19. Cohort Studies

Question 1

what is a cohort?

Answer 1

a group of people that share a similar characteristic

Question 2

biases in cohort studies

Answer 2

If there are difference in how participants are treated based on exposure status, bias will be introduced
§ Information: quality of information is different
§ Non-response: follow up is different
§ Assessment: ascertainment of outcome varies
§ Analytical: preconceptions of data analysts
§ Selection Bias is inherent in observational studies and is the major disadvantage
□ Systematic error due to difference in characteristics between those selected for a study and those not selected
® Eg: healthy volunteer; hospital controls

Question 3

characteristics of cohort study

Answer 3

• Cohort is defined
  ○ Choose target population
  ○ Choose appropriate sample population
• Exposure status is determined and defined
  ○ Do they have certain risk factors?
• Followed prospectively, over time
• Identify and define development of disease/outcome
  ○ Monitor patients for development of outcome
• Compare incidence of disease in two groups

Question 4

how do you identify exposed and unexposed?

Answer 4

1. Select defined population (cohort) before exposures are identified
2. Create a study population by selecting participants based on exposure

Question 5

Realtive risk

Answer 5

use in cohort studies: incidence of outcome in exposed/incidence of outcome in unexposed

incidence in exposed = a/(a+b)
incidence in unexposed = c / (c+d)

a = + outcome, + exposure
b = - outcome, + exposure
c = + outcome, - exposure
d = - outcome, - exposure

Question 6

attributable risk

Answer 6

what proportion of the outcome is caused by the exposure under investigation?

AR = (incidence Ex - Incidence UnEx)/ Incidence Ex

Question 7

advantages/disadvantages of cohort study (compared to case-control study)

Answer 7

Advantages:
- Study design (exposure -> outcome)
○ Can determine temporal sequence
○ Good for identifying natural history of disease
- Good for assessing multiple outcomes after a single exposure
- Can identify incidence and relative risk
○ Case-control, odds ratio only approximates RR if outcome is rare
- Reduces or eliminates recall bias (unless retrospective cohort study)
Disadvantages
- Can be inefficient and costly
○ Rare disease
○ Diseases w/long natural history
- Attrition/loss to follow up
○ If exposed have different rates of attrition than unexposed = BIAS
- Selection bias**
○ Systematic error due to difference in characteristics b/w those selected for a study vs those not selected

Eg healthy volunteer, hospital controls