Also called follow-up studies, incidence studies, longitudinal studies, etc.
Can be prospective or retrospective
Prospective cohort study – healthy people with and without exposure to a factor is found and followed up until disease
At the point of the study only the exposure, not the disease, has occurred
Example: People who smoke and people who don’t smoke are chosen. They’re followed up until disease develops.
Retrospective cohort study – a population with a disease is found. Historical data is used to separate the population into those who had the exposure and those who didn’t
At the point of the study both the exposure and the disease has occurred
Example: People with lung cancer are chosen. Their previous exposure to second-hand smoke is determined.
Retrospective cohort studies are less important than prospective ones
Assessment of exposure can be done in multiple ways
By checking existing data, like medical records
By giving questionnaires
By measuring
Prospective studies require an initially disease-free population
The population must be screened for the disease
Pros
Good for rare exposures
Give the strongest evidence for causation
Give incidence rates, relative risk (RR) and odds radio (OR)
Cons
Not good for rare diseases
Rxpensive
Time-consuming
Needs large sample size
Comparison between retrospective cohort study and case control study
In both, both the exposure and the disease have already occurred
In a retrospective cohort study two similar populations are found, one exposed and one non-exposed. Both populations have the disease in question. The exposed and non-exposed group are compared, and the exposure’s effect on the outcome is determined.
The RR and OR are calculated
Only diseased people are examined
In a case control study two similar populations are found, one with the disease and one without. Both populations are examined for exposure to the factor. The exposure’s effect on the outcome is determined.
Only the OR can be calculated
Both diseased (cases) and non-diseased (control) people are examined
