For DESCRIPTIVE EPIDEMIOLOGY we focus on measures of disease frequency and understanding what those methods are used. These measures come is 5 types of “flavors:” Count, proportion, ratio, rate, risk. In descriptive epidemiology, the commonly used measures are CUMULATIVE FREQUENCY (a proportion), INCIDENCE RATE (rate), and POINT OR PERIOD PREVALENCE (proportion). Students need to be able to calculate and interpret these measures.
Students also need to understand how to compare measure of disease frequency. To do this, the main tasks are to arrange disease frequency data into a two-by-two table (contingency tables), calculate absolute (risk difference and attributable proportion among the exposed) relative measures of comparison (risk ratio, odds ratio), interpret the absolute and relative measures of comparison, and describe what standardization is and its purpose.
For ANALYTIC EPIDEMIOLOGY students first need to distinguish among epidemiologic study designs. There are two broad categories: Observations and experimental. The observational studies include: cohort, case-control, cross-sectional, ecological. Experimental designs include true experimental designs (individual, community, preventive, therapeutic, parallel, crossover, simple, or factorial) or pseudo-experimental (individual or community). Additionally, students are expected to understand the strengths and limitations of each type of study.
Also as part of analytic epidemiology, students are expected to define BIAS in epidemiologic studies and provide examples of selection bias (control selection bias, self-selection bias, loss to follow-up.) The concept of information bias, including recall bias, interviewer bias, and differential and non-differential misclassification are important. Students should be able to describe ways that selection and information bias can be avoided and minimized.
The concept of CONFOUNDING is also important. Students should be able to define and provide examples of confounding; describe methods for assessing the presence of confounding; and describe methods for controlling confounding at both the design stage (randomization, restriction, matching) and the analysis stage (standardization, stratified analysis, matched analysis in case-control studies, and multivariate analysis).
EFFECT MODIFICATION is another concept important in analytic epidemiology. Effect modification is a change in the strength or magnitude of an association between an exposure and an outcome according to the level of a third variable (the effect modifier or the interaction term). Students need to be able to distinguish between an effect modifier and confounding.
CAUSATION in epidemiology – Hill’s criteria (Sir Austin Bradford Hill)
SCREENING in Public Health. Students need to be able to discuss what characteristics make a disease appropriate for screening. For screening, the concepts of sensitivity and specificity are important as are predictive value (predictive value positive and predictive value negative)