Work group – questions and answers – Daniëlle Band
Work group 1 (questions + answers)
Assignment 1
Read the following passages. Identify the study designs and determine what the outcome and exposure
of interest are. Please justify your choices.
a) Questionnaires were mailed to every 10th person listed in the city telephone directory. Each
person was asked to list age, sex, smoking habits, and respiratory symptoms during the
preceding seven days. About 20% of the questionnaires were completed and returned. About
10% of respondents reported having upper respiratory symptoms.
Exposure and outcome are measured at the same time, so it is a cross-sectional study
(observational because no manipulation, and analytical).
• The outcome is respiratory symptoms
• The exposure is smoking habits.
b) A sample of 1,500 employees from an airline company were initially examined in 1951. They
were classified by diagnostic criteria for coronary artery stenosis (CAS). New cases of CAS have
been identified by radiological examination every three years, complemented by ‘cause-of-
death’ certificates. Heart attack rates in different subgroups have been computed.
This is a cohort study (observational, analytical, prospective because forward in time). It uses
time to judge the direction of causality by following a group from the present, assessing the
exposure now and assessing the outcome in the future (outcome has not happened yet).
• The outcome is heart attack
• The exposure is CAS.
b) A random sample of rural community members was selected from four census tracts, and each
person was examined for depression. All persons without disease were randomly assigned to
either a two-years program of physical activity encouragement or a two-years program of
lifestyle advice.
This is an RCT study (experimental, because the situation is manipulated). Participants have
equal chance of being in either group (due to randomisation) (everyone is at risk of getting a
depression). Exposure takes place before outcome (RCTs are always prospective in design).
• The outcome is depression
• The exposure is physical activity encouragement or a two-years program of lifestyle
advice.
c) Patients admitted for carcinoma of the stomach and patients without a diagnosis of cancer are
interviewed about their tobacco chewing history. The question is whether there is an
association between tobacco chewing and gastric cancer.
The cases are matched with the controls, so case control study (observational, analytical). Starts
with the outcome and then looks to determine the exposure (case-control studies are always
retrospective).
• The outcome is carcinoma of the stomach (gastric cancer)
• The exposure is the tobacco chewing.
1
,Work group – questions and answers – Daniëlle Band
d) Researchers retrieved census data at regional level, including median wealth index for the
respective regions, from the Ugandan Bureau of Statistics. These data were matched with air
pollution levels per region which they measured over a one-month period. Data were analyzed
using geographic information systems (GIS), producing maps showing air pollution levels per
region.
You look at regional levels and not individual levels, so it is ecological (observational, descriptive,
ecological)
• The outcome air pollution
• The exposure is wealth level by region.
Assignment 2
You have developed the hypothesis that automobile drivers who regularly sleep less than six hours a
night have higher incidence of fatal accidents compared to those that sleep longer than six hours. Think
about the design of a case-control study to test this hypothesis. Consider:
a) How will you define a ‘case’? Are there any special considerations?
A case is an automobile driver that has been involved in a fatal accident. Special considerations
are that the accident are the primary cause of death and that the cases are drivers and not
passengers.
b) Where would you find the cases?
Cases can be found in police reports or medical records
c) How would you define a ‘control’?
The control are automobile drivers that have not been involved in a fatal accident
d) Name a suitable population from which to choose the controls for your study.
Controls are similar but not too similar to cases, but without the outcome of interest. Controls
can be drivers from similar areas with comparable characteristics and random samples from
motor vehicle license records. The study base would entail everyone who was an active driver
at the time of the accident.
e) What major characteristic must you strive to measure similarly in each study participant (both
the cases and controls)?
The hours they sleep at night, safety features of the car, road conditions, weather, drivers age,
sex, medical history, passenger restraint systems in use.
f) What difficulties could be encountered in measuring these characteristics?
The hours they sleep at night, safety features of the car, road conditions, weather, drivers age,
sex, medical history, passenger restraint systems in use.
2
, Work group – questions and answers – Daniëlle Band
Assignment 3
What is the importance of randomization in an intervention trial, and what does it accomplish?
A randomized controlled trial causes data that is free of bias. It makes it more likely that any differences
in outcomes between groups are a result of the treatment they received. Instead of differences in
outcomes being influenced by other factors like characteristics (age, sex)
Assignment 4
What is meant by the phrase ‘ecological fallacy’?
Ecological fallacy occurs when you make assumptions about individuals based on group-level data. In
other words, it's a mistake to think that characteristics observed in a group apply to every individual
within that group.
Assignment 5
Different study designs have particular advantages and disadvantages. Contrast the casecontrol and
prospective cohort design with respect to the following factors.
a) Costs;
Case-control is inexpensive and prospective cohort study is more expensive
a) Time required for completion of the study;
Case-control can yield important findings in a short time, cohort studies are generally more
time consuming
b) Bias;
Case-control studies are very susceptible for both selection and information bias (participant
cannot recall certain events, who do we select as controls?)
c) What can be estimated (with regards to effect/association measures);
Cohort studies allow for exploration of natural developments between exposure and
outcome, whereas with case-control studies you retrospectively assume two things to be
associated.
