Abstract

While numerous studies have established the adverse independent effects of clinical conditions including neurocognitive dysfunction, psychiatric illness, and substance abuse/dependence on medication adherence among HIV-infected adults, fewer have studied their interactive effects. The current study examined this issue among 204 HIV-infected participants based upon current neurocognitive functioning and DSM-IV-diagnosed psychiatric illness and current substance abuse or dependence. Results confirmed that participants with any of these risk factors demonstrated poorer adherence than individuals with no risk factors. A neurocognitive status × substance abuse/dependence interaction was also identified such that participants with impaired neurocognition and a co-occurring substance abuse/dependence diagnosis demonstrated the poorest adherence. Results confirm the deleterious impact of these risk factors in isolation and also identify a specific interactive effect for individuals with comorbid neurocognitive impairment and a substance abuse/dependence disorder. Findings highlight the need for interventions that simultaneously address these problems.

Introduction

Adherence to highly active antiretroviral therapy (HAART) remains a critical clinical issue in HIV treatment, as suboptimal adherence rates can decrease treatment efficacy and increase risk of adverse patient outcome (Bangsberg, 2006; Lima et al., 2007). It is well established that demographic factors including younger age (Hinkin et al., 2004), fewer years of education (Bottonari et al., 2012), low health literacy (Wolf et al., 2007), lower socio-economic status (Golin et al., 2002), and ethnic minority status (Hatzenbuehler, Nolen-Hoeksema, & Erickson, 2008; Thames et al., 2012) are associated with lower medication adherence rates. Additionally, studies have demonstrated that clinical conditions including neurocognitive impairment, psychiatric illness, and current substance abuse/dependence also have an impact on adherence (Becker, Thames, Woo, Castellon, & Hinkin, 2011; Ettenhofer, Foley, Castellon, & Hinkin, 2010; Hinkin et al., 2002, 2004, 2007; Hendershot, Stoner, Pantalone, & Simoni, 2009; Panos et al., 2013; Thames et al., 2012; Woods, Moore, Weber, & Grant, 2009).

Studies focused on neurocognition have demonstrated that cognitively impaired individuals exhibit a significantly greater risk of poorer adherence compared with cognitively intact adults (Hinkin et al., 2002; Woods et al., 2008). A number of studies have also identified that psychiatric illness, spanning the spectrum from apathy to depression to bipolar disorder, also adversely affects adherence rates among HIV-infected adults (Moore, Posada, et al., 2012; Safren et al., 2001; Thames et al., 2011). With regard to illicit drug use, active substance abuse and dependence have a significant negative impact on adherence, particularly when stimulants are involved (Hinkin et al., 2007; Moore, Blackstone, et al., 2012). Alcohol abuse and dependence have also been linked to poor adherence (Berg, Michelson, & Safren, 2007). These and several other studies conclusively identify that neurocognitive impairment, psychiatric illness, and substance abuse/dependence significantly interfere with treating HIV/AIDS and are important targets for intervention in treatment protocols.

While there is considerable evidence that each of these factors, in isolation, may affect medication adherence in HIV-infected individuals, few studies have examined and compared the concurrent presence of multiple risk factors. However, adherence behaviors are affected by multiple different demographic, psychosocial, and diagnostic factors that in turn influence each other (Ickovics & Meade, 2002). For example, psychiatric illnesses often trigger increased substance use behaviors and also can interfere with cognitive functioning (Porter, Gallagher, Thompson, & Young, 2003; Swendsen & Merikangas, 2000). At the same time, individuals who abuse substances are at an increased risk of developing psychiatric illnesses and cognitive problems over time (Scott et al., 2007; Swendsen & Merikangas, 2000).

Given that many HIV-positive patients exhibit multiple comorbidities, studies examining a range of concurrent risk factors are needed. Particular combinations of co-morbid risk factors may present a greater impediment to adequate adherence than do others. There is evidence that certain drugs (e.g., cocaine) may perturb the function of the blood–brain barrier and influence the progression to AIDS in HIV-infected individuals. Moreover, HIV-infected individuals who abuse substances, including stimulants and alcohol, are at an increased risk of significant neurocognitive impairment (Rippeth et al., 2004; Rothlind et al., 2005). There is also evidence that psychiatric illnesses share similar neurochemical alterations with drug (particularly stimulant) dependence, including alterations in the function of serotonin, dopamine, and peptide systems. Therefore, there is a neurobiological basis for potential additive or interactive effects between psychiatric disease, substance abuse/dependence, and neurocognitive impairment. In line with this, one recent paper reported that HIV-infected methamphetamine users demonstrated poorer self-reported medication adherence when comorbid neurocognitive impairment was present (Moore, Blackstone, et al., 2012) providing support for the aggregate influence of comorbid drug use and compromised neurocognition on adherence. Their paper also reported that a lifetime history of depression independently predicted poorer adherence.

