Tools of the epidemiologist

This review provides a summary of epidemiologic tools to facilitate understanding of the design and analysis of studies of Alzheimer disease (AD) and related disorders. Proportions, ratios, rates, prevalence, incidence, study designs, bias, confounding, effect modification, odds and risk ratios, statistical power, and confidence intervals are defined and discussed. Descriptive epidemiology is concerned with describing the distribution of disease by person, place, and time. It is useful for hypothesis generation, but not generally for hypothesis testing. Observational analytic epidemiology focuses on identifying putative causes for an illness. Although its primary mission is hypothesis testing, it can lead to new hypotheses as well. Finally, experimental analytic epidemiology or clinical trials can provide rigorous tests of presumed causal associations. The strengths and limitations of various designs as they apply to determining causal associations in studies of AD and dementia are reviewed. Over the past 60 years, the epidemiologic study of dementia has evolved from basic descriptive studies of prevalence and incidence to case-control and cohort studies and finally to the first clinical trials to prevent AD.