Differential Effects of Substance Use on HIV Care and Symptom Outcomes: A Longitudinal Analysis Among People with HIV

Abstract

Among people with HIV (PWH), those with suboptimal antiretroviral therapy (ART) adherence face heightened risk for poor clinical outcomes. Substance use is a critical barrier in this population, yet most research treats substance use as a homogeneous risk factor. This study examined the differential effects of specific substances on HIV-related outcomes among PWH with suboptimal ART adherence. Using data from a randomized controlled trial conducted in New York City and Birmingham, AL (N = 300; July 20, 2023 to May 7, 2024), we tested whether baseline use of 10 substance classes predicted ART adherence self-efficacy, self-reported ART adherence, HIV self-management, and symptom distress at 6 and 12 months. Linear regression models were estimated across multiply imputed datasets, adjusting for intervention assignment, site, sociodemographic, and baseline outcomes. Results indicated that baseline cocaine use was significantly associated with lower self-reported adherence at 6 months (β = −5.03, p < 0.050), methamphetamine use predicted poorer HIV self-management at 12 months (β = −2.92, p < 0.050), hallucinogen use predicted lower adherence self-efficacy at 12 months (β = −20.26, p < 0.050), and inhalant use was associated with greater HIV symptom distress at 6 months (β = 8.33, p < 0.050). No significant associations were observed for cannabis, heroin, prescription opioids, stimulants, sedatives, or other substances. Findings highlight the importance of moving beyond generalized models of substance use in HIV care to account for the unique ways different substances affect HIV care and symptom outcomes. Tailoring interventions to substance-specific risk profiles may improve outcomes for PWH.

Introduction

Substance use among people with HIV (PWH) represents a significant global public health concern and poses substantial challenges to HIV care engagement, treatment adherence, and health outcomes.^1–4 The intersection of HIV and substance use creates a complex clinical picture that often results in suboptimal treatment outcomes and increased disease burden.^2,4,5 In the United States, where approximately 1.2 million people live with HIV, the prevalence of drug use disorders is four times higher among PWH compared with the general population.^6,7 This heightened prevalence reflects overlapping risk factors, shared social determinants, and bidirectional influences between HIV and substance use, which collectively contribute to significant health disparities.⁸

Among PWH, individuals with suboptimal antiretroviral therapy (ART) adherence are at particularly high risk for poor health outcomes, including lower rates of viral suppression, immune decline, and HIV-related complications.^9–11 Substance use is one of the most commonly reported barriers to ART adherence among PWH.¹² Multiple studies have demonstrated that PWH who use substances consistently exhibit lower rates of ART adherence compared with nonusers, with certain substances such as methamphetamine associated with up to a threefold increased risk of nonadherence.^13–17 These adherence challenges translate to clinical outcomes, with substance-using PWH demonstrating lower rates of viral suppression and higher CD4 count volatility compared with nonusers.^14,18 Long-term health-related issues include increased opportunistic infections, accelerated disease progression, heightened HIV symptom distress, and greater mortality risk.^9–11 For PWH who already exhibit suboptimal ART adherence, substance use may pose an even more potent barrier to treatment success, making it critical to understand how specific substances affect HIV-related outcomes in this vulnerable subgroup.

A key limitation in the current literature on substance use and HIV outcomes is the tendency to treat substance use as a homogeneous risk factor, rather than examining the effects of specific substances. Although some studies have investigated individual substances—particularly alcohol, cannabis, heroin, and cocaine—findings are inconsistent. For example, alcohol use has been consistently linked to ART nonadherence across multiple studies.^19,20 Heroin and cocaine have also been associated with reduced ART adherence.^20,21 In contrast, findings on cannabis are mixed, with some studies reporting suboptimal adherence among cannabis users and others finding no significant association.^21–23 Evidence on methamphetamine is similarly limited and primarily focused on sexual minority men, restricting generalizability.^24,25 Few studies have systematically compared how different substances influence various domains of HIV care, such as medication adherence, self-management, and symptom burden. Yet the pharmacological effects, behavioral patterns, and social contexts of substance use differ widely across drug classes—factors that likely shape their distinct impacts on the cognitive and motivational processes that underlie HIV self-care.²⁶