Additional studies on how varying combinations of these risk factors impact adherence over time are warranted. The aims of the current study were to objectively determine medication adherence rates among individuals with single and comorbid risk factors that are diagnostically relevant in a clinical setting (i.e., neurocognitive impairment, psychiatric illness, and substance abuse/dependence). We first hypothesized that these factors would individually predict poorer adherence. We also hypothesized that individuals with two risk factors would exhibit poorer adherence than individuals with one risk factor. However, some risk factor combinations, such as substance abuse/dependence and neurocognitive impairment, might have a more pronounced impact than other combinations. We therefore ran additional analyses to determine if specific combinations of risk factors particularly associate with poor adherence.

Methods

Participants

This sample included 204 HIV-infected adults recruited from HIV treatment facilities throughout the greater Los Angeles area. Participants were on average 41.8 years of age (SD = 7.3) and completed 13.0 years of education (SD = 2.0). The majority were male (83.1%) and African American (64.2%). Approximately half (52.5%) met diagnostic criteria for AIDS. All participants were on HAART regimens throughout the study. This work was conducted with the approval of the local institutional review board for protection of human subjects (see Table 1 for complete demographic and clinical information).

Table 1.

Demographic and clinical characteristics of the sample

Demographic 
 Age in years (mean) 41.8 (SD = 7.3) 
 Education in years (mean) 13.0 (SD = 2.0) 
 NAART verbal IQ (mean) 104.8 (SD = 9.5) 
 %Male 83.3 (n = 170) 
Ethnicity 
 %African American 64.3 (n = 131) 
 %Hispanic/Latino 14.2 (n = 29) 
 %Caucasian 15.2 (n = 31) 
 %Asian 2.9 (n = 6) 
 %Native American 0.5 (n = 1) 
 %Multiracial 2.9 (n = 6) 
Medical 
 Most recent CD4 count (median) 189.1 (interquartile range = 418) 
 Lowest CD4 count to date (median) 372.5 (interquartile range = 286) 
 %Current neurocognitive impairment 29.9 (n = 61) 
Neurological 
 %Seizure 7.8 (n = 16) 
 %Stroke 5.9 (n = 12) 
 %Head injury LoC > 60 9.3 (n = 19) 
Current psychiatric illness 21.6 (n = 44) 
 %Current major depressive disorder 15.2 (n = 31) 
 %Current mania or bipolar disorder 1.0 (n = 2) 
 %Current psychosis 9.3 (n = 19) 
Current substance use/dependence 41.2 (n = 84) 
 %Alcohol use 13.2 (n = 27) 
 %Any drug use 36.4 (n = 74) 
 %Cocaine use 27.5 (n = 56) 
 %Amphetamine use 7.4 (n = 15) 
 %Opiate use 2.9 (n = 6) 
 %Cannabis use 10.3 (n = 21) 
 %Sedative use 1.5 (n = 3) 
Comorbid conditions 
 %Neuro + substance 9.8 (n = 20) 
 %Neuro + psych 4.9 (n = 10) 
 %Psych + substance 9.3 (n = 19) 
 %Neuro + psych + substance 1.4 (n = 3) 
Demographic 
 Age in years (mean) 41.8 (SD = 7.3) 
 Education in years (mean) 13.0 (SD = 2.0) 
 NAART verbal IQ (mean) 104.8 (SD = 9.5) 
 %Male 83.3 (n = 170) 
Ethnicity 
 %African American 64.3 (n = 131) 
 %Hispanic/Latino 14.2 (n = 29) 
 %Caucasian 15.2 (n = 31) 
 %Asian 2.9 (n = 6) 
 %Native American 0.5 (n = 1) 
 %Multiracial 2.9 (n = 6) 
Medical 
 Most recent CD4 count (median) 189.1 (interquartile range = 418) 
 Lowest CD4 count to date (median) 372.5 (interquartile range = 286) 
 %Current neurocognitive impairment 29.9 (n = 61) 
Neurological 
 %Seizure 7.8 (n = 16) 
 %Stroke 5.9 (n = 12) 
 %Head injury LoC > 60 9.3 (n = 19) 
Current psychiatric illness 21.6 (n = 44) 
 %Current major depressive disorder 15.2 (n = 31) 
 %Current mania or bipolar disorder 1.0 (n = 2) 
 %Current psychosis 9.3 (n = 19) 
Current substance use/dependence 41.2 (n = 84) 
 %Alcohol use 13.2 (n = 27) 
 %Any drug use 36.4 (n = 74) 
 %Cocaine use 27.5 (n = 56) 
 %Amphetamine use 7.4 (n = 15) 
 %Opiate use 2.9 (n = 6) 
 %Cannabis use 10.3 (n = 21) 
 %Sedative use 1.5 (n = 3) 
Comorbid conditions 
 %Neuro + substance 9.8 (n = 20) 
 %Neuro + psych 4.9 (n = 10) 
 %Psych + substance 9.3 (n = 19) 
 %Neuro + psych + substance 1.4 (n = 3) 

Note: NAART = North American Adult Reading Test; LoC = loss of consciousness.

