Technology has markedly improved access to public
health statistics. These statistics are derived from data sets
which are collections of logically related data arranged in a
prescribed manner. Data may represent information collected at
the national, state, or local levels. Public health data sets
may be conveniently envisioned as falling into two broad categories.
One category includes counts of individual health-related events
or services. Counts are made of individuals who are provided particular
health services. These counts are normally geographically and
chronologically proscribed. For example, one collection of data
might focus on a population in the northeast United States between
1960 and 1980 while another might be limited to citizens of West
Virginia.
Specific events might include hospital emergency
room visits, visits to WIC clinics, deaths attributed to a specific
cause, and preventive services including cancer screenings or
immunizations. Such counts of events, once aggregated, are useful
in assessing general health needs and status, setting reimbursement
levels, determining eligibility, evaluating care and program coverage,
and penetration rates. However, because data collection is limited
to those who seek services, the results may or may not be representative
of the general population.
A second category of data sets describes populations
through the use of sampling techniques. Data collection systems
that create these data sets survey a subset of a reference
population. The reference population could be as broad as all
citizens of the United States or it may be more narrowly constrained.
Examples include many of the federal surveys of health status
and health behaviors and health services utilization. The sampling
techniques are used to identify an appropriate survey population.
Statistical reports including a mix of text, tables,
and figures from data sets are available from an increasing number
of federal, state, and local sources through a variety of electronic
modes, including the Internet.
