Methamphetamine use and HIV risk behaviors in a cohort of homeless-experienced adults who use drugs

Abstract

Background:

Methamphetamine use has increased substantially in the United States, particularly among people experiencing homelessness. While the association between methamphetamine use and HIV risk has been studied in other populations, its impact on HIV risk behaviors among people experiencing homelessness remains less well-studied. The objective of this study was to evaluate the relationship between methamphetamine use and HIV risk behaviors in a cohort of people experiencing homelessness who use drugs.

Methods:

We conducted a cross-sectional survey of adults with recent drug use recruited from three homeless-serving clinical sites in Boston, Massachusetts, between February 2024 and January 2025. The primary exposure was past 3-month methamphetamine use (any vs. none). The primary outcome was a composite measure of HIV risk behavior in the past 3 months, categorized as low (0 events), moderate (1–11 events), or high (≥12 events), based on self-reported condomless sex and shared needle events. Secondary outcomes included sex-related and injection-related HIV risk behaviors, examined separately. We used multinomial logistic regression to assess associations with overall and sex-related HIV risk behaviors, and bivariate analyses for injection-related HIV risk behavior due to sparse outcome data.

Results:

Of 196 participants, the mean age was 48 years. Sixty percent identified as cis-gender male, 56% as White, 31% as Black, and 20% as Hispanic. Methamphetamine use in the past 3 months was reported by 42% of participants. Its use was not independently associated with overall HIV risk behavior (aRR for high vs. low risk: 2.20; 95% CI: 0.84–5.77) or sex-related HIV risk behavior (aRR for high vs. low risk: 1.45; 95% CI: 0.53–4.00). However, in bivariate analyses, methamphetamine use was significantly associated with injection-related HIV risk behaviors: individuals who used methamphetamine were less likely to have low-risk behavior (27% vs. 79%) and more likely to have moderate-risk (43% vs. 8%) and high-risk (31% vs. 4%) behaviors (P < 0.001).

Conclusions:

Among homeless-experienced adults who use drugs, methamphetamine use was associated with higher-risk injection behaviors but not with sex-related HIV risk behaviors. HIV prevention strategies in this population should prioritize safer injection practices, evidence-based methamphetamine treatment, and equitable access to pre-exposure prophylaxis.

1. Introduction

Methamphetamine use has risen sharply in the United States (U.S.) over the past decade, with overdose deaths involving psychostimulants—primarily methamphetamine—increasing 24-fold.¹ Once concentrated in the Western U.S. and among white men, methamphetamine use has expanded both geographically and demographically.^2–4 From 2018–2019, the Northeast U.S. experienced the largest relative surge in psychostimulant-involved mortality.⁴ Between 2015 and 2019, methamphetamine use increased significantly across multiple groups, including heterosexual and sexual minority men and women, and individuals identifying as Black, Hispanic, and White.³

People experiencing homelessness are disproportionately affected by substance use, including methamphetamine use. Recent studies indicate that between one quarter and one third of people experiencing homelessness report regular methamphetamine use.^5,6 This high prevalence has serious public health implications. Methamphetamine-related emergency department visits are on the rise and associated with increased use of chemical restraint (e.g., involuntary administration of a sedating medication for behavioral control) and cardiovascular complications.^7–9 Investigations of recent HIV outbreaks among people experiencing homelessness have identified methamphetamine and opioid co-use as key drivers of HIV transmission.^10–13

This elevated HIV transmission risk is believed to stem from methamphetamine’s effects on the central nervous system, including heightened sexual arousal, increased impulsivity, and reduced inhibition.¹⁴ These mechanisms have been well-documented among men who have sex with men, where methamphetamine use is associated with condomless sex, multiple sexual partners, and HIV seroconversion.^15–21 Associations between methamphetamine use and HIV risk behaviors, including both sex- and injection-related behaviors, have also been demonstrated in people who inject drugs.^22,23 However, research to date has rarely focused on people experiencing homelessness—a distinct high-risk group whose structural vulnerability, unstable housing, limited access to health care, and complex drug use patterns may shape HIV risk differently. Despite overlapping behaviors with other high-risk populations, this group faces unique barriers to HIV prevention and treatment and remains underrepresented in behavioral research.

