Standardized protocols and laboratory procedures were used to conduct the unlinked surveys. Because only information that is routinely collected in the medical records or on intake forms could be used, limited information on demographic characteristics and behavioral risks was abstracted. Only clinics that participated in the surveys each year from 1993–1997 are included in the analyses presented here. When sufficient data were available, trends for each surveillance population were compared by region, sex, race/ethnicity, age group, and HIV risk behavior.

In earlier CDC prevalence reports, clinic data were presented as medians and ranges of HIV prevalence rates. For this report, however, rates are presented as means and, for the trends analyses, were standardized to the 1993 population of each survey type by region, sex, race/ethnicity, and age group in order to control for population changes over time. The standardized prevalence rates in this report summarize data from sites and demographic groups among which these rates may vary. Logistic regression was used to evaluate the variation in time trends across sites and demographic groups for MSM and heterosexual men and women tested in STD clinics and among IDUs tested in DTCs. (See Appendix I for details).

Both for the anonymous unlinked surveys and the routine HIV screening programs included in this report, all seroprevalence rates are expressed as the percentage of persons in each subgroup who had serologic evidence of HIV infection at the time of testing. The participating states and the District of Columbia are divided into four geographic regions (Northeast, Midwest, South, and West) as defined by the U.S. Bureau of the Census (Appendix I). For the regional analyses in this report, data from Puerto Rico are included with the South.

Each of the surveys and screening programs measured HIV seroprevalence, which is used interchangeably with prevalence in this report. Temporal changes in seroprevalence result from an imbalance between the influx and outflow of HIV-infected persons in the population under study. Although the representativeness of each sample population is unknown, collective data from the unlinked clinic surveys and from the screening programs of the Job Corps, the military, and the American Red Cross may indicate geographic and demographic patterns and trends of HIV prevalence.

Go to Summary of Findings