Furthermore, most research examining substance use and HIV outcomes has employed cross-sectional designs that cannot establish temporal relationships or capture the potentially evolving impact of substance use on HIV management over time. Among the longitudinal studies available, few have examined a comprehensive range of substances or considered multiple HIV-related outcomes. This knowledge gap limits the development of tailored interventions that address the specific needs of PWH who use different substances. As health care systems increasingly move toward precision medicine approaches, understanding substance-specific risks becomes essential for optimizing HIV treatment outcomes.

Given these theoretical and clinical considerations, it is plausible that the effects of substance use on HIV-related outcomes may differ by the specific substance used. This study, therefore, focused on substance-specific associations rather than treating substance use as a uniform risk factor and aimed to provide insights into the differential impacts of various substances on HIV self-management, adherence, and symptom experience. This study is a secondary data analysis of a randomized controlled trial (RCT) that evaluated a behavioral intervention for PWH with suboptimal ART adherence in two US cities disproportionately affected by the HIV epidemic—New York City and Birmingham, AL. These diverse contexts (e.g., large urban and less urban, northern and southern), which have distinct HIV epidemiological profiles and substance use patterns, provide an opportunity to examine these relationships across different settings.

Specifically, we examined whether baseline use of distinct substance classes—including cannabis, cocaine, prescription stimulants, methamphetamine, inhalants, sedatives or sleeping pills, hallucinogens, street opioids, prescription opioids, and other drugs—predicted HIV-related outcomes at 6 and 12 months. Outcomes included ART adherence self-efficacy, self-reported ART adherence, HIV self-management, and symptom distress. Findings can inform the development of more nuanced risk assessments and integrated behavioral health interventions that address the challenges faced by PWH who use different substances.

Methods

Data and sample

Data were drawn from a multisite RCT (ClinicalTrials.gov ID NCT05938413) conducted among PWH recruited through purposive sampling from HIV clinical and community-based settings in New York City and Birmingham, AL (July 20, 2023 to May 7, 2024). Eligibility criteria included age 18 years or older, English or Spanish speaking, having been diagnosed with HIV ≥6 months ago, having an HIV-1 RNA level >200 copies/mL, and ART adherence <80% as measured through the single-item self-rating adherence scale. Individuals were excluded if they resided in a nursing home, prison, and/or were receiving inpatient psychiatric care at the time of enrollment, had terminal illness with a life expectancy of <6 months, or planned to move out of the area in the next 12 months.²⁷ The final analytical sample included 300 participants. All study procedures were reviewed and approved by the Institutional Review Boards at the participating institutions. Written informed consent was obtained from all participants before enrollment.

Measures

All study measures were available in both English and Spanish and were administered in Spanish for four participants who were Spanish-speaking only. The instruments used have demonstrated strong validity and reliability in Spanish-speaking populations.^28–30

Substance use

Baseline substance use was measured with the World Health Organization’s Alcohol, Smoking, and Substance Involvement Screening Test (ASSIST), covering 10 substances, including: cannabis, cocaine, prescription stimulants, methamphetamine, inhalants, sedatives or sleeping pills, hallucinogens, street opioids/heroin, prescription opioids, and other drugs. It assesses the frequency of use in the past 3 months for each substance (0 = never, 1 = once or twice, 2 = monthly, 3 = weekly, 4 = daily or almost daily). It has demonstrated strong reliability and validity across diverse populations, including PWH.^31–33

