The data were collected as part of the Supplement to HIV and AIDS Surveillance (SHAS) project, a Centers for Disease Control and Prevention-funded project initiated in 13 state and county health departments in the United States in January, 1990 (9). SHAS is a populationbased survey in which all confirmed and living female AIDS cases reported in LAC are contacted to participate in a detailed assessment of behavioral and sociodemographic characteristics. In LAC, HIVinfected women who have not yet progressed to AIDS and who are receiving medical treatment at the largest public HIV outpatient clinic in the county are also included in SHAS. A confidential questionnaire is administered to all participants in person by trained interviewers.
   The goal of the analysis was to describe the sociodemographics and sexual and drug-using behaviors of African-American women with HIV and AIDS in LAC and compare the characteristics of this subgroup with those of women of other races with HIV and AIDS in LAC. Univariate statistics, stratified by race, are presented for purposes of the descnptive characterization. Odds ratios (OR) and 95% confidence intervals (CI) were computed to examine whether African-American women with HIV and AIDS were more likely than women of other races with HIV and AIDS to report certain behaviors and characteristics. Student's l-tests were conducted to compare the mean age of African-American women with that of women of other races. All analyses were conducted using SAS version 6.10 (10).
   Using an a priori list of confounders, we examined the association between each confounder, race, and behavioral or demographic characteristics. If strong associations were present between the confounder and race and the confounder and the characteristic, or prior information was present, the factor was considered a confounder and controlled for in the analysis.
   Data were analyzed both separately and in combination for HIVinfected women and for those with AIDS to evaluate the impact that the aggregation of the two subgroups had on validity and generalizability. Because the results for the two groups were not significantly different, only the combined analyses are presented. Census data on the annual household income in 1989 (11) for African-American households in LAC were compared with the income distribution of the populationbased sample of Afacan-American women with AIDS only, given that this subset is most representative of the general population.


   In total, 1005 surviving women diagnosed with HIV or AIDS were eligible to participate in a SHAS interview. Of these, 624 (62%) completed a questionnaire, including 379 (61%) who had been diagnosed with AIDS and 245 (39%) who had been diagnosed with HIV infection but had not yet progressed to AIDS.

   This group of African-American women with HIV and AIDS tended to be single parents (64%) who were unemployed (88%), and receiving public assistance (86%). When compared with women of other races with HIV and AIDS, African-American women were more likely to be single parents (OR = 1.6; 95% CI = 1.1, 2.4), reported receiving public assistance more often (OR = 1.8; 95% CI = 1.1, 2.9) and were more apt to live alone. African-American women were also more likely to delay seeking treatment for their HIV infection beyond 6 months of learning of their infection (OR = 2.4; 95% CI = 1.4, 4.3; Table 1).
   A larger proportion of both African-American women (76%) and Latinas (80%) reported annual household incomes of <$10,000 U.S. compared with the incomes of white women (48%), although these differences were not statistically significant. Only 22% of all AfricanAmerican households in LAC in 1989 reported annual household incomes <$10,000 U.S. (11), compared with 72% among the African-American women with AIDS in the study group.
   With respect to high-risk sexual behaviors, 61% of African-American women reported more than one male sexual partner in the last 5 years, a proportion that was significantly higher than that for women of other races (OR = 2.3; 95% CI = 1.6, 3.4). Few African-American women reported that they used a condom every time with their steady sex partners in the year prior to learning of their HIV infection (6%); this finding was also similar among all racial groups. An increased proportion of African-American women reported consistent condom use with their steady sex partner after learning of their HIV infection (37%); however, 35% reported inconsistent condom use following notification of an HIV infection. When controlling for number of sexual partners, AfricanAmerican women were also 2.0 times more likely to trade sex for drugs or money in the past 5 years (95% CI = 1.2, 3.5; Table 2).
   Regarding sexually transmitted disease (STD) history, African-American women were twice as likely to report at least one STD in their lifetimes and had statistically significantly higher reported lifetime histories of genital gonorrhea, syphilis, nongonococcal urethritis, and pelvic inflammatory disease (Table 3).
   In an analysis of injection drug use behaviors, 23% of African-American women reported some lifetime injection drug use, a proportion that did not differ significantly from women of other races. Heroin was the most commonly injected drug for African-American, white, and Latino women; no substantial racial differences were found in the most widely used drugs. Of the women who had ever injected drugs, African-American women were most likely to have shared needles (95%), although this was not statistically different from women of other races (OR = 3.7; 95% CI = 0.9, 16.1; Table 4).

Journal of Acquired Immune Deficiency Syndromes and Human Retrovirology, Vol 19, No. 4, December 1, 1998