530
Higher Adjusted Mortality Rates among Publicly Insured Patients and Blacks in the HIV Outpatient Study
Frank Palella*1, R Baker2, J Chmiel1, E Tedaldi3, R Novak4, K Buchacz5, and J Brooks5
1Feinberg Sch of Med, Northwestern Univ, Chicago, IL, US; 2Cerner Corp, Vienna, VA, US; 3Temple Univ Sch of Med, Philadelphia, PA, US; 4Univ of Illinois at Chicago, Coll of Med, US; and 5CDC, Atlanta, GA, US
Background: Mortality rates among HIV-infected
persons have fallen dramatically and remain low in the HAART era. Identifying
differences in risk for death based upon sociodemographic factors can enhance
public health efforts to further improve survival.
Methods: We analyzed risks associated with
mortality among HIV Outpatient Study (HOPS) patients using multivariable
Poisson models that evaluated the following: sociodemographic factors (age,
sex, race, insurance type), HIV clinical status (CD4 cell count/cm3
at HAART initiation [CD4], prior AIDS-defining illness), and co-morbidities
(chronic infection with viral hepatitis B or C, illicit drug use, smoking).
Results: We included data from 2383 persons seen
from January 1, 1999 to December 31, 2005 who were followed at least 6 months
and had known dates of HAART initiation. Mean duration of follow-up was 5.9
years. The number of deaths observed was 178. We noted discordances in death
rates/100 person-years among persons with public vs non-public insurance (4.0
vs 1.3) and among non-Hispanic blacks vs non-blacks or Hispanics (non-blacks) (3.3
vs 1.9), prompting further evaluation in multivariable analyses. In adjusted
models, persons with public insurance (Medicare, Medicaid, n = 872) had
higher death rates than persons with other types of insurance (n = 1511,
adjusted rate ratio [aRR] 1.77, p <0.001). Other factors
independently associated with greater mortality risk were non-Hispanic blacks
race (aRR 1.52, p = 0.009), advancing age (aRR 1.31 per 10-year
increase, p <0.001), and lower CD4 at HAART initiation (aRR 0.900 per
50-cell increase, p <0.001). CD4 at HAART initiation was lower for
publicly insured versus non-publicly insured patients (257 vs 311, p <0.001);
and for non-Hispanic blacks versus non-blacks (261 vs 301, p <0.001).
Cause of death data were available for 77% of persons with either public or
non-public insurance, 66% for non-Hispanic blacks, and 83% for non-blacks.
Among persons for whom cause of death data were available, non-AIDS-associated
deaths predominated for both publicly and non-publicly insured (63% and 56%)
and for both non-Hispanic blacks and non-blacks (66% and 56%) and included
cardiac, renal, diabetes, and non-AIDS cancer diagnoses.
Conclusions: Public
insurance and non-Hispanic black race were independently associated with
increased risk of death after adjusting for other known HIV-related mortality
risks. Non-AIDS causes of death that included preventable co-morbid illnesses
predominated, suggesting that modifiable health risks may contribute to these
disparities.
|