Adherence self-efficacy

The HIV treatment adherence self-efficacy scale (ASES) is a validated 12-item instrument designed to measure patients’ confidence in managing challenges related to ART adherence or in taking medications in various situations. The scale uses a 10-point response format ranging from 1 (cannot do it at all) to 10 (completely certain can do it), with higher scores indicating greater self-efficacy. The ASES demonstrated excellent psychometric properties with robust internal consistency (ρs > 0.90) and good test–retest reliability at both 3 months (rs > 0.70) and 15 months (rs > 0.40).^34,35

HIV self-management

The HIV self-management scale is a validated 20-item instrument developed to assess the self-management behaviors of PWH. The scale captures three key domains: daily health practices, social support for self-management, and managing the chronic nature of HIV. Each item is rated on a 4-point Likert scale (0 = not applicable, 1 = none of the time, 2 = some of the time, 3 = all of the time), with higher scores indicating greater self-management ability. The scale has demonstrated good reliability (Cronbach’s alpha for subscales ranging from 0.71 to 0.86) and validity in diverse populations, including women with HIV and international cohorts.^36–38

Self-reported adherence

The single-item visual analogue scale (VAS) asked participants to estimate the percentage of prescribed HIV medications taken over the past 30 days by marking a point on a continuous line, ranging from 0% to 100%.³⁹ Research has shown that the VAS significantly correlates with other measures of ART adherence, including the more resource-intensive Medication Event Monitoring System caps and pharmacy refill data, with correlation coefficients ranging from 0.34 to 0.40.⁴⁰ Despite its brevity, the VAS has been shown to be a valid, inexpensive, and minimally intrusive method for assessing medication adherence in both research and clinical settings, and can be effectively administered in multiple languages.⁴¹

HIV symptom distress

The HIV symptom distress module (SDM) is a 20-item self-report instrument designed to assess the distress associated with common HIV-related symptoms, such as fatigue, pain, depression, and sleep problems. Each symptom is rated on a five-point scale from 0 (“No, I did not have this symptom”) to 5 (“Yes, I have this symptom. It bothered me a lot”), with higher scores reflecting greater symptom burden. The SDM has demonstrated strong psychometric properties, including high reliability (α = 0.92), and its validity is supported by significant associations with clinical markers, such as HIV viral load and CD4 count.⁴²

Sociodemographic variables

Variables included age (in years), gender identity (female, male, transgender/queer/other), sex assigned at birth (female, male, intersex), race (Black/African American, White), sexual orientation (homosexual/gay/lesbian, heterosexual/straight, bisexual, other), education level (less than high school, high school diploma, some college/associate degree, college or more), and annual income (<$10,000, $10,000–$19,999, >$20,000).

Data analysis

Analyses were conducted using R Studio.⁴³ Descriptive statistics were used to characterize the sample. Multiple imputation by chained equations (MICE) was used to handle missing data.⁴⁴ A total of five imputed datasets were generated using predictive mean matching, with imputation models incorporating all study variables, including demographics, site, intervention assignment, baseline predictors, and outcomes at each time point.⁴⁵ Imputation convergence was verified graphically and by inspecting logged events. Pooled estimates were obtained using the with() and pool() functions from the MICE package. Linear regression models were used for all outcomes (6- and 12-month ART adherence self-efficacy, self-reported ART adherence, self-management, and symptom distress). All regression models included the 10 substance variables simultaneously as predictors to adjust for potential polysubstance use and isolate the unique association of each substance with HIV outcomes. All models were adjusted for intervention assignment, site, age, gender identity, sex assigned at birth, race, sexual orientation, education level, annual income, and baseline outcome values. Intervention assignment (intervention vs. control) was included as a covariate in all regression models to adjust for potential intervention effects on outcomes.