Procedure

After providing informed consent, participants completed a demographic questionnaire and structured clinical interview, and then a battery of neuropsychological tests. Participants were prescribed self-administered HAART throughout the course of the study. Trained psychometrists conducted all testing under supervision of a board-certified neuropsychologist while psychiatric interviewing was conducted under the supervision of a licensed clinical psychologist. Participants were instructed how to use the medication adherence monitoring system, as described below. Participants were then scheduled to return at 1-month intervals over 6 months after the baseline testing.

Participants were excluded if they had an adherence rate of <5% at the 1-month follow-up (n = 5), resulting in a final sample of 204 participants. All procedures were approved by local institutional review board panels.

Measures

Medication adherence

Medication adherence was measured through both self-report and Medication Event Monitoring System (MEMS) measures (see Hinkin et al., 2002, 2004, 2007 for additional details on MEMS cap procedures). Consistent with previous studies (Hinkin et al., 2002, 2004), a single antiviral medication per participant was selected for both MEMS monitoring and self-report. Participants were informed to only open the MEMS cap while taking a dose and to refill the bottle at a time they ordinarily took a dose. They were explicitly instructed to not use pill organizers or rely on pocket dosing. At the 1-month follow-up visits, data from the MEMS caps were downloaded. Adherence rate was calculated as the percent of doses taken relative to total doses prescribed. A proportion of the sample (n = 31; 13.4%) had 1 or 2 months of missing MEMS data over the 6-month period. To include them in the main analyses, missing data were imputed by averaging the adherence rates of the months before and after the missing data point. Missing data at the 6-month point was imputed by reentering data from the fifth month point. Participants with 3 or more missing months of MEMS data and those missing two consecutive months of MEMS data were excluded from the study.

Neuropsychological Assessment

Participants completed a neuropsychological test battery (see Table 2 for a list of tests employed). Deficit scores for each variable were computed in the manner developed by Heaton et al. (2004) in which deficit scores were assigned to T-scores as follows: T > 39 = 0; 39 ≥ T ≥ 35 = 1; 34 ≥ T ≥ 30 = 2; 29 ≥ T ≥ 25 = 3; 24 ≥ T ≥ 20 = 4; T < 20 = 5. A global deficit score (GDS) was then calculated by averaging deficit scores into a single score (see Heaton et al., 2004 for a detailed description of this approach). To maximize specificity and ensure that participants in the cognitively impaired group demonstrated significant neurocognitive impairment, we employed a conservative cut-point of 1.0. The GDS score was dichotomized into impaired (GDS ≥ 1.0) and unimpaired (GDS < 1.0) categories. Sixty-one participants (29.9%) were classified as neurocognitively impaired.

Table 2.

Neuropsychological measures and normative data

Domain Test name Variable Normative data 
Information processing speed WAIS-III Digit symbol Taylor and Heaton (2001
 WAIS-III Symbol search Taylor and Heaton (2001
 TMT Part A Heaton et al. (2004
Learning and memory CVLT-II Trials 1–5 total Test Manual 
 CVLT-II LDFR trial Test Manual 
Attention WAIS-III Letter–number Taylor and Heaton (2001
 WAIS-III Digit span Taylor and Heaton (2001
 PASAT Series 1 Diehr et al. (2003
Verbal fluency COWAT FAS Heaton et al. (2004
Executive functioning TMT Part B Heaton et al. (2004
 Stroop Interference trial Golden (1976) 
 WCST-64 Perseverative errors Test Manual 
Motor functioning GPT Dominant hand Heaton et al. (2004
 GPT Nondominant hand Heaton et al. (2004
Domain Test name Variable Normative data 
Information processing speed WAIS-III Digit symbol Taylor and Heaton (2001
 WAIS-III Symbol search Taylor and Heaton (2001
 TMT Part A Heaton et al. (2004
Learning and memory CVLT-II Trials 1–5 total Test Manual 
 CVLT-II LDFR trial Test Manual 
Attention WAIS-III Letter–number Taylor and Heaton (2001
 WAIS-III Digit span Taylor and Heaton (2001
 PASAT Series 1 Diehr et al. (2003
Verbal fluency COWAT FAS Heaton et al. (2004
Executive functioning TMT Part B Heaton et al. (2004
 Stroop Interference trial Golden (1976) 
 WCST-64 Perseverative errors Test Manual 
Motor functioning GPT Dominant hand Heaton et al. (2004
 GPT Nondominant hand Heaton et al. (2004

Note: WAIS-III = Wechsler Adult Intelligence Scale-Third Edition; TMT = Trail Making Test; CVLT-II = California Verbal Learning Test Second Edition; LDFR = Long Delay Free Recall; PASAT = Paced Auditory Serial Addition Test; COWAT = Controlled Oral Word Association Test; WCST = Wisconsin Card Sorting Test; GPT = Grooved Pegboard Test.

