Transcript Patients with Low Adherence

PREDICTORS OF LOW ADHERENCE TO HAART AMONG INDIGENT HIV-INFECTED MEN WHO HAVE SEX
WITH MEN, IN MIAMI/DADE COUNTY’S HIV DRUG ASSISTANCE PROGRAM
Jason Gagnon, M.A., M.P.H.,1 Seth Welles, Sc.D., Ph.D.,2 Anthony J. Japour, MD3
1Division of Graduate Medical Science, Boston University School of Medicine, Boston, MA, 02118, 2 Department of Epidemiology, Boston University School of
Public Health, Boston, MA,
Abstract 156575
Background: This study evaluated whether demographic
and behavioral factors, current illicit drug use and history of
addiction, HIV disease indicators and treatment history are
associated with low HAART adherence in populations
seeking free healthcare.
Results: Ninety-seven HIV+ MSM enrolled in an AIDS
Drug Assistance Program in Miami participated in focus
groups to identify barriers to HAART adherence and condom
use. Surveys administered during focus groups collected
data on demographics, illicit drug use/history of addiction,
and history of HAART treatment adherence.
Findings: Patients recruited were mature (median age: 40
yrs.) with lower level of education (77% ≤ high school), and
high unemployment (79%); more than half were men of
color. Greater than 50% of subjects reported being
noncompliant to HAART ever, and over one-third reported
noncompliance in the prior weekend. Factors associated
with low adherence included marijuana use [odds ratio
(OR): 4.67, p = 0.01], and years since HIV diagnosis [OR:
1.09 (9% increase per year), p< 0.03], and perceived
barriers to ARV adherence [OR: 14.05; 95% CI: 2.40,
82.16; p= 0.003, for quartile 2 vs quartile 1 of total score
for inventory of barriers; and OR:10.02; 95% CI:1.45,
69.13), for quartile 3 vs quartile 1 of total score].
Conclusion: In this patient group of HIV+ men receiving
free HIV medications through government supported AIDS
Drug Assistance program, it was current drug use and
addiction history (not demographic profiles or ART
treatment parameters) that were associated with HAART
non-adherence. Our findings suggest that treatment of
current illicit drug use may have optimal impact on
treatment adherence in groups of lower income HIV+ men
with long-term infection.
Introduction
Adherence to the HAART regimen is inarguably the most
important determinant for long term virologic suppression
and health in HIV+ individuals. Previous studies have
identified a broad array of factors associated with lower
HAART adherence, including:
•demographics (younger age, black race, education level,
income, employment status1, transmission route)
3Department
of Medicine, University of Miami Miller School of Medicine, Miami FL, 33101
Methods and Data Collection
Results (Continued)
Participants
•97 HIV-infected men recruited through Miami’s ADAP Program
•54% are White/Hispanic, 29% Non-Hispanic Black, and 6% are
Black/Haitian. 98% are at least 20 years old.
Table 2. HIV/AIDS Disease Indicators and Treatment Parameters, by Level of ART Adherence
Eligibility
•Men were eligible if they were at least 18 years old, HIV-positive,
sexually active, self-identified as men who have sex with men (MSM),
able to speak, read, and converse in English, Spanish, or Creole, and
able to provide written informed consent.
Data Collection
•Data for analyses were collected from 97 focus group participants.
Data that were collected included: regarding self-reported adherence
to HIV medications, barriers to adherence, recreational drug use, and
condom use.