Results

Table 1 presents the sample baseline characteristics. The analytic sample consisted of 300 participants with a mean age of 48.1 years [standard deviation (SD): ±11.7]. The sample was predominantly non-Hispanic Black (69.9%), with smaller proportions identifying as non-Hispanic White (9.0%), non-Hispanic Other (7.7%), Hispanic Black (5.7%), Hispanic White (5.0%), and Hispanic other race (1.7%). Nearly half identified as male (48.7%), 44.3% as female, and 6.7% as transgender, queer, or another gender identity. Most participants (51.0%) identified as heterosexual, with 32.7% identifying as homosexual and 10.3% as bisexual. Educational attainment varied, with 28% having less than a high school education and 34.3% reporting some college or an associate degree. Over half (51.0%) reported annual incomes under $10,000. Participants were recruited from New York City (48.3%) and Birmingham, AL (51.7%), with equal representation across intervention and control groups.

Table 1.

Sample Baseline Characteristics (N = 300)

	Mean/N	SD/%
Age	48.1	11.7
Gender identity
Female	133	44.3%
Male	146	48.7%
Transgender/queer/other	20	6.7%
Sex assigned at birth
Female	134	44.7%
Male	162	54%
Intersex	1	0.3%
Race/ethnicity
Non-Hispanic Black	207	69.9%
Non-Hispanic White	27	9.0%
Non-Hispanic other	23	7.7%
Hispanic Black	17	5.7%
Hispanic White	15	5.0%
Hispanic other	5	1.7%
Sexual orientation
Bisexual	31	10.3%
Heterosexual	153	51.0%
Homosexual	98	32.7%
Other	14	4.7%
Education
Less than high school	84	28.0%
High school	88	29.3%
Some college/associate degree	103	34.3%
College or more	24	8.0%
Annual income
Less than $10,000	153	51.0%
$10,000–$19,999	55	18.3%
$20,000 or more	54	18.0%
Study site
Birmingham, AL	155	51.7%
New York City, NY	145	48.3%
Intervention group
Control	150	50.0%
Intervention	150	50.0%
Adherence self-efficacy	88.17	30.16
HIV self-management	43.45	11.30
Self-reported adherence—VAS	77.92	28.20
Self-reported adherence—SRSI
Very poor	13	4.3%
Poor	27	9.0%
Fair	54	18.0%
Good	50	16.7%
Very good	68	22.7%
Excellent	88	29.3%
HIV symptom distress	19.33	16.99
ASSIST score (% moderate–high risk)
Cannabis	23.8%	—
Cocaine	2.0%	—
Stimulant	2.0%	—
Methamphetamine	1.0%	—
Inhalants	1.0%	—
Sedative	6.1%	—
Hallucinogens	0.3%	—
Heroin	0.3%	—
Prescription opioids	3.1%	—
Other drug	2.2%	—

ASSIST, Alcohol, Smoking, and Substance Involvement Screening Test; SD, standard deviation; VAS, visual analogue scale; SRSI, Single-Item Self-Rating Adherence Scale.

At baseline, cannabis was the most commonly reported substance at moderate-to-high risk levels, with 23.8% of participants screening above the clinical threshold on the ASSIST. Other substances with notable moderate-to-high risk prevalence included sedatives (6.1%), prescription opioids (3.1%), and other drugs (2.2%). At the same time, rates for cocaine, stimulants, methamphetamine, inhalants, hallucinogens, and heroin each remained at or below 2.0%. Participants also reported moderate levels of HIV-related functioning. The mean ART adherence self-efficacy score was 88.17 (SD: ±30.16), HIV self-management was 43.45 (SD: ±11.30), self-reported ART adherence averaged 77.92 (SD: ±28.20), and HIV symptom distress averaged 19.33 (SD: ±16.99).

Table 2 shows the association of different substance use with 6- and 12-month HIV outcomes. Hallucinogen use at baseline predicted lower ART adherence self-efficacy at 12 months (β = −20.26, p < 0.050). Methamphetamine use predicted poorer HIV self-management at 12 months (β = −2.92, p < 0.050). Baseline cocaine use was significantly associated with lower self-reported medication adherence at 6 months (β = −5.03, p < 0.050). Additionally, inhalant use was associated with greater HIV symptom distress at 6 months (β = 8.33, p < 0.050).

