What is a cohort study?
Comparisons of groups of a population who were a) exposed b) not exposed to a certain thing
e.g those exposed to radiation after Hiroshima/Nagasaki atomic bombings
In a cohort study, you start with a population with different incidences of a thing and work (forwards / backwards) to figure out what caused this difference.
In a cohort study, you can either
a) start with a population and follow them going forwards
b) go back and examine differences in a population
What are these types of study called?
Prospective cohort study
Retrospective cohort study
A cohort study allows you to estimate the ___ of something occuring in a population.
In a cohort study, the population you have needs to be a ___ population.
otherwise you don't know what's typical/rare
What size of population do you need to pick up rare diseases in a cohort study?
Large sample size
What is relative risk?
Relative risk = Probability of event (incidence) when exposed / Probability of event (incidence) when not exposed
What is relative risk also known as?
When there is no difference between two populations, the relative risk is _.
If the risk ratio is < 1, the exposed population are (protected from / vulnerable to) the effect.
If the risk ratio is > 1, the exposed population is (protected from / vulnerable to) the effect.
What is absolute risk?
Absolute risk = Risk in exposed group - Risk in non-exposed group
What is a case control study?
Find two populations, one with the condition and one not
Take identical histories and compare them
Cohort studies tend to go ___ in time while case-control studies go ___.
Case control studies are good for studying ___ diseases.
What statistic is associated with case control studies?
How is an odds ratio calculated?
Odds ratio = Odds of event when exposed / Odds of event when not exposed
An odds ratio of < 1 means that the risk is (lower / higher) in the exposed group.
An odds ratio of > 1 means that the risk in the exposed group is (lower / higher) than the risk in the non-exposed group.
Can causation be proved by a case-control study?
histories rely on memory