Data Analysis
18,500
(1,165, 240,000)
1,160
(25, 25,000)
11,000
(505, 135,000)
0.05
Median CD4 cell count/mm3
(25%-75% interquartile range)
300
(165, 432)
416
(303, 603)
350
(219, 554)
0.02
Median Months on ARV
(25%-75% interquartile range)
60
(36, 120)
48
(24, 96)
60
(24,120)
0.21
NRTI-Containing
39/46 (84.7)
41/43 (95.3)
80/89 (90.0)
0.16
NNRTI-Containing
14/46 (30.4)
11/43 (25.6)
25/89 (28.1)
0.65
PI-Containing
30/46 (65.2)
25/43 (58.1)
55/89 (61.8)
0.52
3/46 (6.5)
2/43 (4.7)
5/89 (5.6)
1.00
One
5/46 (10.9)
1/43 (2.3)
6/89 (6.7)
0.09
Two
17/46 (37.0)
15/43 (34.9)
33/89 (37.1)
Three
19/46 (41.3)
18/43 (41.9)
39/89 (43.8)
Four
5/46 (10.9)
9/43 (20.9)
16/89 (18.0)
Median Plasma HIV-1 RNA copies/ml
(25%-75% interquartile range)
Overall ARV Regimen (%)
Fuzeon-Containing
No. ARVs in Rx Regimen (%)
Frequency of Likelihood (%) for Each Barrier
Possible Barriers to Adherence
1
Highly Unlikely
2
3
No Likely or Unlikely
4
5
Highly Likely
Alcohol
52 (54.2)
11 (11.5)
8 (8.3)
2 (2.1)
24.0 (23)
Marijuana
57 (60.0)
6 (6.3)
7 (7.4)
2 (2.1)
24.2 (23)
•The dependent variable: Self-reported lower levels (<50%) of ART
adherence
Other Recreational drug use
59 (62.8)
5 (5.3)
8 (8.5)
2 (2.1)
21.3 (20)
Organization
57 (61.3)
10 (10.8)
9 (9.7)
3 (3.2)
15.1 (14)
Forgetting
36 (37.5)
23 (24.0)
9 (9.4)
9 (9.4)
19.8 (19)
•Contingency table analysis to identify differences: Demographics, level of
HIV knowledge, perceived barriers to ART adherence
Think you took meds but did not
40 (42.6)
24 (25.5)
7 (7.5)
7 (7.5)
10.0 (16)
Don’t believe in taking medications
69 (71.9)
8 (8.3)
2 (2.1)
3 (3.3)
14.6 (14)
Use of alternative non-prescription drugs
65 (69.2)
8 (8.5)
6 (6.4)
5 (5.3)
10.6 (10)
Number of pills
49 (52.1)
11 (11.7)
22 (23.4)
4 (4.3)
8.5 (8)
Toxic side effects
48 (50.5)
11 (11.6)
13 (13.7)
8 (8.4 )
15.8 (15)
•Categorical data: Tests for differences included Pearson Chi-square and
Fischer Exact tests, as appropriate.
Physical side effects
41 (42.7)
12 (12.5)
17 (17.7)
7 (7.3)
19.8 (19)
Work Schedule
62 (65.3)
10 (10.5)
8 (8.4)
6 (6.3)
9.5 (9)
Cost of medications
65 (68.4)
5 (5.3)
7 (7.4)
7 (7.4)
11.6 (11)
•Simple and multiple logistic regressions were conducted to identify
independent risk factors for associations with low levels of ART adherence.
All odds ratios are presented with associated 95% confidence intervals.
Selling medications to cover other expenses
73 (77.7)
5 (5.3)
1 (1.1)
1 (1.1)
14.9 (14)
Selling medications to buy recreational drugs
74 (78.7)
2 (2.1)
3 (3.2)
1 (1.1)
14.9 (14)
Homelessness
58 (61.7)
6 (6.4)
6 (6.4)
7 (7.5)
18.1 (17)
Fear of people learning your HIV status
55 (57.3)
8 (8.3)
9 (9.4)
5 (5.2)
19.8 (19)
Food Requirements
56 (58.3)
10 (10.4)
13 (13.5)
4 (4.2)
13.5 (13)
•Continuous data: Tests for difference between groups included Wilcoxon
rank-sum tests
Table4. Associations of Substance Use, HIV Treatment History and Disease Indicators, and Levels of Perceived Barriers to ART Adherence
with Self-Reported Low Adherence
Results
Factor
Table 1. Demographics and Measures of Sexual Orientation/Homosexual Behaviors
Race (%)
•rigor of treatment, level of satisfaction with HAART
information, fear of side effects, occurrence of side
effects (e.g. including lipodistrophy causing the person to
be “outed” as HIV+) and pill burden are major predictors
of adherence.