Table 2.

Substance Use’s Association with 6- and 12-Month HIV Outcomes (N = 300)

	6-month		12-month
	Estimate (SE)	p	Estimate (SE)	p
Adherence self-efficacy
Cannabis	−0.65 (1.09)	0.551	0.44 (1.08)	0.685
Cocaine	1.81 (2.14)	0.398	2.07 (2.17)	0.340
Stimulant	−5.83 (3.22)	0.071	2.95 (3.47)	0.398
Methamphetamine	−1.45 (2.52)	0.565	0.49 (3.27)	0.883
Inhalants	−7.64 (6.18)	0.24	1.04 (5.96)	0.864
Sedative	0.52 (1.94)	0.793	−0.11 (2.06)	0.958
Hallucinogens	10.90 (8.25)	0.189	−20.26* (9.37)	0.039
Heroin	4.87 (10.12)	0.633	14.73 (10.29)	0.160
Prescription opioids	0.06 (2.35)	0.979	−1.39 (2.84)	0.630
Other drug	−2.29 (2.97)	0.443	−1.99 (3.01)	0.510
HIV self-management
Cannabis	0.35 (0.41)	0.387	−0.55 (0.44)	0.219
Cocaine	1.58 (1.02)	0.135	−0.08 (0.87)	0.924
Stimulant	−0.50 (1.30)	0.704	−0.92 (1.44)	0.528
Methamphetamine	−1.05 (1.14)	0.360	−2.92* (1.24)	0.025
Inhalants	−1.72 (2.00)	0.394	1.16 (1.19)	0.546
Sedative	−0.52 (0.80)	0.518	0.19 (0.72)	0.791
Hallucinogens	0.38 (4.49)	0.933	−0.35 (6.03)	0.956
Heroin	−2.26 (4.95)	0.654	3.76 (5.88)	0.538
Prescription opioids	−0.01 (1.06)	0.993	0.04 (1.07)	0.967
Other drug	−1.03 (1.30)	0.433	0.18 (1.18)	0.881
Self-reported adherence
Cannabis	0.99 (0.96)	0.306	1.80 (0.94)	0.057
Cocaine	−5.03* (2.10)	0.018	2.71 (2.21)	0.224
Stimulant	−1.92 (3.17)	0.545	2.23 (3.25)	0.496
Methamphetamine	0.24 (3.21)	0.942	2.21 (2.71)	0.420
Inhalants	4.12 (4.76)	0.392	0.26 (6.13)	0.967
Sedative	0.15 (1.58)	0.925	0.39 (1.95)	0.845
Hallucinogens	−14.59 (8.08)	0.074	−7.32 (14.19)	0.623
Heroin	2.41 (10.17)	0.814	0.70 (11.87)	0.954
Prescription opioids	0.46 (2.31)	0.843	−0.25 (2.57)	0.924
Other drug	−0.78 (2.70)	0.772	−3.47 (3.50)	0.336
HIV symptom distress
Cannabis	0.88 (0.60)	0.149	0.52 (0.60)	0.387
Cocaine	−1.09 (1.37)	0.433	1.24 (1.51)	0.421
Stimulant	0.46 (1.75)	0.794	1.26 (1.82)	0.491
Methamphetamine	2.37 (1.85)	0.220	1.72 (1.53)	0.268
Inhalants	8.33* (3.08)	0.015	1.35 (2.70)	0.621
Sedative	−0.20 (0.95)	0.829	−1.12 (1.20)	0.366
Hallucinogens	−11.21 (7.27)	0.162	−3.03 (5.69)	0.602
Heroin	7.90 (7.01)	0.280	−2.10 (6.49)	0.750
Prescription opioids	−0.65 (1.35)	0.633	−1.22 (1.24)	0.325
Other drug	1.57 (1.86)	0.409	−0.12 (1.52)	0.936

All regression models included the 10 substance variables simultaneously as predictors to adjust for potential polysubstance use and isolate the unique association of each substance with HIV outcomes. All models were adjusted for intervention assignment, site, age, gender identity, sex assigned at birth, race, sexual orientation, education level, annual income, and baseline outcome values.

