摘要：Objectives. We described the racial/ethnic disparities in survival among people diagnosed with AIDS in Florida from 1993 to 2004, as the availability of highly active antiretroviral therapy (HAART) became widespread. We determined whether these disparities decreased after controlling for measures of community-level socioeconomic status. Methods. We compared survival from all causes between non-Hispanic Blacks and non-Hispanic Whites vis-a-vis survival curves and Cox proportional hazards models controlling for demographic, clinical, and area-level poverty factors. Results. Racial/ethnic disparities in survival peaked for those diagnosed during the early implementation of HAART (1996–1998) with a Black-to-White hazard ratio (HR) of 1.72 (95% confidence interval [CI] = 1.62, 1.83) for males and 1.40 (95% CI = 1.24, 1.59) for females. These HRs declined significantly to 1.48 (95% CI = 1.35, 1.64) for males and nonsignificantly to 1.25 (95% CI = 1.05, 1.48) for females in the 2002 to 2004 diagnosis cohort. Disparities decreased significantly for males but not females when controlling for baseline demographic factors and CD4 count and percentage, and became nonsignificant in the 2002 to 2004 cohort after controlling for area poverty. Conclusions. Area poverty appears to play a role in racial/ethnic disparities even after controlling for demographic factors and CD4 count and percentage. The HIV/AIDS epidemic has disproportionately affected the non-Hispanic Black population in the United States. In 2008, an estimated 545 000 non-Hispanic Blacks were living with HIV/AIDS. 1 The estimated prevalence of HIV infection for 2008 among non-Hispanic Blacks was 18.2 per 1000 population, more than 7 times higher than the estimated rate for non-Hispanic Whites (non-Hispanic Whites; 2.4 per 1000). 1 Non-Hispanic Blacks as a group not only have a higher prevalence of HIV/AIDS, but once infected also have a lower survival rate. The 3-year survival rate in the United States for people diagnosed with AIDS between 2001 and 2005 was 80% among non-Hispanic Blacks compared with 84% for non-Hispanic Whites, 83% for Hispanics, and 88% for Asians, 2 further contributing to the disparities in the HIV mortality rate of 16.8 per 100 000 among non-Hispanic Blacks compared with 1.6 per 100 000 among non-Hispanic Whites during 2007. 3 Race/ethnicity is a fundamentally social as opposed to a biological construct, 4,5 and survival disparities between non-Hispanic Blacks and non-Hispanic Whites have not generally been seen in settings with universal health care access such as in the Veterans Administration health care system, 6 the military health care system, 7 or a health maintenance organization. 8 Therefore, potentially modifiable social explanations for the observed racial/ethnic disparities in survival should be examined. A most remarkable advance in medical treatment in the past century was the development of highly active antiretroviral therapy (HAART). It led to a significant improvement in survival from HIV/AIDS, 9–12 but racial/ethnic disparities in HIV/AIDS survival remain 2, 13–16 and in New York City appeared to widen. 17 Despite these well-recognized health disparities, there is a critical gap in the knowledge about why the disparity exists. Two population-based studies, both in San Francisco, California (a city that has provided free HIV care for those who cannot afford it), found that Black race was no longer associated with survival between 1996 and 2001 after controlling for neighborhood socioeconomic status (SES), and that this SES effect seemed to be related to HAART use. 18,19 However, in a study using HIV surveillance data from 33 states, racial/ethnic disparities in 5-year survival after HIV diagnosis between 1996 and 2003 persisted after adjusting for county-level SES and other factors. 20 The objective of this study was to describe the racial/ethnic disparities in AIDS survival in Florida among people diagnosed with AIDS between 1993 and 2004 (a period spanning the time before and during the widespread availability of HAART) and to determine if these disparities decrease after controlling for community-level SES.