Substance Abuse/Dependence and Psychiatric Status

Psychiatric status was assessed through a modified version of the mood and psychotic spectrum modules from the Structured Clinical Interview for the DSM-IV (SCID-IV). Participants who met criteria for a current major depressive episode, bipolar disorder, manic episode, or psychotic episode were classified as having a psychiatric disorder. Current drug and alcohol abuse/dependence were also assessed with the SCID-IV. Drugs assessed included cocaine, amphetamines, opiates, cannabis, and sedatives. Individuals who met DSM-IV criteria for current abuse or dependence were classified as having a substance abuse/dependence disorder. Using the SCID-IV, we identified 44 participants (21.6%) with a current psychiatric condition and 84 (41.2%) participants with current substance abuse or dependence. Regarding comorbid risk factors, 20 participants (9.8%) had both neurocognitive impairment and current substance abuse/dependence, 10 participants (4.9%) had comorbid neurocognitive impairment and psychiatric illness, and 19 participants (9.3%) had current psychiatric illness and substance abuse/dependence.

Data Analysis

First, the Jonckheere–Terpstra (J–T) test of incremental risk was run to confirm that two risk factors was worse than one risk factor, which was in turn worse than no risk factors on overall adherence. Next, the key analysis examined the main and interactive effects of risk factors on overall medication adherence using a three-way ANCOVA. We controlled for age, years of education, ethnicity, and comorbid neurological illness, as prior studies have reported that these variables can affect adherence (Hinkin et al., 2002, 2004; Kalichman, Ramachandran, & Catz, 1999; Thames et al., 2012). Comorbid medical conditions that were controlled for included active seizure disorder and history of stroke or head trauma resulting in a loss of consciousness exceeding 60 min. Significant two-way interaction effects were plotted and interpreted. We did not interpret the three-way interaction given the small number of participants with all three conditions (n = 3).

Results

Jonckheere–Terpstra Analysis

The J–T test for all three groups was significant (J–T statistic = −2.9, p = .004). Mean adherence for no risk factors was 77.3% (SD = 17.5), for one risk factor was 64.4% (SD = 27.0), and for two risk factors was 62.8% (SD = 28.0). Based on these means, it is evident that the no risk factor group had the best adherence. Contrary to expectations, the J–T test between one and two risk factors was not significant (J–T statistic = −0.2, p = .855), indicating that two risk factors did not contribute to poorer adherence than one risk factor.

Main Analysis

ANCOVA was run to examine adherence levels in individuals with different combinations of the risk factors. The ANCOVA identified main effects in the expected direction (see Table 3). The substance abuse/dependence main effect was significant, F(1, 192) = 4.1, p = .044, partial η² = 0.021, as was the psychiatric illness variable, F(1, 192) = 4.5, p = .036, partial η² = 0.023. There was a trend toward significance for the neurocognitive status variable, F(1, 192) = 2.8, p = .094, partial η² = 0.015. However, when covariates were removed, neurocognitive status was significant, F(1, 196) = 4.4, p = .038, partial η² = 0.022.

Table 3.

Summary of results

 F p η2 
Covariates 
 Age 3.50 .063 0.018 
 Ethnicity 7.09 .008 0.036 
 Years of education 4.11 .044 0.021 
 Neurological illness 2.35 .127 0.012 
Main effects 
 Neurocognitive status 2.83 .094 0.015 
 Substance abuse/dependence 4.10 .044 0.021 
 Psychiatric illness 4.45 .036 0.023 
Two-way interactions 
 Neurocognitive status by substance abuse/dependence 7.87 .006 0.039 
 Neurocognitive status by psychiatric illness 0.21 .646 0.001 
 Psychiatric illness by substance abuse/dependence 0.18 .674 0.001 
 F p η2 
Covariates 
 Age 3.50 .063 0.018 
 Ethnicity 7.09 .008 0.036 
 Years of education 4.11 .044 0.021 
 Neurological illness 2.35 .127 0.012 
Main effects 
 Neurocognitive status 2.83 .094 0.015 
 Substance abuse/dependence 4.10 .044 0.021 
 Psychiatric illness 4.45 .036 0.023 
Two-way interactions 
 Neurocognitive status by substance abuse/dependence 7.87 .006 0.039 
 Neurocognitive status by psychiatric illness 0.21 .646 0.001 
 Psychiatric illness by substance abuse/dependence 0.18 .674 0.001 

