| |
Large nationally representative
surveys on health and health care can answer critical research
questions about health outcomes, quality of life, and quality
of care for specific conditions as well as for populations of
interest (e.g. elderly, minority, low income, disabled) (Federman
et al., 2001; Egede
et al., 2002; Schneider
et al., 2002). These surveys commonly contain items
on health and functional status, including symptoms such as shortness
of breath, chest discomfort, and difficulty walking. Analyses
of nationally representative surveys can be descriptive or analytic,
hypothesis-generating or hypothesis-testing. Studies conducted
using national surveys may be cross-sectional (Stafford
and Cyr, 1997), longitudinal (Saliba
et al., 2001; Porell
and Miltiades, 2001), or examine trends over time (Olfson
et al., 2002).
Advances in web technology
have made survey data increasingly available to researchers. Web
access to databases can allow casual users the freedom to explore
some actual data without the cost, paperwork, and lead time that
has been characteristic of data acquisition. Click here
for an example.
Advances in computer
technology, coupled with improvements in statistical software,
have made it possible to analyze large data sets with personal
computers using the sophisticated analytic techniques needed to
address complex survey designs and sampling methods. This chapter
provides an introduction to the secondary analysis of large nationally
representative survey databases primarily using two surveys, the
Medical Expenditure Panel Survey, conducted by the Agency for
Healthcare Research and Quality (AHRQ), and the Medicare Current
Beneficiary Study (MCBS), conducted by the Centers for Medicare
& Medicaid Services (CMS) for illustration. Selected examples
from additional surveys are also used to illustrate important
concepts.
|