Cohort studies.

Question 1

What is the best type of observational study?

Answer 1

A cohort study.

Question 2

What simple exposure category is often used in cohort studies?

Answer 2

Yes/No.

Question 3

Why is the final population in a cohort study slightly smaller than the original population?

Answer 3

Death/ loss to follow up.

Question 4

When are cohort studies particularly useful?

Answer 4

When the time period between the exposure and the disease is relatively short.

Question 5

Why can it be useful to use occupational populations (such as the Navy) as target populations?

Answer 5

They tend to be more compliant.

Question 6

Name 6 types of target populations used in a cohort study.

Answer 6

1. Population based on geographical region.
2. Population based on a certain occupation.
3. Population based on disease status.
4. Population based on risk group.
5. Population based on an exposure event.
6. Population based on a birth cohort (eg those born in a certain year).

Question 7

If a cohort study is small enough what can be used to encourage people to comply to follow up?

Answer 7

Payment.

Question 8

What 4 methods can be used to assist with follow up?

Answer 8

1. Get individuals to supply a next of kin.
2. Use the national postal service for address updates.
3. Social security records.
4. Motor vehicle records.

Question 9

Cut points used when categorising continuous data always need to have a biological significance. True or false?

Answer 9

False. They can be arbitrary.

Question 10

If an outcome is subjective what do you need to do when carrying out a cohort study?

Answer 10

Define this outcome upfront.

Question 11

What two types of outcomes can be measured in a cohort study?

Answer 11

1. Single/ discrete events

2. Multiple occurrence events.

Question 12

When looking at the outcome of a discrete event in a cohort study do you consider disease subtypes?

Answer 12

No.

Question 13

Name two examples of discrete/single outcomes that can be looked at in a cohort study.

Answer 13

1. Mortality.

2. First occurrence of a disease.

Question 14

What transition is looked at in a cohort study when the outcome occurs multiple times?

Answer 14

The transition from healthy status to disease status.

Question 15

What needs to be uniform in exposed and non exposed individuals in a cohort study to reduce bias?

Answer 15

Measures for detecting outcome.

Question 16

Is it easier to ‘clean’ a case control or a cohort study?

Answer 16

Case control as can do at the time of collection.

Question 17

What did the NIH-AARP diet and health care study investigate?

Answer 17

Cancer incidence in 500,000 50-70 year olds. Produced 53,000 incident cases.

Question 18

How many publications have come from EPIC?

Answer 18

100+

Question 19

What case control ratio was used in EPIC?

Answer 19

2:1.

Question 20

Smaller studies can have less power. What is an advantage of using a smaller study?

Answer 20

Repeats of measurements can be taken.

Question 21

If a cohort study is both retrospective and prospective what is it called?

Answer 21

Ambidirectional.

Question 22

An advantage of cohort studies is the exposure is measured for the reason you want to in the way you want it to be. True or false?

Answer 22

False. This is not the case with retrospective cohort studies.

Question 23

Retrospective studies are much quicker to carry out than prospective studies meaning they are more often used. True or false?

Answer 23

False. They are quite rare.

Question 24

Future follow up can occur in a retrospective cohort studies although this is rare. True or false?

Answer 24

False. If this does happen this is an ambidirectional cohort study.

Question 25

Why are retrospective cohort studies rare?

Answer 25

As the data is not collected for their specific purpose meaning it can be difficult to use.

Question 26

Exclusion criteria of who to include in recruitment and analysis of cohort studies should be stated upfront. Name reasons why individuals may be excluded (5 things).

Answer 26

1. No histology recorded. 2. Insufficient data at endpoint. 3. Prevalence of a different disease at baseline (still recorded and used to investigate questions not related to that specific disease). 4. Factors which mean they can not be 'at risk' from your disease. (still recorded and used to investigate questions not related to that specific disease). 5. Questionable data.

Question 27

Do reference groups used in cohort studies tend to be internal or external?

Answer 27

Internal.

Question 28

When would an external reference group be used in a cohort study?

Answer 28

When everyone in the cohort is exposed. The outcome being measured also has to be present in the external reference group.

Question 29

When external references are used what is it hard to do?

Answer 29

Adjust for confounding.

Question 30

Are external reference groups more commonly used in prospective or retrospective cohort studies?

Answer 30

Retrospective.

Question 31

What is the equation for the 95% confidence interval for RR?

Answer 31

e^(lnRR+-(1.96x (B/A(A+B) + D/C(C+D)^1/2))

Question 32

What is the equation for attrituble risk?

Answer 32

Incidence in exposed- Incidence in unexposed.

Question 33

What sort of regression can adjust for potential confounding in cohort studies?

Answer 33

Multiple logistic regression.

Question 34

When will an individual contribute to a survival analysis calculation?

Answer 34

When they still have the potential to become a case.

Question 35

What sort of regression is used in survial analysis?

Answer 35

Cox proportional hazards regression- time to event used as the dependant variable.

Question 36

What is information bias?

Answer 36

When different quality and/or quantity of information is collected from the exposed/non exposed group.

Question 37

What would you call this type of bias? ' Misclassification of exposure or disease status.'

Answer 37

Misclassification bias.

Question 38

What would you call this type of bias? 'Individuals who partake in trials tend to be more affluent and health conscious than the average person from your population meaning they can be unrepresentative of the population as a whole'.

Answer 38

Healthy worker effect selection bias.

Question 39

What sort of validity is affected if the healthy worker effect bias is present?

Answer 39

External validity effected (Internal is not).

Question 40

Are cohort studies good for rare diseases?

Answer 40

No.

Question 41

Cohort studies should not be used if the exposure changes over time. True or false?

Answer 41

Yes, they can be. This is a disadvantage of cohort studies.

Question 42

Can both survival curves and hazard ratios be calculated from cohort studies?

Answer 42

Yes.

Question 43

What is an advantage of using observational studies?

Answer 43

They do not have the same ethical concern as analytical studies.

Question 44

Can you use case control or cohort studies be used to prepare for RCT's?

Answer 44

Cohort.

Question 45

What sort of bias is reduced in cohort studies compared to case control studies?

Answer 45

Survival bias.

Question 46

What is guaranteed in cohort studies (and not in case control)?

Answer 46

Exposure occurring before disease status.