We next examined the ANCOVA for two-way interaction effects. A neurocognitive status by substance abuse/dependence interaction effect was identified, F(1, 192) = 7.9, p = .006, partial η² = 0.039, such that participants who had comorbid neurocognitive impairment and substance abuse/dependence exhibited disproportionately poorer adherence than either condition in isolation (see Fig. 1). There was no neurocognitive status × psychiatric illness interaction effect, F(1, 192) = 0.2, p = .646, and no psychiatric illness × substance abuse/dependence interaction effect, F(1, 192) = 0.2, p = .674.

Fig. 1.

Substance use × neurocognitive status interaction effect. Note: MEMS = Medication Event Monitoring System. The y-axis has been truncated for effect.

Fig. 1.

Substance use × neurocognitive status interaction effect. Note: MEMS = Medication Event Monitoring System. The y-axis has been truncated for effect.

Discussion

Neurocognitive impairment, substance abuse/dependence, and psychiatric illness have all been associated with lower rates of medication adherence among HIV-infected adults. To date, however, few studies have evaluated the aggregate impact of varying combinations of these three major risk factors. Our findings address this gap and highlight issues regarding the impact of these risk factors on medication adherence in HIV. Consistent with the literature (Hinkin et al., 2002, 2007; Moore, Blackstone, et al., 2012; Woods et al., 2009), individuals with current cognitive dysfunction, substance abuse, or psychiatric illness demonstrated poorer medication adherence than did those without any of the abovementioned risk factors. Although neurocognitive dysfunction only demonstrated a trend toward significance for this analysis, results were significant when the covariates were taken out. This finding suggests, at least in this sample, that the other variables accounted for more of the variance that explained adherence. Given the broad literature reporting the impact of neurocognitive impairment on adherence (Hinkin et al., 2002; Woods, Moore, et al., 2009), it is premature to conclude that neurocognitive status is less important than substance abuse/dependence and psychiatric illness. Furthermore, and as discussed below, results do highlight the importance of neurocognitive dysfunction in combination with comorbid substance abuse/dependence.

Contrary to our hypothesis, two risk factors did not predict poorer adherence than one risk factor. However, the ANCOVA identified an effect for two risk factors although this effect was masked when the risk factors were aggregated together. The two-way interactions allowed us to examine different combinations of these variables. We did not identify any interactive effects between psychiatric illness and other risk factors. This result is consistent with a previous investigation that reported that lifetime depression contributed to poorer adherence but did not interact with other risk factors (Moore, Blackstone, et al., 2012). We also identified a specific neurocognitive impairment by substance abuse/dependence interaction effect such that individuals with both risk factors had the worst adherence among all the groups. Active substance abuse/dependence, particularly involving stimulants, is a well-established predictor of poor adherence (Arnsten et al., 2002; Hinkin et al., 2004, 2007; Moore, Blackstone, et al., 2012), and this appears to be particularly potent when combined with neurocognitive impairment.

It has previously been reported that HIV-infected individuals with stimulant dependence and neuropsychological impairment exhibit poorer adherence (Moore, Blackstone, et al., 2012). There may be a neurobiological relationship between certain illicit drugs and neurocognition in HIV which in turn further worsens medication adherence. Although the mechanisms behind this relationship remain speculative at this time, both the HIV virus and some drug classes (e.g., methamphetamines) result in increased excitotoxicity, due to both direct mechanisms as well as indirect neuroinflammatory responses to both the virus and drugs (Rippeth et al., 2004). Substance use-related neurocognitive compromise and HIV-associated neurocognitive dysfunction can both contribute to poor adherence. From a behavioral standpoint, it is also likely that individuals who exhibit difficulty remembering and organizing their medication regimes simply have greater difficulty doing so when they are intoxicated. Another interpretation is that poor adherence leads to a worsening of HIV-associated immunocompromise, which in turn results in additional declines in neurocognitive functioning (Ettenhofer et al., 2010).

Treatment implications of these data warrant attention. As the HIV population ages, the prevalence of individuals with neurocognitive impairment is likely to rise, and those who continue to have substance abuse/dependence diagnoses are at an increased risk of poorer medication adherence. This subset of individuals may require more managed care and regular follow-up with health-care professionals to maintain adequate adherence rates. Clinicians would therefore benefit from formally evaluating their patients' neuropsychological status while assessing for frequency of substance use as part of treatment planning. Specific treatments that simultaneously target substance use and adherence, such as motivational interviewing (Safren et al., 2001), may be viable intervention strategies for individuals with substance use disorders. However, when neurocognitive impairment is also present, compensatory strategies that require top-down planning and prospective memory, such as medication monitoring, are less effective (Woods et al., 2008). Such patients may require a more structured managed-care setting where other individuals can assist with their daily regimens. Even less drastic measures, like bolstering social support networks, have been shown to improve adherence (Gonzalez et al., 2004). For those with neurocognitive compromise, an added benefit might be the assistance of family members or companions who can assist with medication adherence. Studies identifying possible treatment combinations for this vulnerable subset of individuals are certainly warranted.