Ethnicity (%)
Age (yrs)
Patients with Low
Adherence (n=47)
45 (43,50)
Patients with Higher
Adherence (n=45)
45 (41, 48)
Black
26/44 (59.1)
20/42 (47.6)
Non-Black
18/44 (40.9)
22/42 (52.4)
Haitian
Total study
population
45 (41,49)
46/86 (53.5)
P-value
0.53
0.29
Adjusted OR
(95% CI)
P-value
(adjusted model)
Marijuana Use
6.05 (2.12, 17.25)
4.67 (1.43, 15.27)
0.01
Cocaine Use
5.00 (1.51, 16.60)
- - -
Ever been in Drug Treatment Program
2.62 (0.99, 6.90)
- - -
Years since HIV Diagnosis
(increased risk per year)
1.07 (>1.00, 1.14)
1.09 (1.0, 1.19)
CD4 cell count
(increased risk per 100 cells)
0.90 (0.80, 1.10)
- - -
Viral Load (copy number)
1.00 (1.00, 1.00)
- - -
1.0 (referent)
- - -
500 -9999 copies
0.67 (0.21, 2.16)
- - -
> = 10000 copies
2.34 (0.92, 5.97)
- - -
Quartile 4 (Highest)
2.91 (0.70, 12.0)
3.02 (0.53, 17.24)
0.21
Quartile 3
8.0 (1.84, 34.79)
10.02 (1.45, 69.13)
0.02
Quartile 2
8.0 (1.93, 33.18)
14.05 (2.40, 82.16)
0.003
Quartile 1
1.00 (referent)
1.00 (referent)
< 500 copies
3/40 (7.5)
3/44 (6.8)
6/84 (7.1)
Hispanic
19/40 (47.5)
25/44 (56.8)
44 /84(52.4)
Non-Hispanic
18/40 (45.0)
16/44 (36.4)
34/84 (40.5)
Gay
19/46 (41.3)
18/44 (40.9)
37/90 (41.1)
Bisexual
12/46 (26.1)
17/44 (38.6)
29/90 (32.2)
Heterosexual
15/46 (32.6)
9/44 (20.5)
24/90 (26.7)
0.69
0.03
Perceived Barriers to Adherence
Summary Score Quartile (%)
High School
32/39 (82.1)
30/41 (73.2)
62/80 (77.5)
College/Grad
School
7/39 (17.9)
11/41 (26.8)
18/80 (22.5)
Part-time or
unemployed
10/43 (23.3)
9/44 (20.5)
19/87 (21.8)
Full-time
33/43 (76.7)
33/44 (79.5)
68/87 (78.2)
Always
13/42 (31.0)
13/40 (32.5)
26/82 (31.7)
Usually
4/42 (9.5)
9/40 (22.5)
13/82 (15.9)
Sometimes
9/42 (21.4)
8/40 (20.0)
7/82 (20.7)
Rarely
4/42 (9.5)
2/40 (5.0)
6/82 (7.3)
12/42 (28.6)
8/40 (20.0)
20/82 (24.4)
0.31
Conclusions
Education (%)
0.34
Employment (%)
0.75
Frequency MSM
activity (%)
When under
influence /high
Unadjusted OR
(95% CI)
Viral Load (categorical)
Orientation (%)
We estimate rates of ART adherence and identify
independent predictors of adherence in a clinic-based
sample of men in Miami; the sample was intentionally
enrolled to be reflective of the community served by the
Miami-Dade County Health Department’s AIDS Drug
Assistance Program, which includes approximately 50%
white/Hispanic and 50% black/African-American.
P-value
Table 3. Perceived Barriers to ART Adherence, for all Participants
•factors associated with HIV/AIDS disease progression,
treatment, and clinical care (medical care status, higher
plasma HIV-1 RNA level, treatment type/regimne)
•Levels of adherence changes over time along with
changes of attitudes, beliefs, physical ability, and
cognitive changes.
Total
•Study variables: Age in years, race/ethnicity, sexual orientation identity,
education, employment status, sexual behaviors, condom use during
intercourse, recreational substance abuse, HIV medical history (most recent
CD4 cell count and plasma HIV-1 viral load), and parameters of current HIV
treatment regimen (types and numbers of specific ARVs).
Demographic
•Depression is an independent predictor of adherence
associated with low self-efficacy, inability to manage side
effects, medication fatigue, and inadequate
understanding that non-adherence can advance disease
status—again, all of which are additional predictors of
adherence.
Patients with Higher
Adherence (n=45)
HIV Disease Indicators
•mental health indicators and correlates (depression,
drug use, partner intimacy, sexual functioning)
•Mental health indicators and recreational drug use both
impact adherence to antiretroviral therapy [ART] and are
potentially modifiable.
Patients with Low
Adherence (n=47)
HIV Disease/Treatment Parameter
•Study participants who used marijuana were 4 times as likely to be low HAART
adherers
•Length since HIV diagnosis was significantly associated with low HAART
adherence
•Demographics and other disease indicators were not associated with low
adherence in this group.
•Findings suggest that treatment of current illicit drug use may have optimal
impact on treatment adherence in groups of lower income HIV+ men with longterm infection.
0.28
Acknowledgements
This project would not have been possible without the continued support and expert contributions of my principal advisor and
mentor, Dr. Seth Welles. Additionally, I would like to thank Dr. Anthony Japour for his assistance in making available the raw data
available used in this research project as well as for his insights and contributions into the manuscript. Finally, I would like to
express gratitude for the research subjects who ultimately made this study possible.