Opiate substitution treatment has been associated with better adherence to lifesaving antiretroviral medications, but the impact of other substance abuse treatment on adherence is unknown.
In this study, 215 patients who had been in adherence-focused research studies provided electronically-measured adherence data and a measure of whether the patient had recently been in substance abuse treatment. Recent engagement in substance abuse treatment was independently associated with significantly higher adherence, after covarying for recent substance use and other factors potentially affecting adherence.
The findings suggest that substance abuse treatment is associated with better adherence. Potential mechanisms by which substance abuse treatment improves adherence, such as more stability or more future-orientation, require further study.
Keywords:Medication adherence; AIDS; Substance abuse; Treatment
There are considerable data suggesting that current substance abuse is associated with both worse adherence to antiretroviral medications and worse attendance at HIV treatment visits . HIV-positive patients who use drugs or alcohol may be referred to standard substance abuse treatment with the hope that medication adherence will improve with abstinence.
However, among substance abuse treatments, only opiate substitution treatment has consistently been associated with better adherence, as summarized in a 2008 meta-analysis . Both buprenorphine [3,4] and methadone maintenance [5,6] have been consistently associated with better adherence. Patients who received both opiate substitution treatment and directly observed therapy have shown even better adherence over time than with those receiving opiate substitution alone [7,8].
It is not known whether enrollment in substance abuse treatments other than opioid substitution is associated with better adherence. In this study, we used data pooled from two clinical trials of adherence-focused interventions to test the hypotheses that HIV-infected people who had recently been in substance abuse treatment would have better antiretroviral adherence, and better adherence over time, than those who had not had recent substance abuse treatment.
Overview and sampling
The data were drawn from two of the sixteen studies in the MACH14 cohort . The MACH 14 cohort is a dataset pooled from 16 studies conducted at 14 sites across 12 states. Each study in MACH14 used electronic data monitoring (EDM) pillcaps to objectively measure participants’ adherence to antiretroviral medication. The focus of this study was on non-methadone substance abuse treatment so studies conducted in methadone maintenance programs were not considered in this analysis. From the 1579 participants in the MACH14 dataset, we identified 215 from two studies based outside methadone clinics [10,11] because only these two studies’ participants had both EDM and substance abuse treatment status data. Written informed consent was obtained for participation in the parent studies, and the Yale Institutional Review Board approved the secondary analyses.
Patients were asked about engagement in substance abuse treatment and use of specific substances for varying preceding timeframes: one of the two studies (n=47) asked about participation in substance abuse treatment during the past 90 days and use of specific substances over the past 30 days, while the other study (n = 168) asked about treatment over the past 30 days and substance use over past 14 days. To aggregate substance use data across studies, variables representing use of specific substances were defined as the proportion of days within the asked-about timeframe the person had used each of several substances.
This analysis used data collected at the first time-point at which participants had EDM data for the preceding four weeks, had also been asked about being recently enrolled in substance abuse treatment, and were not enrolled in a methadone-clinic-based study. To estimate the effect of substance abuse treatment on adherence, adherence was calculated for the four weeks up to and including the date recent substance abuse treatment enrollment was assessed, as well as for the four weeks after the substance abuse treatment determination. Adherence in each week was calculated by dividing the weekly number of doses taken by the weekly number of prescribed doses for each medication, with adherence to each medication capped at 100%. Adherence for a patient on multiple anti-retrovirals was calculated by averaging across prescribed medications.
The effects of substance abuse treatment on adherence were determined in multivariate analyses that included a grouping variable denoting whether the patient was enrolled in substance abuse treatment and a variable reflecting substance abuse treatment over time. The analyses were conducted controlling for sociodemographic characteristics that might differ between patients in, and not in, substance abuse treatment. To control for the anticipated finding that patients in substance abuse treatment would have more active drug use than a reference group including people who had never had significant substance use, analyses included a measure representing the largest proportion of days during which participants had used either cocaine, opiates, or stimulants. Cannabis use was not included in this measure of illicit drug use because in a separate analysis of the MACH14 dataset  and in an earlier study  recent cannabis use was not associated with worse adherence.
Analyses were run with SAS 9.2. The model included random effects for intercept and slope as this model had better fit to the data than models with fixed effects only.
Overall, the sample was 65.7% male, and was composed of 49.3% African American, 2.8% Hispanic, and 19.5% Caucasian participants. Only 40.1% were employed. Altogether, 39.5% were men who have sex with men (MSM). Mean CD4+ count was 232.2 (+/− 206.9) and 8.5% of the 199 patients with viral load results available had undetectable HIV viral loads (<400); these measures of HIV disease did not differ between those in and those not in substance abuse treatment. Mean adherence during the first week was 82% (+/− 30%) overall, with 57.2% (123/215) having 100% adherence, and 12.6% (27/215) having adherence over 90% but less than 100%.
Of the 215 participants, 13 had recently been in substance abuse treatment--- one had been in residential treatment and the rest did not specify the type of treatment attended. Table 1 compares those in substance abuse treatment to those not in it; those in treatment were less likely to be African American (Fisher’s Exact Test =.04), and had used alcohol on a lower percentage of days (t=5.9, p<.0001). Only one of the thirteen patients in substance abuse treatment had used alcohol recently and none had used other substances, whereas there were substantial rates of recent use of several substances among those out of treatment.
Table 1. Baseline group characteristics
The multivariate analysis estimated the effects of sociodemographic measures (age, gender, race/ethnicity, MSM-status, employment), time in weeks, and maximum proportion days reported using an illicit drug (Table 2). As expected, adherence was slightly and significantly worse over time (estimate =−.02, p=.002). Being in substance abuse treatment was associated with better overall adherence (estimate =0.09, p= .05). The week by treatment interaction was not statistically significant and was removed from the model; substance abuse treatment status did not significantly change adherence over time.
