Drug Treatment Disparities Among African Americans Living with HIVAIDS

1
Drug Treatment Disparities Among African
Americans Living with HIV/AIDS
Carleen H. Stoskopf, Sc.D. William Pearson,
Ph.D. Jong Deuk Baek, Ph.D. Yunho Jeon, M.S.
2
Background

• Many studies have identified disparities in
  health status, health care access, and health
  care utilization by race/ethnicity.
• In the late 1990s, clinical trials found the high
  efficacy of Highly Active Antiretroviral Therapy
  (HAART).
• Researchers found that African Americans were
  significantly less likely to use newer
  antiretroviral regimens (e.g., protease
  inhibitors and NNRTIs).

3
Background continued

• Several studies found that racial/ethnic
  minorities were less likely to use drugs for
  opportunistic diseases than whites (e.g., PCP and
  TB/MAC prophylaxis).
• This study explores changes in use of HAART and
  drugs for opportunistic diseases and to determine
  factors that influence the use of HAART by
  race/ethnicity.

4
HIV Cost and Services Utilization Study (HCSUS)
data

• Nationally representative sample
• Adults in care for HIV
• Multi-stage design (n 4,042)
• Geographical
• Medical provider
• Patients

5
HCSUS data

• Panel study
• Baseline January 96 - March 97
• N 2,864 (71)
• First follow-up December 96 - July 97
• N 2,466 (61)
• Second follow-up August 97 - January 98
• N 2,267 (56))

6
Independent Variables

• Age
• Gender
• Race
• Mode of HIV exposure
• Household composition
• Employment status
• Educational attainment

7
Independent Variables

• Insurance status
• Household income
• Lowest reported CD4 cell count

8
Dependent Variables

• HAART Coded as a positive response if the
  person indicated taking HAART in the past six
  months
• Combinations of NRTI plus certain PI or NNRTI
• 99 of the sample met the published eligibility
  criteria in 1996 for HAART
• CD4 lt 500 cells/µl OR HIV RNA gt 10,000 copies/ml
  OR symptomatic HIV or AIDS
• Handout 1

9
Dependent VariablesOpportunistic Infections Rx

• Cytomegalovirus (CMV)
• Common herpes virus causing retinitis and colitis
• Pneumocystic Carinii Pneumonia (PCP)
• Infection of the lungs caused by Pneumocystis
  carinii
• Tuberculosis (TB)
• Bacterial infection, Mycobacterium tuberculosis

10
Dependent VariablesOpportunistic Infections Rx

• Mycobacterium Avium Complex
• Bacterial infections, Mycobacterium avium,
  Mycobacterium intracellulare
• Fungal Infections
• Herpes Simplex Viruses
• HSV-1, HSV-2
• Immune System Boosters

11
Analysis

• Description of Sample
• National estimates
• Bivariate analysis
• Multivariate analysis (Odds Ratio)

12
Results of Analyses
13
Sample Characteristics of Respondents

• Distribution of the sample respondents is
  consistent for all three surveys
• All sample frequencies decrease across surveys
  except for Medicare recipients and some CD4
  counts.
• Table 1

14
Sample Characteristics of Respondents
15
Sample Characteristics of Respondents
16
Sample Characteristics of Respondents
17
Sample Characteristics of Respondents
18
Sample Characteristics of Respondents
19
Sample Characteristics of Respondents
20
Sample Characteristics of Respondents
21
Sample Characteristics of Respondents
22
Sample Characteristics of Respondents
23
Sample Characteristics of Respondents
24
HAART Use in the Three Consecutive Surveys

• Test of Independence (Chi square)
• Proportions of HAART use were dramatically
  increased (sample frequencies and estimated
  population weighted percentages) across all three
  surveys.
• All independent variables are statistically
  significant except age in the two follow-ups.
• Table 2

25
Weighted Population Estimate () of HAART Use by
Race/Ethnicity

• African Americans are less likely to use HAART
  than other racial/ethnic groups.

26
Weighted Population Estimate () of HAART Use by
Gender

• Females are less likely to use HAART than males,
  but the gap closes over the three surveys.

27
Weighted Population Estimate () of HAART Use by
Employment Status

• Those who are employed full time, or those who
  are disabled, are more likely to use HAART.

28
Weighted Population Estimate () of HAART Use by
Education

• Those who are more highly educated are more
  likely to use HAART.

29
Weighted Population Estimate () of HAART Use by
Insurance Status

• Those persons who have private insurance are
  more likely to use HAART.

30
Weighted Population Estimate () of HAART Use by
Income

• Those with higher incomes are more likely to use
  HAART.

31
Weighted Population Estimate () of HAART Use by
CD 4 Cell Count

• Those with lower CD4 counts are more likely to
  use HAART.

32
Rx for Opportunistic Diseases in the Three
Consecutive Surveys

• Test of Independence (Chi square) for
  Race/Ethnicity
• African Americans are the least likely group to
  use drugs for opportunistic diseases.
• This finding is true across all three surveys,
  except for TB treatment in the second follow-up
  survey.
• Table 3

33
Multivariate Logistic RegressionUse of HAART

• The multivariate logistic Regression allows for
  controlling the influence of the various
  independent variables.
• Race (being African American) is consistently
  statistically significant across all three
  surveys. When compared to Whites, AA are
  significantly less likely to have used HAART in
  the last six months.
• Odds ratios are 0.32, 0.54, and 0.70,
  respectively
• Table 4

34
Multivariate Logistic RegressionUse of HAART

• Other significant findings include
• Men having sex with men are more likely to use
  HAART at the second follow-up
• Those who are unemployed or not working are less
  likely to use HAART at the baseline survey, those
  not working are still less likely to use HAART at
  the first follow-up.

35
Multivariate Logistic RegressionUse of HAART

• As compared to the uninsured, those with
  Medicaid, private insurance, private HMO, or
  Medicare were significantly more likely to be
  receiving HAART at the baseline survey.
• At the second survey (first follow-up) only those
  with private insurance had a statistically
  significant advantage
• As expected, those with the lowest CD4 counts
  were significantly more likely to receive HAART
  therapy.

36
Multivariate Logistic Regression Rx for
Opportunistic Diseases
Baseline Survey African Americans were
significantly less likely than Whites to receive
drug treatment for these disease
categories Cytomegalovirus Pneumocystis Tuberc
ulosis Fungal Infections Herpes Immune
System Booster Table 5 - 10
37
Multivariate Logistic Regression Rx for
Opportunistic Diseases

• Second survey (first follow-up)
• At the second survey, African Americans were
  significantly less likely to receive drug
  treatment for all disease except pneumocystis.
  These include
• Cytomegalovirus Tuberculoses
• Fungal Infections Herpes

38
Multivariate Logistic Regression Rx for
Opportunistic Diseases

• Third survey (second follow-up)
• By the third survey, African Americans were still
  significantly less likely than Whites to receive
  drug treatment for
• Cytomegalovirus Tuberculosis
• Fungal Infections

39
Conclusions

• African Americans were consistently less likely
  to receive appropriate treatment for HIV/AIDS and
  other infections associated with this disease as
  HAART was introduced.
• Over time, the racial disparities in HAART use
  decrease, but statistical differences remain
  between African Americans and Whites.