Several limitations bear mention. The psychiatric and substance abuse/dependence variables are aggregates of multiple individual disorders. While grouping together several psychiatric disorders or several drugs of abuse under one factor has benefits with regards to data analysis, it may obscure the unique contribution of specific psychiatric diagnoses or drug of abuse. MEMS caps have their own limitations as well. Use of such caps is cumbersome, leading some participants to cease using them to dispense their medication, particularly over time. It is therefore possible that the data underestimate actual adherence rates. Our sample of individuals who presented with all three risk factors was small, which limited our ability to formally evaluate this group. In addition, only 12 participants exhibited both psychiatric illness and neurocognitive impairment, which may account for the lack of any significant interaction effects. We also do not have information about treatment history for any psychiatric illnesses that may have influenced results. Additional studies on individuals with psychiatric illness and other comorbidities are required before we can determine whether a certain combination of risk factors (e.g., bipolar disorder with substance abuse/dependence) has a greater impact on adherence behaviors than others.

In addition, it should be noted that there was a large proportion of unexplained variance in the models, as several other known adherence risk factors, such as poor health literacy, regime complexity, length of illness, low social support, and poor self-efficacy were not formally evaluated. As such, we cannot comment on how they might have interacted with the measured risk factors. Finally, a few patients had comorbid neurological conditions. While we did statistically control for this variable in the ANCOVA, additional influences of these conditions might still be present, particularly in another sample that might have a greater proportion of neurologically compromised participants. It may also be that some of the observed relationship between medication adherence and risk factors were driven by the neurological illnesses and are not specific to HIV itself.

Our results demonstrate that HIV medication adherence is greatly compromised when neurocognitive impairment is combined with substance abuse/dependence. Future research should establish possible protective mechanisms that may mitigate against negative outcomes in the face of these otherwise deleterious risk factors. This in turn may allow for the development of treatment protocols specifically designed to reduce risk and improve health outcomes for this particularly vulnerable population.

Conflict of Interest

None declared.

Acknowledgements

This work was supported by National Institutes of Health grant R01 DA13799 and the National Institutes of Health Ruth L. Kirschstein National Research Service award T32 MH19535.

References

Arnsten
J. H.
Demas
P. A.
Grant
R. W.
Gourevitch
M. N.
Farzadegan
H.
Howard
A. A.
, et al.  . 
Impact of active drug use on antiretroviral therapy adherence and viral suppression in HIV-infected drug users
Journal of General Internal Medicine
 , 
2002
, vol. 
17
 
5
(pg. 
377
-
381
)
Bangsberg
D. R.
Less than 95% adherence to nonnucleoside reverse-transcriptase inhibitor therapy can lead to viral suppression
Clinical Infectious Diseases
 , 
2006
, vol. 
43
 
7
(pg. 
939
-
941
)
Becker
B. W.
Thames
A. D.
Woo
E.
Castellon
S. A.
Hinkin
C. H.
Longitudinal change in cognitive function and medication adherence in HIV-infected adults
AIDS and Behavior
 , 
2011
, vol. 
15
 
8
(pg. 
1888
-
1894
)
Berg
C. J.
Michelson
S. E.
Safren
S. A.
Behavioral aspects of HIV care: Adherence, depression, substance use, and HIV-transmission behaviors
Infectious Disease Clinics of North America
 , 
2007
, vol. 
21
 
1
(pg. 
181
-
200
)
Bottonari
K. A.
Tripathi
S. P.
Fortney
J. C.
Curran
G.
Rimland
D.
Rodriguez-Barradas
M.
, et al.  . 
Correlates of antiretroviral and antidepressant adherence among depressed HIV-infected patients
AIDS Patient Care and STDs
 , 
2012
, vol. 
26
 
5
(pg. 
265
-
273
)
Diehr
M. C.
Cherner
M.
Wolfson
T. J.
Miller
S. W.
Grant
I.
Heaton
R. K.
HIV Neurobehavioral Research Center, T
The 50 and 100-item short forms of the Paced Auditory Serial Addition Task (PASAT): Demographically corrected norms and comparisons with the full PASAT in normal and clinical samples
Journal of Clinical and Experimental Neuropsychology
 , 
2003
, vol. 
25
 
