Clinical Epidemiology by Diederick E. Grobbee Arno W. Hoes

Author:Diederick E. Grobbee,Arno W. Hoes
Language: eng
Format: epub
Publisher: Jones & Bartlett Publishers, Inc.
Published: 2014-04-10T04:00:00+00:00

TAXONOMY OF EPIDEMIOLOGIC DATA COLLECTION

Like many young scientific disciplines, epidemiology suffers from the use of confusing and inconsistent terminology. Many epidemiologists use the same wording to describe different studies or use different words for the same research approach. Particularly problematic is the naming of studies by words that seem to have a qualitative implication. As indicated in this chapter, by itself the word observational is a clean term that applies to any form of empirical research. Too often it is used to suggest a limitation of the research.

The word descriptive has a similar history of misuse. In several textbooks, a distinction between analytic and descriptive research is made, where descriptive studies supposedly do not provide definitive answers. We use the term descriptive as contrasted with causal to indicate whether the determinant–outcome relationship under study is meant to explain causality or is only meant to describe the strength of the association.

All research is analytic by nature. In our view, epidemiologic studies should be classified according to three dimensions: (1) time, referring to the time (zero or > 0) between measurement of the determinant and the outcome as well as to the prospective or retrospective nature of the data collection; (2) census or sampling; and (3) experimental or nonexperimental. We recommend that you use all three elements in the nomenclature in texts describing the nature of data collection. This removes the need to rely on vague, suggestive, and noninformative jargon such as retrospective study, prospective study, survey, follow-up study, and the like. Note that a prospective study is a study in which the data are collected after the researchers decided to address a specific research question and the term can thus refer to many types of data collection, including a cohort study, case-control study, cross-sectional study, or randomized trial. Also, the meaning of the term longitudinal study is unclear. All studies, except diagnostic studies, address longitudinal associations.

Thus, the characteristics of the main approaches to data collection in clinical epidemiology can be summarized as follows (see Table 7–1):

• A cohort study has a time dimension greater than zero; analyses are based on a census of all subjects in the study population, and the data collection can be conducted prospectively or retrospectively. The study can be observational or experimental, but if it is experimental it usually takes the form of a randomized trial.

• A dynamic population study has a time dimension greater than zero; analyses are based on a census of all subjects in the study population for the time they are members of the population, and the data collection can be conducted prospectively or retrospectively. Such studies are typically nonexperimental (i.e., observational). Because the term dynamic population study is hardly ever applied in the literature, we use the term cohort study to indicate both studies involving dynamic populations and cohorts.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.