Nested studies. Flashcards
What is the main advantage of using nested studies?
They are effective ways to utilise resources.
What is an advantage of using a nested case control study over a normal case control study?
Exposure is assessed prior to the outcome.
Nested case control studies reduce the wasting of resources. What is another advantage of them?
Reduce cost in data entry.
Giving people the option to right written answers in questionaries can provide useful, detailed information. What is a disadvantage of these questions?
Free text fields are very difficult to code.
Ideally controls would be selected from the study population. If this is not the case and an external group must be used what must you ensure is the case?
That the external group should have the same amount of exposure to the cases. This is rarely plausible.
Should controls be selected independently or dependently of their exposure status?
Independently.
What rate applied to controls should not differ dependant on exposure status?
The sampling rate.
What three types of control sample strategies exist for nested controls?
- Cumulative sampling.
- Incidence density sampling.
- Case cohort/ baseline.
Which of the following are sampling methods used for nested studies?
- Cumulative sampling.
- Incidence density sampling.
- Case cohort/ baseline.
- Cumulative sampling.
2. Incidence density sampling.
What is another name for incidence density sampling?
Risk set sampling.
Describe cumulative sampling.
Cases are identified and controls are matched from survivors accordingly at the end of followup/ when you decide to do the study.
Describe incidence density sampling.
Controls are selected at the same time that cases are diagnosed.
Describe case cohort studies.
Non cases are a sub cohort of the original cohort.
Which of the following sampling techniques can result in the same individual being classified as both a case and a control?
- Cumulative sampling.
- Incidence density sampling.
- Case cohort.
Incidence density sampling.
Case cohort.
What bias can be introduced in cumulative sampling and not in incidence density sampling?
‘Super healthy’ individuals have been selected from the ordinal cohort.
Which of the following method of control selection is used for assessing multiple outcomes?
- Cumulative sampling.
- Incidence density sampling.
- Case cohort.
Case cohort.
Which of the following method of control selection allows cases to be matched to any number of controls?
- Cumulative sampling.
- Incidence density sampling.
- Case cohort.
All of them.
Control selection for case cohort studies should happen at the start of the study. True or false?
False, it can happen at any time. All individuals at the start of the study should however be considered for selection.
In terms of matching for age what is an advantage of using a large cohort?
You have more chance of having sufficient data to make narrower age categories.
What is the relationship between effect size and sample size?
The large the effect size the smaller the sample needed to detect an effect.
Roughly how much cheaper is a nested case control study compared to analysing the whole cohort?
More than 99% cheaper.
What is the issue of drawing blood samples for case control studies?
The blood sample may change after the development of a disease.
What sort of ratio is calculated form cumulative density sampling?
Odds ratio (like normal case control).
What sort of ratio is calculated form incidence density sampling?
Rate ratio.
What sort of ratio is calculated form case cohort sampling?
Risk ratio.
Case control studies are good for rare outcomes. Nested studies are not. True or false?
False, they are.
Why are nested case-control studies flexible?
As you can do them at any time.
Why does a nested case control study reduce bias?
Selection bias is reduced as the same population is used for cases and controls/ cases and non cases.
Case cohort studies are good at excluding what sort of cases?
Cases that are logistically difficult to exclude as they have a large subclinical phase. To exclude an individual they need to be screened (which is easier when you are only using a subset for the control group).
‘Case cohort studies are always ideal for testing multiple outcomes from a cohort as they are cheaper and waste less resources.’ Why is this statement false?
As excluding cases reduces statistical efficacy. If it is possible to use the whole cohort you should do so.