4
(pg. 
571
-
585
)
Ettenhofer
M. L.
Foley
J.
Castellon
S. A.
Hinkin
C. H.
Reciprocal prediction of medication adherence and neurocognition in HIV/AIDS
Neurology
 , 
2010
, vol. 
74
 
15
(pg. 
1217
-
1222
)
Golden
C. J.
Identification of brain disorders by the Stroop Color and Word Test
Journal of Clinical Psychology
 , 
1976
, vol. 
32
 (pg. 
654
-
658
)
Golin
C. E.
Liu
H.
Hays
R. D.
Miller
L. G.
Beck
C. K.
Ickovics
J.
, et al.  . 
A prospective study of predictors of adherence to combination antiretroviral medication
Journal of General Internal Medicine
 , 
2002
, vol. 
17
 
10
(pg. 
756
-
765
)
Gonzalez
J. S.
Penedo
F. J.
Antoni
M. H.
Durán
R. E.
McPherson-Baker
S.
Ironson
G.
, et al.  . 
Social support, positive states of mind, and HIV treatment adherence in men and women living with HIV/AIDS
Health Psychology
 , 
2004
, vol. 
23
 
4
(pg. 
413
-
418
)
Hatzenbuehler
M. L.
Nolen-Hoeksema
S.
Erickson
S. J.
Minority stress predictors of HIV risk behavior, substance use, and depressive symptoms: Results from a prospective study of bereaved gay men
Health Psychology
 , 
2008
, vol. 
27
 
4
(pg. 
455
-
462
)
Heaton
R. K.
Marcotte
T. D.
Mindt
M. R.
Sadek
J.
Bently
H.
McCutchan
J. A.
, et al.  . 
The impact of HIV-associated neuropsychological impairment on everyday functioning
Journal of the International Neuropsychological Society
 , 
2004
, vol. 
10
 (pg. 
317
-
331
)
Hendershot
C. S.
Stoner
S. A.
Pantalone
D. W.
Simoni
J. M.
Alcohol use and antiretroviral adherence: Review and meta-analysis
Journal of Acquired Immune Deficiency Syndromes (1999)
 , 
2009
, vol. 
52
 
2
(pg. 
180
-
202
)
Hinkin
C. H.
Barclay
T. R.
Castellon
S. A.
Levine
A. J.
Durvasula
R. S.
Marion
S. D.
, et al.  . 
Drug use and medication adherence among HIV-1 infected individuals
AIDS and Behavior
 , 
2007
, vol. 
11
 
2
(pg. 
185
-
194
)
Hinkin
C. H.
Castellon
S. A.
Durvasula
R. S.
Hardy
D. J.
Lam
M. N.
Mason
K. I.
, et al.  . 
Medication adherence among HIV+ adults effects of cognitive dysfunction and regimen complexity
Neurology
 , 
2002
, vol. 
59
 
12
(pg. 
1944
-
1950
)
Hinkin
C. H.
Hardy
D. J.
Mason
K. I.
Castellon
S. A.
Durvasula
R. S.
Lam
M. N.
, et al.  . 
Medication adherence in HIV-infected adults: Effect of patient age, cognitive status, and substance abuse
AIDS
 , 
2004
, vol. 
18
 
Suppl. 1
(pg. 
S19
-
S25
)
Ickovics
J. R.
Meade
C. S.
Adherence to HAART among patients with HIV: Breakthroughs and barriers
AIDS Care
 , 
2002
, vol. 
14
 
3
(pg. 
309
-
318
)
Kalichman
S. C.
Ramachandran
B.
Catz
S.
Adherence to combination antiretroviral therapies in HIV patients of low health literacy
Journal of General Internal Medicine
 , 
1999
, vol. 
14
 
5
(pg. 
267
-
273
)
Lima
V. D.
Geller
J.
Bangsberg
D. R.
Patterson
T. L.
Daniel
M.
Kerr
T.
, et al.  . 
The effect of adherence on the association between depressive symptoms and mortality among HIV-infected individuals first initiating HAART
Aids
 , 
2007
, vol. 
21
 
9
(pg. 
1175
-
1183
)
Moore
D. J.
Blackstone
K.
Woods
S. P.
Ellis
R. J.
Atkinson
J. H.
Heaton
R. K.
, et al.  . 
the HNRC Group and the TMARC Group
Methamphetamine use and neuropsychiatric factors are associated with antiretroviral non-adherence
AIDS Care
 , 
2012
, vol. 
24
 
12
(pg. 
1504
-
1513
)
Moore
D. J.
Posada
C.
Parikh
M.
Arce
M.
Vaida
F.
Riggs
P. K.
, et al.  . 
HIV-infected individuals with co-occurring bipolar disorder evidence poor antiretroviral and psychiatric medication adherence
AIDS and Behavior
 , 
2012
, vol. 
16
 
