Observational Studies in CVD

Question 1

What are descriptive studies?

Answer 1

• describe a particular risk factor or disease
• e.g. observational studies: cases, ecological, cross-sectional

Question 2

What are analytical studies?

Answer 2

• study cause and effect
• e.g. observational: case-control, cohort; interventional: clinical trials

Question 3

Non-longitudinal studies

Answer 3

Measures at one point in time with no follow up

e.g. case, ecologcial, cross-sectional, case-control

Question 4

Longitudinal studies

Answer 4

Have a follow-up after a period of time

e.g. cohort, clinical trials

Question 5

How is data collected for CS studies?

Answer 5

• questionnaires
• examinations
• investigations
• produce mostly descriptive outputs related to prevalence

Question 6

What is the limitation of a CS study?

Answer 6

• associations among variables, not temporal relationships
• weak evidence of causality
• hypothesis generating rather than supporting
• cannot determine cause and effect

Question 7

What is an example of a CS study?

Answer 7

AusDiab: The Australian Diabetes, Obesity and Lifestyle Study

2000, 11000 people, CHD, diabetes and other diseases

prevalence data

Question 8

Case control studies compare

Answer 8

Previous exposure status to a particular risk factor between cases and controls

Question 9

In retrospective study designs

Answer 9

• cohort or case-control studies
• outcome of interest has already ocurred at initiation of the study
• looking to the past to identify a commong risk factor

Question 10

What are the benefits of case-control studies?

Answer 10

• explicit knowledge of temporal relationships between exposure and outcome
• useful for rare outcomes

Question 11

What are the drawbacks of case-control studies?

Answer 11

• Cannot investigate the true temporal relationsjip between the exposure and the outcome
• cannot infer incidence

Question 12

What is an odds ratio?

Answer 12

• approximation of relative risk of outcome conferred by exposure
• output of case-control studies

Question 13

What is the key output of a case-control study?

Answer 13

The odds ratio: an approximation of relative risk

Question 14

How is the odds ratio calculated?

Answer 14

OR= [odds of exposure to non-exposure among cases]/[odds of exposure to non-exposure among controls]

Question 15

What is an example of a case-control study?

Answer 15

The INTERHEART study

• effect of modifiable risk factors associated with MI differs across coutnries
• e.g. smokers have 2.87x risk of MI

Question 16

What are the hallmarks of a cohort study?

Answer 16

• longitudinal (w/follow-up)
• incidence data
• comparison of outcomes between/subgroups (e.g. exposed vs non-exposed)
• derive relative risk

Question 17

What is a prospective cohort study?

Answer 17

• people classified by exposure (risk factor) prior to outomce
• allows knowledge of temporal relationship between exposure and outcome

Question 18

What is a retrospective cohort study?

Answer 18

• use data already available for a cohort and the cohort has developed the outcomes
• more cost-effective
• e.g. registered patient data at a heart failure clinic

Question 19

What are the advantages to a cohort study?

Answer 19

• Study multiple exposures and outcomes can be assessed
• Can research hypotheses post-hoc (i.e. retrospective)

Question 20

What are the disadvantages of cohort studies?

Answer 20

• difficult for rare outcomes
• not cheap or easy

Question 21

Example of a cohort study?

Answer 21

Framingham Heart Study

• 5000 people, 1948
• est. risk factors for CVD (lipids, somking, BP, etc.)
• incidence of CVD, coronary HD, and stroke
• recruitment of offspring continued the study

Question 22

What is active outcome ascertainment?

Passive?

Answer 22

Physical validation of an outcome by assessment (e.g. hospital visit)

Obtained from databases of hospital, clinic visits etc.

Question 23

What is bias?

Answer 23

Unintentional error that causes a systematic difference between or among groups

Question 24

What is selection bias type I?

Answer 24

systematic difference in people selected vs. not selected

• ‘worried well’ are more likely to participate but are usually more active and motivated in some way
• results are tf not generalizable

Question 25

What is selection bias type II?

Answer 25

• Systematic differences in subjects within groups being compared
• may attribute to differences observed in results rather than exposure
• e.g. recruiting all cases from the hospital and all controls from outside

Question 26

What is information bias?

Answer 26

• systematic differences in the way information is collected between/among groups being compared
• differences (partly) responsible for results
• e.g. BP cuffs on healthy and obese pt - cuffs are tighter on obese and tf influence BP measurements

Question 27

Recall bias is

Answer 27

• a type of information bias
• people with the condition of interest are more likely to remember exposure to a risk factor

Question 28

What is confounding?

Answer 28

• mistaken attribution for the relationship between two variables
• a confounder can independently alter the outcome
• it can be related to the exposure
• e.g. age, sex

Question 29

How can confounding be minimized?

Answer 29

• match by confounders (e.g. age, sex)
• restriction criterea (e.g. only males)
• stratification analysis (by sub-groups)
• multivariate analyses