Antiretroviral Medication Diversion among HIV-positive Substance Abusers in South Florida

Medication adherence is critical in the management of many chronic illnesses, including HIV.¹ Antiretroviral (ARV) non-adherence increases the risk of treatment failure, drug resistance, and disease transmission.² Our recently completed field research among HIV+ patients in urban South Florida documented modest levels of ARV adherence, and a related, yet understudied, phenomenon – the diversion of ARV medications. Diversion – the unlawful channeling of regulated pharmaceuticals from legal sources to illicit markets³ – has typically been studied in relation to prescription opioids and other medications with significant abuse potential.⁴ Nevertheless, the diversion of ARVs by patients is apparent in South Florida,⁵ and has direct implications for non-adherence, making it a critical issue for patient care and public health. We examined factors that impact vulnerability to ARV diversion among highly marginalized HIV+ individuals.

Methods

Guided by targeted sampling in geographic areas with high HIV prevalence and poverty indices,⁶ direct outreach was utilized to recruit indigent HIV+ substance abusers in South Florida between 2010 and 2012. Research staff distributed study cards and flyers in street venues and HIV service organizations. We enrolled approximately equal numbers diverting their personal ARV(s) (N = 251) and not (N = 252); due to a lower average recruitment rate, the time period required to achieve the full sample of diverters was 24% longer. Eligibility criteria were: age 18 or older; cocaine or heroin use 12 or more times in the prior 3 months; documented HIV+ status; and, current ARV prescription. In addition, diverters endorsed at least one occasion of ARV diversion in the prior 3 months. Systematic screening was conducted to verify eligibility: 2,112 individuals were screened, 599 met study eligibility criteria, 503 were enrolled. The primary reason for non-enrollment (84%) was repeated failure to present for the interview. Following informed consent, participants completed a single face-to-face interview based on the GAIN⁷ instrument, assessing demographics, substance use, DSM-IVR dependence and mental health status. Standardized instruments assessed HIV diagnosis/treatment history,⁸ attitudes toward HIV care providers,⁹ health literacy,¹⁰ HIV-related stigma,¹¹ ARV knowledge and attitudes,^{12, 13} and ARV adherence¹⁴; a new instrument assessed ARV diversion. Participants received a $30 stipend upon interview completion. Bivariate logistic regression models examined the factors associated with ARV diversion in this sample; significant predictors from the bivariate analyses were included in the multivariate model.

Results

Median age was 46; 59.4% were male. 55.3% met DSM-IVR criteria for past year substance dependence. 81.4% reported monthly income below $1,000; 39.2% were homeless in the prior 3 months. Diverters reported a median of 7 lifetime episodes of ARV diversion. Self-reported ARV adherence in the past week was 95% or better among 54.1% of participants; diverters had .26 times lower odds of achieving 95% adherence ([95% CI 0.18, 0.37]; p ≤ .001). Table 1 displays the bivariate and multivariate predictors of ARV diversion. Male gender, severe depression, severe anxiety, substance dependence, recent homelessness, and high HIV-related stigma were each associated with increased odds of diverting prescribed ARV medications. The odds of diversion were lower among participants reporting higher HIV-related treatment knowledge, higher health literacy, and those achieving 95% ARV adherence. In the multivariate model, recent homelessness and male gender were associated with increased odds of ARV diversion; higher HIV-related treatment knowledge, better adherence, and higher health literacy remained as significant protective factors.

Table.