*p < .050.

SE, standard error.

Discussion

This study contributes to the growing body of literature on the intersection of substance use and HIV health by examining substance-specific associations with ART adherence self-efficacy, self-reported ART adherence, self-management, and symptom distress among PWH with suboptimal ART adherence over 12 months. While prior studies have demonstrated the negative impact of substance use on HIV outcomes, few have used longitudinal data to systematically compare the effects of multiple substances across different domains of HIV care. Our findings provide empirical support for the need to move beyond treating substance use as a monolithic risk factor and highlight the clinical utility of substance-specific risk assessments.

Consistent with prior research, cocaine use was significantly associated with lower self-reported ART medication adherence at 6 months. This finding aligns with previous studies documenting the association between cocaine use and nonadherence to ART.^14–16 Cocaine’s physiological effects on executive functioning and impulse control, coupled with a chaotic lifestyle often associated with its use, may reduce individuals’ capacity to engage consistently in daily treatment routines.^46,47 These findings underscore the importance of incorporating substance-specific assessments into HIV care to flag patients at heightened risk for adherence challenges.

While cocaine use was significantly associated with lower self-reported ART adherence, other substances did not show significant associations in the same model. This may reflect differences in the pharmacological effects, use patterns, and social contexts associated with each substance. For instance, prescription opioids, even when misused, may have less immediately disruptive effects on daily functioning because their sedative properties can dull physical discomfort and emotional distress without acutely impairing cognition or behavior in the same way as cocaine.^48,49 Interestingly, methamphetamine use, heroin use, and other opioid use were not significantly associated with ART adherence in our study, despite prior evidence documenting such associations, particularly among sexual minority men and people who inject drugs.^20,21,24,25 This discrepancy may stem from our sample’s composition, as all participants already exhibited suboptimal ART adherence at baseline, potentially restricting variability in adherence outcomes and limiting the detection of substance-specific effects. Moreover, the prevalence of methamphetamine, heroin, and opioid use was relatively low in our cohort, and differences in demographics and behavioral patterns compared with prior research may also contribute. Additionally, we found no significant association between cannabis use and ART adherence, which aligns with the mixed findings in prior research, where some studies report associations with suboptimal adherence and others find no significant relationship.^21–23 These results highlight the complexity of substance use in the context of HIV care and underscore the need for future research to disentangle not only the type of substance used but also patterns of polysubstance use, intensity, and context of use, particularly in diverse populations.

Managing HIV extends far beyond adherence to ART—it encompasses a broad array of behaviors including physical activity, stress management, dietary changes, spiritual coping, medical engagement, and social support utilization. In this study, methamphetamine use was uniquely associated with poorer HIV self-management at 12 months, even after adjusting for baseline levels and covariates. This finding builds on limited prior research linking methamphetamine use to suboptimal ART adherence, particularly among sexual minority men,^13,16,50 and expands this evidence to demonstrate its predictive impact on a wider range of self-care behaviors in a more demographically diverse sample. Methamphetamine’s known effects on neurocognitive functioning, emotional regulation, and behavioral inhibition may disrupt individuals’ ability to initiate and sustain the varied and goal-directed behaviors required for long-term HIV management.^13,50,51 These findings suggest that HIV self-management interventions may be strengthened by incorporating cognitive-behavioral components that directly address the specific challenges posed by methamphetamine use, particularly for populations at elevated risk.

