Epidemiology: Study Types
Cohort Study
Analytic Observational Study -multiple observations at diff. timepoints (follow-up with participants) -Prospective: predicting if people will get disease -Primary data (some secondary) Applications: study chronic disease etiology, estimate risk, study multiple outcomes, know directionality of disease! RELATIVE RISK
Ecologic Study
Analytic Observational Study -one observation at one time point -secondary data sources used: census data, hospital records Applications: figure out the cause of a disease -Can't infer directionality! -Ecologic Fallacy!!! (implied causal relationships that are incorrect) CORRELATION/CHI-SQUARED
Case-Control Study
Analytic Observational Study -single point of observation -Retrospective: people already have a disease, go back and ask about exposure -primary and secondary data -individual observation Applications: chronic diseases w/ unknown etiology, infectious disease outbreaks -no incidence data! ODDS RATIO
Cross-Sectional Study
Descriptive Observational (Prevalence) Study -one observation of the whole population at one time point -simultaneous selection, assignment, and assessment of research subjects -primary data sources: survey, questionnaires Applications: describe disease distributions, magnitude of problem, and trends -Can't infer directionality! PREVALENCE ESTIMATES
Clinical Trial
Experimental Study -evaluate drug applicability (or a program potentially) -individually focused -RANDOMIZATION -MANIPULATION -can last days-years -highly selected group of individuals Outcomes: measure rates of disease, death or recovery
Community Trial
Experimental Study -test the benefit of new educational program or policy -community focused NO randomization -MANIPULATION -longer duration than clinical trials -targets all members of a community
Nested Case-Control Study
A case-control study where cases and controls are drawn from the population in a cohort study controls=subset of cohort without disease cases=subset of cohort with disease -proved some control over confounding factors -reduce cost because exposure information is collected from only a subset of the cohort