3
Work group 1 (questions + answers)
Assignment 1
Read the following passages. Identify the study designs and determine what the outcome and exposure
of interest are. Please justify your choices.
a) Questionnaires were mailed to every 10th person listed in the city telephone directory. Each
person was asked to list age, sex, smoking habits, and respiratory symptoms during the
preceding seven days. About 20% of the questionnaires were completed and returned. About
10% of respondents reported having upper respiratory symptoms.
Exposure and outcome are measured at the same time, so it is a cross-sectional study
(observational because no manipulation, and analytical).
• The outcome is respiratory symptoms
• The exposure is smoking habits.
b) A sample of 1,500 employees from an airline company were initially examined in 1951. They
were classified by diagnostic criteria for coronary artery stenosis (CAS). New cases of CAS have
been identified by radiological examination every three years, complemented by ‘cause-of-
death’ certificates. Heart attack rates in different subgroups have been computed.
This is a cohort study (observational, analytical, prospective because forward in time). It uses
time to judge the direction of causality by following a group from the present, assessing the
exposure now and assessing the outcome in the future (outcome has not happened yet).
• The outcome is heart attack
• The exposure is CAS.
b) A random sample of rural community members was selected from four census tracts, and each
person was examined for depression. All persons without disease were randomly assigned to
either a two-years program of physical activity encouragement or a two-years program of
lifestyle advice.
This is an RCT study (experimental, because the situation is manipulated). Participants have
equal chance of being in either group (due to randomisation) (everyone is at risk of getting a
depression). Exposure takes place before outcome (RCTs are always prospective in design).
• The outcome is depression
• The exposure is physical activity encouragement or a two-years program of lifestyle
advice.
c) Patients admitted for carcinoma of the stomach and patients without a diagnosis of cancer are
interviewed about their tobacco chewing history. The question is whether there is an
association between tobacco chewing and gastric cancer.
The cases are matched with the controls, so case control study (observational, analytical). Starts
with the outcome and then looks to determine the exposure (case-control studies are always
retrospective).
• The outcome is carcinoma of the stomach (gastric cancer)
• The exposure is the tobacco chewing.
1
,Work group – questions and answers – Daniëlle Band
d) Researchers retrieved census data at regional level, including median wealth index for the
respective regions, from the Ugandan Bureau of Statistics. These data were matched with air
pollution levels per region which they measured over a one-month period. Data were analyzed
using geographic information systems (GIS), producing maps showing air pollution levels per
region.
You look at regional levels and not individual levels, so it is ecological (observational, descriptive,
ecological)
• The outcome air pollution
• The exposure is wealth level by region.
Assignment 2
You have developed the hypothesis that automobile drivers who regularly sleep less than six hours a
night have higher incidence of fatal accidents compared to those that sleep longer than six hours. Think
about the design of a case-control study to test this hypothesis. Consider:
a) How will you define a ‘case’? Are there any special considerations?
A case is an automobile driver that has been involved in a fatal accident. Special considerations
are that the accident are the primary cause of death and that the cases are drivers and not
passengers.
b) Where would you find the cases?
Cases can be found in police reports or medical records
c) How would you define a ‘control’?
The control are automobile drivers that have not been involved in a fatal accident
d) Name a suitable population from which to choose the controls for your study.
Controls are similar but not too similar to cases, but without the outcome of interest. Controls
can be drivers from similar areas with comparable characteristics and random samples from
motor vehicle license records. The study base would entail everyone who was an active driver
at the time of the accident.
e) What major characteristic must you strive to measure similarly in each study participant (both
the cases and controls)?
The hours they sleep at night, safety features of the car, road conditions, weather, drivers age,
sex, medical history, passenger restraint systems in use.
f) What difficulties could be encountered in measuring these characteristics?
The hours they sleep at night, safety features of the car, road conditions, weather, drivers age,
sex, medical history, passenger restraint systems in use.
2
, Work group – questions and answers – Daniëlle Band
Assignment 3
What is the importance of randomization in an intervention trial, and what does it accomplish?
A randomized controlled trial causes data that is free of bias. It makes it more likely that any differences
in outcomes between groups are a result of the treatment they received. Instead of differences in
outcomes being influenced by other factors like characteristics (age, sex)
Assignment 4
What is meant by the phrase ‘ecological fallacy’?
Ecological fallacy occurs when you make assumptions about individuals based on group-level data. In
other words, it's a mistake to think that characteristics observed in a group apply to every individual
within that group.
Assignment 5
Different study designs have particular advantages and disadvantages. Contrast the casecontrol and
prospective cohort design with respect to the following factors.
a) Costs;
Case-control is inexpensive and prospective cohort study is more expensive
a) Time required for completion of the study;
Case-control can yield important findings in a short time, cohort studies are generally more
time consuming
b) Bias;
Case-control studies are very susceptible for both selection and information bias (participant
cannot recall certain events, who do we select as controls?)
c) What can be estimated (with regards to effect/association measures);
Cohort studies allow for exploration of natural developments between exposure and
outcome, whereas with case-control studies you retrospectively assume two things to be
associated.
3