Unlike other substance use disorders, such as opioid use disorder, there are currently no FDA-approved medications for methamphetamine use disorder. Behavioral interventions, such as contingency management, remain the primary evidence-based treatment.²⁴ This treatment gap underscores the urgency of addressing methamphetamine-related harms through tailored prevention strategies. At the same time, the HIV prevention landscape has evolved: pre-exposure prophylaxis (PrEP)—including both daily oral and long-acting injectable formulations—offers powerful protection, but remains underutilized among people who use drugs.^25,26

A better understanding of how methamphetamine use relates to HIV risk behaviors among homeless-experienced individuals is essential to inform tailored prevention strategies.²⁷ The objective of this study was to examine the association between methamphetamine use and HIV risk behaviors in a cohort of homeless-experienced adults who use drugs. We hypothesized that methamphetamine use would be associated with increased HIV risk behaviors, including both sex- and injection-related behaviors.

2. Methods

2.1. Setting and Participants

We recruited participants from three clinical sites affiliated with Boston Health Care for the Homeless Program (BHCHP): BHCHP’s headquarters facility adjacent to an academic medical center (“headquarters site”), a clinic embedded within a local day shelter (“day shelter site”), and an addiction-focused clinic deployed through a mobile outreach unit (“mobile site”). The BHCHP headquarters facility provides comprehensive health care for individuals experiencing homelessness. We recruited participants from its lobby as well as from the 24-hour medical respite program housed within the same building. The day shelter serves over 500 guests daily and offers medical care, behavioral health support, and housing assistance. The mobile unit delivers addiction and primary care to homeless-experienced individuals at overdose hotspots throughout Boston.

We recruited English and Spanish-speaking adults (≥18 years) with a lifetime history of homelessness who had used drugs in the past 3 months and were not known to be living with HIV. We included both currently and formerly homeless individuals because individuals accessing homeless health care services at recruitment sites often transition between periods of homelessness and housing, and those recently housed often remain at elevated risk for adverse health outcomes. HIV negativity was required since our primary outcome, HIV risk behaviors, may be influenced by an individual’s HIV status, and the goal of this study was to provide preliminary data for an intervention aimed at preventing HIV acquisition. Individuals were excluded if they were unable to correctly answer 3 basic knowledge questions about the study during the informed consent process, following methods used in prior research conducted with this population.^28–30

2.2. Study design and procedures

We prospectively recruited participants from February 26, 2024, to January 29, 2025. After providing verbal informed consent, participants completed a structured quantitative survey and received a $10 gift card for their time. The study was approved by the Mass General Brigham Institutional Review Board.

2.3. Measures

Study measures were primarily drawn from the Centers for Disease Control and Prevention’s (CDC) Medical Monitoring Project (MMP), a national surveillance project that collects data on clinical care, health behaviors, and health outcomes of people living with HIV in the U.S.^31–34 Most items were adapted to shorten the recall period from the original 12 months to 3 months. This timeframe was selected to balance the need for capturing meaningful behavior patterns with the goal of minimizing recall bias, particularly in a population with lower health literacy and higher behavioral complexity.^35,36

2.3.1. Primary exposure

The primary exposure of interest was past 3-month methamphetamine use, dichotomized as “any use” vs. “no use”, regardless of the route of administration. One participant reported “Don’t know” in response to past 3-month methamphetamine use, which was categorized as no use in the analyses. In a sensitivity analysis, we dichotomized past 3-month methamphetamine use as “daily” vs. “less than daily” to assess whether more frequent methamphetamine use impacted HIV risk behaviors differently.

2.3.2. Outcomes

The primary outcome was overall HIV risk behavior in the past 3 months, defined as the total number of condomless sex events (including receptive and insertive anal and vaginal sex) and shared needle events during that period. We categorized this outcome into three levels: low (0 events), moderate (1–11 events, corresponding to less than one event per week on average), and high (12 or more events, corresponding to at least one event per week on average). We categorized the outcome into these clinically meaningful thresholds to mitigate the influence of extreme outliers and to account for potential recall bias in self-reported event counts.

