Glossary of Epidemiology: 13) Case-Control Studies: Design

Start Studying! Add Cards ↓

2 reasons case-conrtol trials are susceptible to bias
1) Extra step to select controls. 2) Both E and D are known and can influence both selection of subjects information accuracy
3 conditions necessary for valid case-control study
1) Identification of the study base. 2) Controls are random sample of study base. 3) Accurate nondifferential info on exposure and confounders.
Identification of the study base
1) All cases. 2) Random sample of cases if large number
6 steps in a case-control study
1) Identify Study Base. 2)Identify cases. 3) Identify controls. 4) Exposure. 5) Confounders. 6)Analysis
3 keys for Disease: Cases
1)Aspire for high Sens and Spec. 2) Restrict to newly developed disease (or dealing with survival). 3)Try and get ALL CASES (avoids selection bias)
random samples from study base
Exposure: issues for case-control
subjects enter study after disease occurrence. Pre-disease data needed.
OR and RR estimates
Study Base
Underlying cohort and time-period where all cases are identified and controls are randomly samples. AKA - source cohort, reference population
What IS NOT they study base
Cases + Controls
2 types of study base
1) Primary. 2) Secondary
Primary Study Base
Explicit cohort where all cases and controls can be identified. Only info needed for idenficiation is: person and dates
Nested case-control designs
Random sample of controls from a primary study base. Could have been a cohort study at greater cost.
EG. of primary study base
1) Study conducted within an existing cohort or rct population. 2) Study conducted with medical care DB where a retro-cohort could be done
EG. of NOT primary study base
1) Population where cases are ascertained is ambiguously defined. 2) Well dfined but rosters do not define study base
Seconday Study base
Study base identified conceptually but no definitive roster. Selcting controls approximate random sampling.
Danger of using a secondary study base
Sampling is not random and introduces bias
5 reasons to use a case-control study
1) When cohort study not feasible (expense, time). 2)Brief transient exposures. 3) Outbreaks. 4)Primary study base is present. 5) Seconday study base approximates random sampling.

Add Cards

You must Login or Register to add cards