Class session 13 Case-control studies Epidemiology 503
Hierarchy of Study Designs Generate Hypotheses Establish Causality Good place to start investigating a research question Ecologic Studies Cross Sectional Studies Case Control Studies Cohort Studies Randomized Trials Gold standard for establishing causality
Randomized Trials and Cohort Studies Study Population Randomization or Self-Selected Exposure Exposed to factor Not exposed to factor Event No event Source Population
Limitations: RCTs and Cohort Studies Expensive Lengthy time of follow-up Inefficient for studying rare diseases
Case-Control Study Design Cases Controls Disease No Disease Exposed Unexposed Exposed Unexposed Controls should represent exposure in the population that gave rise to the cases
Selection of Cases and Controls Case Definition? Prevalent or incident? Where from? Hospitals Disease registries Community Controls Where from? General population controls Friend controls Hospital controls Matching? For selection of controls, weigh pros and cons: ease of identification vs ability to draw from same source population as cases
Measures of Association: CCS Can we get incidence measures from a case-control study? Cases Controls TOTAL Exposed a b a+b No c d c+d Can’t calculate risk ratio or rate ratio in a case-cohort study In cohort study, risk or rate ratio is: Incidence in exposed (a/(a+b))/incidence in unexposed (c/(c+d)) Ratio of disease to no disease is set by investigator
Odds Ratio: CCS OR = ad / bc Cases Controls Exposed a b No c d Odds of exposure among cases = a/c Odds of exposure among controls = b/d Cases Controls Exposed a b No c d OR = ad / bc Interpretation (OR=1.9) In study, the odds of exposure were1.9 times greater (or 90% greater) among those with disease (cases) than among those without disease (controls).
Cohort vs. Case Control Cohort Study Case Control Study Research Question Are persons exposed more likely to get disease vs. unexposed? Are persons with disease more likely to be exposed vs. persons without disease? Selection of Subjects With and without exposure of interest. With and without disease of interest. Interpretation of OR Those exposed to XYZ were 2.5 times more likely to develop disease vs. those not exposed to XYZ. Those with disease were 2.5 times more likely to be exposed to XYZ than those without disease. Strengths Good for rare exposures Can calculate incidence Can study multiple outcomes Good for rare diseases Good for long latency periods Can study multiple exposures Limitations Inefficient for rare diseases Expensive Cannot calculate incidence Inefficient for rare exposures