Lecture 3: Observation studies in CVD

Question 1

Define a cross sectional study

Answer 1

• a point in time of studying a particular population or pattern of disease/risk factors
• study can go for years but each n contributes only once
• no follow ups
• data collected as questionnaires, examination, investigation
• descriptive prevalence outputs

Question 2

Pros and cons of cross sectional study

Answer 2

Pro: cheap, easy to run, can explore associations

Cons: no temporal relationship, no cause and effect (weak causality)

Question 3

Define case control studies

Answer 3

comparison of PREVIOUS exposure status between:
– subjects with outcome of interest (cases)
– subjects without outcome of interest (controls)

• controls are matched with cases, 1:1 or n:1
• matching by confounders - eg: age, sex

Question 4

Define a cohort study

Answer 4

• longitudinal, with follow-up of subjects
• collect incidence data
• comparison of outcomes between/among subgroups eg, not exposed vs exposed to risk factor
• derive relative risks

Question 5

Pro and con of cohort study

Answer 5

pro: cohorts can be established as part of routine clinical care (e.g. stroke service at RMH)
con: expensive, not easy, difficult to study rare outcomes

Question 6

Define selection bias

Answer 6

systematic difference in characteristics of people selected for study and those not selected - specifically, people whose data were used for analysis and people whose data were not

observed result may not reflect the true situation, and/or may not be generalisable

ALSO
systematic difference(s) in characteristics of subjects within groups being compared

these DIFFERENCES are (partly) responsible for the observed study results

Question 7

Define information (measurement) bias

Answer 7

systematic difference(s) in the way information is collected between/among groups being compared

• differences are (partly) responsible for the observed study results
• arises when there is variability (especially subjectivity) in methods for collecting information

Question 8

How would you minimise selection bias?

Answer 8

• careful recruitment (rep sample, case and controls from same source)
• maximise response
• minimise lost to follow up

Question 9

How would you minimise information bias?

Answer 9

• standardised method of data collection (uniform between/among groups)
• objective assessment

Question 10

What is a confounder?

Answer 10

• confounding effects on exposure and outcome

Question 11

What are examples of universal confounders?

Answer 11

• gender

- age

Question 12

How would you minimise confounders?

Answer 12

By design

• matching by confounder
• restriction

By analysis
- restriction
stratification
multivariate (regression) analyses to remove confounders and look at true relationships