Substance use treatment may improve adherence by several mechanisms. Although the analyses controlled for illicit drug use, it is possible that our self-report measures of substance use understated the impact of substance abuse treatment on substance abuse and that it is in fact abstinence that facilitates adherence. In one of the few randomized controlled studies of HIV-positive drug users in which abstinence was the target outcome, there was a trend towards a significant correlation between consecutive weeks of toxicology-tested abstinence during the intervention and reductions in viral load . There is also evidence from a naturalistic longitudinal cohort study that attendance at HIV treatment, a sine qua non for adherence, appears to improve with newly-achieved abstinence .
Substance abuse treatment might improve adherence by mechanisms other than facilitating abstinence from using drugs. Substance abuse treatment typically involves case management  to address the unstable housing characteristic of drug users . Stable housing arrangements during substance abuse treatment would be expected to foster adherence, in that stable routines have been associated with better adherence . Substance abuse treatment also focuses patients on future goals, an orientation that has been described as fostering adherence , and substance abuse treatment can involve re-arranging social networks in ways that also might foster better adherence .
It is possible that enrollment in substance abuse treatment reflects a lurking un-measured variable associated with both being in substance abuse treatment and better adherence. The finding of better adherence among people in substance abuse treatment was not buttressed by finding better adherence over time among patients in treatment. However, it might have been difficult to detect the time course of benefit from substance abuse treatment because the data did not specify when patients were entering, continuing, or finishing substance abuse treatment.
There are several other caveats in interpreting these findings. Substance abuse was measured by self-report, and it is possible that substance abuse was disproportionately under-reported by people out of substance abuse treatment, thus exaggerating the impact of substance abuse treatment on adherence. The type of substance abuse treatment was not specified and the findings may not apply to all types of substance abuse treatment. Finally, the sample size was modest, and the number of participants in substance abuse treatment was small. It is noteworthy that although adherence decreased on average over time, the course of adherence varied significantly by person. Further analyses should test variables that may account for individual differences in adherence over time.
These findings lend some support to the clinical practice of addressing substance use in an effort to improve adherence. The crucial next step is to develop and prospectively test substance abuse-focused interventions for patients with both substance abuse and adherence problems.
The authors declare that they have no competing interests.
MR designed the analyses and drafted the manuscript. All the authors read and approved the final manuscript. All the authors except ACB and YW contributed patient data to the MACH 14 dataset. ACB helped design the analyses, and the aggregated dataset from which they were drawn. YW and HHL conducted the statistical analyses.
This research was supported by the multi-site adherence collaboration in HIV (MACH14) grant R01MH078773 from the National Institute of Mental Health (NIMH), Office on AIDS. The original grants of individual participating studies are: R01DA11869, MH54907, R01 NR04749, R01 NR04749, MH68197, R01 DA13826, K23MH01862, MH01584, R01 AI41413, R01 MH61173, NIH/ NIAID AI38858, AI069419, K02 DA017277, R01 DA15215, NIMH P01 MH49548, MH58986, RO1MH61695, CC99-SD003, CC02-SD-003 and R01 DA015679. We would like to thank all the patients who participated in each of the individual studies. The content of the paper is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Altice FL, Bruce RD, Lucas GM, Lum PJ, Korthuis PT, Flanigan TP, Cunningham CO, Sullivan LE, Vergara-Rodriguez P, Fiellin DA, Cajina A, Botsko M, Nandi V, Gourevitch MN, Finkelstein R, Collaborative B: HIV treatment outcomes among HIV-infected, opioid-dependent patients receiving buprenorphine/naloxone treatment within HIV clinical care settings: results from a multisite study.
Palepu A, Tyndall MW, Joy R, Kerr T, Wood E, Press N, Hogg RS, Montaner JSG: Antiretroviral adherence and HIV treatment outcomes among HIV/HCV co-infected injection drug users: The role of methadone maintenance therapy.
Altice FL, Maru DS-R, Bruce RD, Springer SA, Friedland GH: Superiority of directly administered antiretroviral therapy over self-administered therapy among HIV-infected drug users: a prospective, randomized, controlled trial.
Berg KM, Litwin A, Li X, Heo M, Arnsten JH: Directly observed antiretroviral therapy improves adherence and viral load in drug users attending methadone maintenance clinics: A randomized controlled trial.
Liu H, Wilson IB, Goggin K, Reynolds NR, Simoni JM, Golin CE, Rosen MI, Gross R, Wagner G, Schneiderman N, Remien RH, Erlen JA, Arnsten JH, Bangsberg DR, or the MACH14 Group MACH14: A Multi-site Collaboration on ART Adherence among 14 Institutions.
Wagner GJ, Kanouse DE, Golinelli D, Miller LG, Daar ES, Witt MD, Diamond C, Tilles JG, Kemper CA, Larsen R, Goicoechea M, Haubrich RH: Cognitive-behavioral intervention to enhance adherence to antiretroviral therapy: a randomized controlled trial (CCTG 578).
Kushel MB, Colfax G, Ragland K, Heineman A, Palacio H, Bangsberg DR: Case management is associated with improved antiretroviral adherence and CD4(+) cell counts in homeless and marginally housed individuals with HIV infection.
AIDS Behav 2012. Publisher Full Text
AIDS Educ Prev 1995, 7:429-442. PubMed Abstract
Remien RH, Hirky AE, Johnson MO, Weinhardt LS, Whittier D, Le GM: Adherence to Medication Treatment: A Qualitative Study of Facilitators and Barriers Among a Diverse Sample of HIV+ Men and Women in Four U.S. Cities.
Am J Addict 2003, 12:455-458. PubMed Abstract