Observational Study Designs Flashcards
Unique Aspect of Case Control Study
Start w/ outcome, then look back to find exposure (make 2 groups, 1 w/ outcome 1 w/out, compare the previous exposures of the 2). So very good for rare diseases
CC Measure of Association (& why)
Odds Ratio: (ad)/(bc)
Because you have a case group and a similarly picked and separate control group, so a+b and c+d not reflective of the total population
Notable Point about Selecting Subjects in CC Studies
Both cases and controls need to come from the same source population
Biggest Problem in CC Studies
Selection bias
Recall Bias
Cases more likely to recall exposures
2 Types of Matched CC Study
Frequency and individual matching
Cohort Study
Evaluate sequentially the exposure then outcome over extended period of time (selected subjects begin study free of outcome)
Cohort Measure of Association (& why)
Relative risk, bc sample hopefully reflective of population
3 Cohort Benefits
Get true incidence
Temporal
Representative data of some pop
3 Cohort Comparison Groups
Internal comparison - divide single cohort into groups based on exposure
Separate control cohort - 2 different groups compared
Compare separate cohort w/ available population data
Diff b/w CC and Retrospective Cohort
R Cohort starts w/ exposure and then later determines outcome, so can get RR and true incidence among E/non-E pops
SMR
Observed incidence of mortality vs. that which is expected
Cohort Selection Bias
Created by who drops out IF they have special relationship to outcome (IV drug users dropping out of HIV study)
2 Major Cohort Biases
Misclassification & attrition (lost to followup can create selection)
Cohort Study Exposure/Outcome Strength/Weakness
Good for rare exposures, bad for rare outcomes
