BOX 4-1
An Overview of Epidemiological Study Designs

Epidemiology can be defined as the study of the distributions, determinants, and outcomes of health and disease in populations. These populations may be geographically defined, but they may also be clinical or institutional groups. Full population definition usually requires further demographic characterization, at a minimum. The determinants or causes of health and disease states are often referred to as risk factors. Outcomes of interest span a wide spectrum, including, for example, mortality, symptom severity, disease progression, other new mental and physical conditions (morbidity), the costs of illness, disability and dysfunction, and satisfaction with medical care. Epidemiology may be observational, wherein the world is observed as it is, or experimental and interventional, wherein experimental or quasi-experimental interventions are invoked Within observational epidemiology, descriptive epidemiology generally refers to the quantitative occurrence of health and disease phenomena in characterized populations, often using rates over a specified period of time, and requiring a known population denominator. For example, a prevalence rate describes how frequently the event of interest, regardless of time of onset, occurs at any point within a specified time interval. An incidence rate is the rate of new events in the population occurring within that interval.

Epidemiology can also be analytical as well as descriptive. There are many analytical study designs to assess population phenomena; three are basic to assessing health and disease in populations. Cross-sectional assessments or surveys, at one defined point in time, can be used to determine prevalence rates, characterize those in whom the events of interest exist, and explore statistical associations among the study variables of interest.

The case-control, or retrospective, study takes many forms and is mainly intended to identify risk factors for diseases and conditions. In its simplest application, new cases of a disease are identified, a demographically similar control group is designated, and differences in prior “exposure” rates to the risk factors of interest are calculated. The cohort or longitudinal study in essence starts with a defined population in whom putative risk factor information has been well characterized, but in whom the disease or condition of interest has not yet occurred. The population is then followed for incident events of interest, which are then related to previously acquired risk factor information. Each type of study can be elaborated and each has strengths, weaknesses, and methodological nuances. Full descriptions of these epidemiological methods for both community and clinical settings can be found in standard textbooks of epidemiology and clinical epidemiology.


While clinical descriptions of elder mistreatment are present in historical texts, even now there have been few population-referent, geographically based studies of elder mistreatment occurrence in the modern, peer-reviewed literature. The latter third of the twentieth century saw the descrip-

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement