41 - Observational Studies in CVD Flashcards

1
Q

Cross-sectional studies

A

Sample of population, taken at one point in time (non-longitudinal)
Data taken from questionnaires, examinations, investigations, etc.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Drawbacks of cross-sectional studies

A

Can’t really prove causation, as have no temporal component.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Advantages of cross-sectional studies

A

Cheap, fast, relatively easy

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Case-control studies

A

Compare previous exposure status to risk factor between those known to have outcome (cases), and those without outcome (controls).
Non-longitudinal.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Steps in case-control studies
1)
2)

A

1) Match cases and controls according to confounders (age, sex, etc).
2) Determine previous exposure among subjects

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Study type that is useful for studying rare outcomes

A

Case-control studies

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Why are case-control studies useful for studying rare outcomes?

A

Can select cases who have rare condition, instead of recruiting a large enough group of people to study so that one of them might develop condition

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Key output of case-control studies

A

Odds ratio (approximation of relative risk)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Why is the key output of case-control studies an odds ratio, and not a relative risk?

A

Case-control studies are non-longitudinal. Relative risk requires longitudinal data to generate (relative risk is rate exposed/rate unexposed, and rate is the probability of disease occurring in an undiseased population over cumulative time of study).

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

How is odds ratio calculated?

A

[odds of exposure to non-exposure among cases] ÷

[odds of exposure to non-exposure among controls]

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Cohort studies

A

Longitudinal studies.
Compare outcomes among subgroups (EG: disease in exposed/unexposed).
Collect incidence data

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Data output of cohort studies

A

Relative risk

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Difference between prospective and retrospective cohort study

A

Prospective is where subjects are recruited before outcome has occurred.
Retrospective is where subjects are recruited after outcome has occurred.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Advantages of cohort studies

A

explicit (often-detailed) knowledge about temporal
relationship between exposure and outcome
• can include multiple exposures and outcomes
• research hypotheses can be addressed post hoc in
established cohorts

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Disadvantages of cohort studies

A

1) Hard to study rare outcomes

2) Difficult to carry out, expensive

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Study that gave data for interactions between CVD risk factors

A

Framingham study

17
Q

Bias

A

Unintentional error leading to systematic difference between or among groups

18
Q

Two main types of bias

A

Selection bias

Information bias

19
Q

Selection bias

A

Systematic difference in people selected for study (EG: type of people willing to enter study might be unrepresentative).

20
Q

Example of selection bias

A

Difference between subjects who stay in Framingham study, and those who drop out

21
Q

Ways to minimise selection bias
1)
2)
3)

A

1) Careful recruitment (cases and controls from the same place)
2) Maximise results.
3) Careful follow up

22
Q

Information bias

A

Systematic differences in way that data is gathered between or among groups.
Particularly arises when there is variability (subjectivity) in how data are collected.

23
Q

Example of information bias

A

Differential in how CVD was measured between males and females in Framingham study

24
Q

Universal confounders

A

Age, sex