is
least likely in a cohort study?
a. information bias
b. absence of temporality
c. loss to follow-up
d. none of the above
b
Among the items listed below, which design feature or bias is NOT similarly
found in both case-control and cohort studies?
a. both make comparisons between exposed and unexposed groups
b. confounding
c. both measure exposure before disease occurs
d. both can have misclassification of exposure
c
Exposure is measured __________ disease occurrence in cohort studies.
a. before
b. after
a
In a cohort study, the association between smoking and coronary heart disease
was investigated by selecting for study a group of 3000 smokers and 5000 non-
smokers who were free of disease at baseline. Both groups were followed for the
development of coronary heart disease. 84 smokers developed coronary heart
disease and 87 non smokers did too. 4913 persons did not develop disease and
did not smoke. The measure of association relating exposed to unexposed would
be:
(84/3000)/(87/5000)
, True or false: Attributable risk tells us how much disease can be prevented if
exposure is removed from the population.
True
In a 20 year cohort study with 30 cases of heart disease in exposed group and
54,308 person years, while the unexposed group had 60 cases and 51,477 person-
years. The measure of association showed 0.47. The best interpretation of this
finding was:
a. exposed group had a greater risk of heart disease than unexposed group
b. unexposed group had reduced risk of heart disease compared to exposed
group
c. exposed group had a reduced risk of heart disease compared to the unexposed
group
c
Which of the following is an advantage of cohort studies?
a. can examine multiple outcomes from a single exposure
b. can match on confounders
c. can avoid recall bias
d. can enroll cases at beginning of study
a
True or false: A case control study can be embedded within a cohort?
True
What are the two examples of case-control design embedded within a cohort?
Nested case control and case-cohort
What is the design feature unique to experimental studies?
Randomization
True or false: The purpose of randomization is to eliminate conscious bias by
provider on patient selection for treatment group.
True
What are the two reasons why randomization is preferred to non-randomized
studies?
1. eliminates conscious bias by provider on patient selection for treatment group
2. averages unknown confounders and makes groups similar
The primary purpose of a double blind study is to:
a. achieve comparability of treated and untreated subjects
b. reduce efforts of sampling error
c. avoid provider bias and participant bias
d. avoid participant bias and increase recruitment
c
Randomized controlled trials do not include which design:
a. randomization
b. blinding