Hallucinogen use was significantly associated with lower ART adherence self-efficacy at 12 months. The HIV Treatment ASES captures confidence in one’s ability to manage treatment in the face of challenges such as side effects, disrupted routines, stigma, discouragement, and logistical barriers. Hallucinogens, such as lysergic acid diethylamide (LSD) and psilocybin, induce lasting changes in sensory processing and self-perception through serotonin 2A receptor agonism,⁵² which may interfere with the ability to sustain treatment adherence confidence over time. That no other substances showed a similar association may reflect the unique impact of hallucinogens on introspective and perceptual domains that are especially relevant for self-efficacy beliefs.⁵³ Our findings contribute to the limited literature linking hallucinogen use to HIV treatment adherence self-efficacy, a key cognitive mechanism underlying long-term engagement with care.³⁴ These findings point to the potential value of tailoring adherence interventions for PWH who use hallucinogens by integrating strategies to strengthen self-efficacy, such as cognitive restructuring, motivational interviewing, and behavioral rehearsal techniques.

We also found that inhalant use predicted increased HIV symptom distress at 6 months. While inhalant use is less commonly studied in the context of HIV care, this finding is consistent with prior literature linking inhalants to adverse immunological and psychological effects. Laboratory studies have demonstrated that even brief exposure to amyl nitrite vapors can induce functional immune suppression, which may exacerbate symptom expression in immunocompromised individuals.⁵⁴ Furthermore, some studies suggest that inhalants may not only impair immune defense but also directly promote HIV viral replication, leading to more pronounced symptom burden.⁵⁵ Psychologically, inhalant use has been associated with anxiety sensitivity and panic symptomatology, which may heighten distress through misinterpretation or amplification of common HIV-related sensations.^56–58 The overlap between inhalant side effects (e.g., dizziness, nausea, shortness of breath) and HIV symptoms further complicates symptom attribution, potentially reinforcing a cycle of symptom-focused anxiety and distress.^57,59,60 This underscores the need for symptom monitoring and harm reduction interventions that account for inhalant use, particularly in settings serving marginalized populations.

Several limitations warrant consideration. First, we used baseline substance use to predict HIV outcomes at 6 and 12 months. Future longitudinal studies incorporating repeated assessments of substance use are needed to capture dynamic risk trajectories and to better understand the reciprocal relationship between substance use and HIV-related outcomes. Second, while the sample was sociodemographically diverse, it was drawn from participants enrolled in a behavioral intervention study, which may limit generalizability to PWH not engaged in care or research. Relatedly, findings may not extend beyond the urban Northeast and Southern US contexts represented in this study, where structural and cultural factors may differ from other regions. Third, although we controlled for a wide range of baseline covariates—including intervention arm, study site, baseline HIV outcome scores, and key sociodemographic characteristics—residual confounding may still be present. Fourth, alcohol was not included in the ASSIST assessment. Given the known impact of alcohol on ART adherence and HIV outcomes, future research should incorporate alcohol to provide a more comprehensive understanding of substance-specific risks. Fifth, although the majority of participants completed study procedures in English, four participants completed measures in Spanish. While we used instruments that have been previously validated in Spanish-speaking populations, administering measures in multiple languages could introduce potential measurement variability. This represents a minor limitation, particularly given the small number of Spanish-only participants. Finally, although we intentionally sought to enroll a nearly equal number of men and women to address the underrepresentation of women in HIV research, this sampling strategy resulted in a higher proportion of women than is typical in the US HIV epidemic, potentially limiting direct comparisons with national epidemiological data and requiring caution when generalizing findings to male-dominated populations. These limitations underscore the need for more robust, multisite longitudinal research to further elucidate substance-specific pathways influencing HIV outcomes.

In conclusion, the findings demonstrate that specific substances are differentially associated with HIV-related outcomes. By identifying distinct substance-specific risks, this study provides a foundation for more personalized approaches to HIV care. Integrated models that combine HIV treatment with behavioral health services tailored to substance use profiles may improve health outcomes for PWH and reduce disparities in care engagement and treatment success.

Disclaimer

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

Research reported in this publication was supported by the National Institute of Nursing Research of the National Institutes of Health under award number R01NR019758.