Research Design Flashcards
What is the hierarchy of evidence?
Systematic reviews Randomized controlled trials Cohort studies Case Control Case Reports Expert opinion
What is the Hawthorne effect?
People changing behavior in a study in order to please people. Use placebo!
What is intent-to-teat analysis?
Analysis of results including those who dropped out of the study early
Makes study more generalizable
Can RCTs measure incidence?
Yes.
Incidence -> number of new cases of disease over a given period of time.
How do we determine risk?
Absolute risk = incidence
(# of people w/ disease / total # of people)
Relative risk or Risk ratio:
(Incidence in those exposed/incidence in those unexposed)
What are the differences in RR and OR? What are similarities?
Both estimate the risk of disease/outcome in exposed to unexposed samples.
OR is odds: odds in one group/(odds in another group)
RR: incidence in one group/incidence in another group
What makes Cohort and Case control studies observational?
We cannot manipulate the independent variables
Cannot control extraneous factors.
Can you determine incidence from Cohort studies?
Cohort is a group of people with something in common and are observed for a period of time to see what happens to them.
studying “predictor variables”
Yes, we can infer incidence from these studies.
How do cohort studies progress?
Prospectively or retrospectively
I.e. Cohort without disease —> (exposure / non-exposure to somerisk) —->was the disease/outcome developed?
Define absolute risk in terms of incidence:
They are equal.
What is RR in cohort studies?
RR=Incidence in exposed/incidence in unexposed
How is a case-control study produced?
Look backward to compare people with and w/o a condition.
We know we have those with x diz from a population and controls (w/o diz) from a population.
Look backward to see if an exposure to a risk factor had an impact.
We can try to estimate the relative risk from this
What are some instances when Cohort studies are particularly beneficial?
Diseases with long latency
Rare disease
Can you measure incidence in Case-control studies?
NO! You can estimate it.
Really just the odds that a diz. Group was exposed.
What is a type of bias specific to case-control studies?
Recall bias.
