Biostatistics-Observational Studies Flashcards
What measure of association is commonly used in a case control study?
Usually an odds ratio. This is because case control studies tend to not get as good of values for incidence.
What tests are used to determine statistical significance of means? Proportions?
Mean comparison = T-test, Analysis of Variance. Proportion comparison = Chi square, McNemar’s (paired data).
Why are hospital studies not always the best place to do a study?
External validity is compromised because your population is not a good representation of the outside community. Internal validity is compromised because of recall bias (sick people recall events leading to their condition better)
What type of study would minimize misclassification of people as controls?
Cohort. You don’t know whether their going to be a case or control at the beginning of the study.
How is confounding addressed?
Matching (attempts to compare people in the same scenario to eliminate confounders), Stratification (separate your measures of association into sub-groups to expose and separate confounders) and Adjustment.
What measure of association is used in a cohort study?
Risk ratio. This is because you get a good level of incidence.
What are standard biases in cohort studies?
Exposed and unexposed groups are not equally susceptible for reasons unrelated to the exposure, group crossover between subjects, drop-out and tendency to measure exposed group more closely than unexposed.
