|
With
these methodologic perspectives as background, we now provide
an overview of different epidemiologic study designs.
Lilienfeld
and Lilienfeld (1980) classified epidemiologic study designs
as:
- Observational
(relying on naturally occurring variation in risk factors) or
experimental (introducing exposure to protective factors
or treatments by design);
- Individual
(the person is the unit of observation) or ecological
(a social group, community, state, or nation is the unit of
analysis); and
- Relying
on cross-sectional (or retrospective) data collection
or longitudinal data collection.
Observational
studies may be designed to draw samples from the total population
without respect to disease or risk factor status; to compare cases
and controls; or to compare persons exposed to a risk factor to
persons not exposed. Experimental studies introduce, on a randomized
basis, a protective factor to persons at risk of disease (a preventive
trial) or treatment to persons with disease (a clinical trial).
A community
trial is a preventive trial conducted on an ecological basis in
which a protective factor is introduced to all persons in a set
of social units such as worksites or towns, with comparable social
units serving as controls.
In the following
sections we discuss each of the variations on individual level
study design. Because ecological studies (e.g. community trials
or analyses of morbidity data at the level of the social group,
community or other aggregation) are rarely used in TMD, we will
not review ecological studies here. Morgenstern
(1982) has reviewed the use of ecologic analysis in epidemiologic
research, while Von
Korff et al. (1992a) have considered conceptual issues in
epidemiologic inference from ecologic data. For each type of study
design, we provide examples from the epidemiologic literature
on TMD or other chronic/recurrent pain conditions.
|