Bivariate and Multivariate Logistic Regression Models Predicting Past 90 Day ARV Diversion among HIV+ Substance Abusers, (N = 503)

Bivariate Models ARV Diversion1	Diverters n = 251	Non-Diverters n = 252	Odds Ratio	95% CI
Demographics
Age, mean (SD)	45.9 (7.9)	46.2 (7.6)	1.00	0.97, 1.02
Male Gender1, n (%)	166 (66.1)	133 (52.8)	1.75**	1.22, 2.50
Years HIV diagnosis, mean (SD)	13.0 (7.4)	13.5 (7.2)	0.99	0.97, 1.02
HS Education1, n (%)	132 (52.6)	152 (60.3)	1.37	0.96, 1.95
Health Status
Good self-reported health2, n (%)	124 (49.4)	125 (49.6)	1.01	.71, 1.43
95% ARV Adherence1, n (%)	95 (37.8)	177 (70.2)	0.26***	0.18, 0.37
Risk Factors
Homeless in past 90 days1, n (%)	120 (47.8)	77 (30.6)	2.08***	1.45, 3.00
Substance Dependence1, n (%)	155 (61.8)	123 (48.8)	1.69**	1.19, 2.41
Severe Depression1, n (%)	155 (61.8)	120 (47.6)	1.78**	1.25, 2.53
Severe Anxiety1, n (%)	93 (37.1)	65 (25.8)	1.69**	1.16, 2.48
HIV-related Stigma Score, mean (SD)	24.6 (7.7)	23.1 (7.5)	1.03*	1.00, 1.05
Protective Factors
HIV Treatment Knowledge, mean (SD)	5.6 (1.5)	6.0 (1.0)	0.76***	0.66, 0.88
ARV Medication Attitudes, mean (SD)	8.3 (2.3)	8.6 (1.4)	0.94	0.85, 1.03
Health Literacy, mean (SD)	8.5 (3.6)	9.6 (2.9)	0.91***	0.86, 0.96
Attitudes toward HIV Providers, mean (SD)	43.4 (5.7)	44.2 (4.9)	0.97	0.94, 1.00
Multivariate Model
Male Gender1			1.99***	1.34, 2.96
95% ARV Adherence1			0.29***	0.20, 0.43
Homeless in past 90 days1			1.64*	1.09, 2.47
Substance Dependence1			1.30	0.85, 1.98
Severe Depression1			1.23	0.76, 1.99
Severe Anxiety1			1.08	0.65, 1.79
HIV-related Stigma			0.99	0.96, 1.02
HIV Treatment Knowledge			0.85*	0.72, 0.99
Health Literacy			0.92**	0.86, 0.98

¹Reference category is ‘no’

²Reference category is poor/fair.

^*p ≤ .05;

^**p ≤ .01;

^***p ≤ .001.

Discussion

By comparison to controlled substances, the diversion of legend prescription drugs garners less attention and resources from law enforcement, the health practitioner and scientific communities, policy makers, educators and industry. Nevertheless, diversion of non-controlled substances, such as ARVs, has clear implications for the integrity of the medication supply chain, for patient safety, and for public health.¹⁵ Our data show that indigent substance abusing patients face challenges with ARV adherence, and moreover, that non-adherence is linked to the diversion of ARVs. The targeting of indigent, street-based HIV+ patients by pill brokers or “collectors” offering small financial incentives to purchase monthly supplies of ARV medication(s) appears to be common in South Florida^5,16; even among non-diverters, 37.4% reported being approached for this purpose (H. Surratt, unpublished data 2011). The pricing structure of ARV medications makes them highly profitable in the illicit market,¹⁷ in effect incentivizing diversion and the targeting of HIV+ individuals with high levels of competing needs. In our sample, men appeared to be especially vulnerable to ARV diversion. Although the reason is unclear, males were more likely to report recent homelessness, which may increase their exposure to street-level pill brokers and dealers. Interventions to enhance medication adherence and reduce diversion should be tailored for the most vulnerable, homeless HIV+ individuals. Knowledge of ARV treatment regimens and health literacy factors are clearly amenable targets for individual level practitioner-based interventions; HIV treatment specialists and other health practitioners should be aware of the potential for ARV diversion in order to better support their patients. Housing insecurity requires increased attention at a structural level in order to reduce the exploitation of vulnerable HIV+ individuals.