Cohort Studies

Question 1

What is a COHORT?

Answer 1

A defined group of people with known exposures/personal characteristics

Question 2

How is a relative risk between two internal sub-cohorts calculated?

Answer 2

• Incidence rate ratio (IRR)
• IR (A)/IR (B)
• A is the exposed, B is the unexposed

Question 3

How is risk estimated within a cohort?

Answer 3

• Incidence Rate IR (observed cases/total population x time)

- Standardised Mortality Rate SMR (observed/expected x100)

Question 4

What is the difference between INTERNAL and EXTERNAL comparisons in cohort studies?

Answer 4

• INTERNAL is subdividing the cohort into two groups of different levels of exposure
• EXTERNAL is comparing the cohort group to an external reference population

Question 5

What problems may arise with internal comparisons?

Answer 5

• Sub-cohorts may be radically different sizes so may lead to requirement of very large studies
• May be subject to confounding e.g. Age, sex

Question 6

How would you calculate the 95% confidence interval for an SMR value?

Answer 6

• Lower = observed SMR / e.f.

- Upper = observed SMR x e.f.

Question 7

Identify the major problem with observing large cohorts over a long period of time

Answer 7

During the study period the cohort members age and age-specific morbidity/mortality rates in the reference population change

Question 8

Describe the relationship between sample size and error factor

Answer 8

As the number of observed events (O) increases, the error factor decreases

Question 9

Why is the error factor for internal comparisons larger than that of external comparisons?

Answer 9

Cohorts are subdivided into smaller groups - smaller sample sizes so larger error factors for each group

Question 10

Why does the error factor calculation of the SMR only involve the observed cases (O)?

Answer 10

• Uncertainty in the unexposed reference population is negligible due to the large numbers it is based on
• Therefore the only uncertainty in the SMR comes from the rate of the exposed (O)

Question 11

State 2 advantages of using external comparisons over internal comparisons

Answer 11

• Tend to have less random variation

- Useful when it is not possible to use sub-cohorts

Question 12

State 3 disadvantages of using external comparisons in a cohort study

Answer 12

• limited data pertaining to reference population e.g. only have mortality data
• SELECTION BIAS e.g. healthy worker effect
• Sometimes only binary comparisons are possible e.g. comparing exposed cohort to non-exposed reference population

Question 13

What are the advantages of internal comparison cohort studies?

Answer 13

• Can establish that exposures precede outcomes
• Can study rare exposures
• Collect more detailed information for DOSE RESPONSE (can measure a range of exposures) e.g. Comparing risk of smoking with lung cancer

Question 14

What are the disadvantages of internal comparison cohort studies?

Answer 14

• Large numbers required
• Study must be followed over a long period of time
• SURVIVOR BIAS - risk of high number of losses due to follow up

Question 15

What is meant by SURVIVOR BIAS?

Answer 15

• Risk of high number of losses due to follow ups over a long period of time
• Those who remain in the study may be more health conscious than those who leave study

Question 16

Describe the effect of the “healthy worker effect”

Answer 16

Selection bias when comparing occupational cohorts with reference population

• Occupational cohorts are on average more healthier than the general population
• SMR may be well below 100

Question 17

What are the main characteristics of cohort studies?

Answer 17

• COMPARATIVE
• LONGITUDINAL
• Identifies exposures and compares risk outcomes
• Start with disease free individuals and follow up over a period of time

Question 18

How do you calculate MORTALITY RATE? Explain whether or not this is a rate

Answer 18

• MR = no. of deaths / total population

- NOT A RATE as it is not per unit time