8
(pg. 
2257
-
2266
)
Panos
S. E.
Del Re
A. C.
Thames
A. D.
Arentsen
T. J.
Patel
S. M.
Castellon
S. A.
, et al.  . 
The impact of neurobehavioral features on medication adherence in HIV: Evidence from longitudinal models
AIDS Care
 , 
2013
, vol. 
26
 
1
(pg. 
79
-
86
)
Porter
R. J.
Gallagher
P.
Thompson
J. M.
Young
A. H.
Neurocognitive impairment in drug-free patients with major depressive disorder
The British Journal of Psychiatry
 , 
2003
, vol. 
182
 
3
(pg. 
214
-
220
)
Rippeth
J. D.
Heaton
R. K.
Carey
C. L.
Marcotte
T. D.
Moore
D. J.
Gonzalez
R.
, et al.  . 
Methamphetamine dependence increases risk of neuropsychological impairment in HIV infected persons
Journal of the International Neuropsychological Society
 , 
2004
, vol. 
10
 
01
(pg. 
1
-
14
)
Rothlind
J. C.
Greenfield
T. M.
Bruce
A. V.
Meyerhoff
D. J.
Flenniken
D. L.
Lindgren
J. A.
, et al.  . 
Heavy alcohol consumption in individuals with HIV infection: Effects on neuropsychological performance
Journal of the International Neuropsychological Society
 , 
2005
, vol. 
11
 
01
(pg. 
70
-
83
)
Safren
S. A.
Otto
W. M.
Worth
J. L.
Salomon
E.
Johnson
W.
Mayer
K.
, et al.  . 
Two strategies to increase adherence to HIV antiretroviral medication: Life-steps and medication monitoring
Behaviour Research and Therapy
 , 
2001
, vol. 
39
 
10
(pg. 
1151
-
1162
)
Scott
J. C.
Woods
S. P.
Matt
G. E.
Meyer
R. A.
Heaton
R. K.
Atkinson
J. H.
, et al.  . 
Neurocognitive effects of methamphetamine: A critical review and meta-analysis
Neuropsychology Review
 , 
2007
, vol. 
17
 
3
(pg. 
275
-
297
)
Swendsen
J. D.
Merikangas
K. R.
The comorbidity of depression and substance use disorders
Clinical Psychology Review
 , 
2000
, vol. 
20
 
2
(pg. 
173
-
189
)
Taylor
M. J.
Heaton
R. K.
Sensitivity and specificity of WAIS-III/WMS-III demographically corrected factor scores in neuropsychological assessment
Journal of the International Neuropsychological Society
 , 
2001
, vol. 
7
 
7
(pg. 
867
-
874
)
Thames
A. D.
Becker
B. W.
Marcotte
T. D.
Hines
L. J.
Foley
J. M.
Ramezani
A.
, et al.  . 
Depression, cognition, and self-appraisal of functional abilities in HIV: An examination of subjective appraisal versus objective performance
The Clinical Neuropsychologist
 , 
2011
, vol. 
25
 
2
(pg. 
224
-
243
)
Thames
A. D.
Moizel
J.
Panos
S. E.
Patel
S. M.
Byrd
D. A.
Myers
H. F.
, et al.  . 
Differential predictors of medication adherence in HIV: Findings from a sample of African American and Caucasian HIV-positive drug-using adults
AIDS Patient Care and STDs
 , 
2012
, vol. 
26
 
10
(pg. 
621
-
630
)
Wolf
M. S.
Davis
T. C.
Osborn
C. Y.
Skripkauskas
S.
Bennett
C. L.
Makoul
G.
Literacy, self-efficacy, and HIV medication adherence
Patient Education and Counseling
 , 
2007
, vol. 
65
 
2
(pg. 
253
-
260
)
Woods
S. P.
Dawson
M. S.
Weber
E.
Gibson
S.
Grant
I.
Atkinson
J. H.
Timing is everything: Antiretroviral nonadherence is associated with impairment in time-based prospective memory
Journal of the International Neuropsychological Society
 , 
2009
, vol. 
15
 
01
(pg. 
42
-
52
)
Woods
S. P.
Moore
D. J.
Weber
E.
Grant
I.
Cognitive neuropsychology of HIV-associated neurocognitive disorders
Neuropsychology Review
 , 
2009
, vol. 
19
 
2
(pg. 
152
-
168
)
Woods
S. P.
Moran
L. M.
Carey
C. L.
Dawson
M. S.
Iudicello
J. E.
Gibson
S.
, et al.  . 
Prospective memory in HIV infection: Is “remembering to remember” a unique predictor of self-reported medication management?
Archives of Clinical Neuropsychology
 , 
2008
, vol. 
23
 
3
(pg. 
257
-
270
)