Our secondary outcomes included disaggregated sex-related HIV risk behaviors and injection drug use-related HIV risk behaviors, each assessed over the past 3 months. Sex-related HIV risk behavior was defined as the total number of condomless sex events (including receptive and insertive anal and vaginal sex) and categorized into three levels (low, moderate, and high) using the same thresholds as the primary outcome described above. Given the relatively low frequency of reported shared needle events, injection drug use-related HIV risk behavior categories were defined differently: low risk was defined as no injection drug use in past 3 months, moderate risk as injection drug use in past 3 months without shared needle use, and high risk as injection drug use in past 3 months with shared needle use at least once.

2.3.3. Other measures

Demographic variables included age, race, ethnicity, sex, gender, and sexual orientation. Social determinants included incarceration in the past 3 months and unsheltered homelessness in the past 3 months. In addition to methamphetamine use, substance use measures also captured frequency of past 3-month opioid and cocaine use.

2.4. Statistical analysis

We used descriptive statistics to summarize cohort characteristics, stratified by past 3-month methamphetamine use. To assess the association between methamphetamine use and overall HIV risk behavior (primary outcome), we conducted multinomial regression, modeling HIV risk behavior as a 3-level outcome (low, moderate, and high), with “low” as the reference category. An ordinal logistic model was initially considered; however, the proportional odds assumption was not met, prompting the use of a multinomial approach. Covariates were selected a priori based on prior literature and clinical judgment, with careful consideration of model parsimony given the limited sample size.

We conducted a separate multinomial regression model to examine the association between methamphetamine use and sex-related HIV risk behavior (secondary outcome), using the same covariates as in the primary model. For the multinomial models, we report adjusted risk ratios (aRRs) and 95% confidence intervals (CIs), comparing moderate- vs. low-risk and high- vs. low-risk groups.

For the secondary outcome of injection drug use-related HIV risk behavior, we chose to present bivariate analyses because the number of events was small in many covariate subgroups, leading to unstable estimates in adjusted models. We used analysis of variance for continuous variables and Fisher’s exact tests for binary and categorical variables.

Finally, we conducted sensitivity analyses with the drug use variables recategorized as “daily” vs. “less than daily use”.

3. Results

3.1. Cohort characteristics

We enrolled a total of 206 participants. Ten were excluded from the analytic sample: five due to post-enrollment ineligibility (four initially screened eligible but were later determined ineligible based on survey responses, and one was identified as a duplicate enrollee) and five due to incomplete survey completion (<50% completed). Thus, 196 participants were included in the final analytic sample.

The average age was 48 years (SD 10.6), and most were male (62.8%). Just over half (55.6%) identified as White, nearly one third (30.6%) as Black, and 20.4% as Hispanic (Table 1).

Table 1.

Characteristics of a cohort of homeless-experienced adults who use drugs by past 3-month methamphetamine use status

Characteristic	Entire cohort (n=196) N (%)	Methamphetamine use (n=82) N (%)	No methamphetamine use (n=114a) N (%)
Age, mean (SD)	48.0 (10.6)	44.2 (9.4)	50.7 (10.5)
Sex
Male	123 (62.8)	57 (69.5)	66 (57.9)
Female	73 (37.2)	25 (30.5)	48 (42.1)
Gender
Cisgender male	118 (60.2)	53 (64.6)	65 (57.0)
Cisgender female	70 (35.7)	24 (29.3)	46 (40.4)
Transgenderb	8 (4.1)	5 (6.1)	3 (2.6)
Race
Black	60 (30.6)	15 (18.3)	45 (39.5)
White	109 (55.6)	59 (72.0)	50 (43.9)
Otherc	7 (3.6)	2 (2.4)	5 (4.4)
Unknown	20 (10.2)	6 (7.3)	14 (12.3)
Ethnicity
Hispanic	40 (20.4)	18 (22.0)	22 (19.3)
Non-Hispanic	154 (78.6)	63 (76.8)	91 (79.8)
Unknown	2 (1.0)	1 (1.2)	1 (0.9)
Sexual orientation
Heterosexual	160 (81.6)	65 (79.3)	95 (83.3)
Non-heterosexuald	36 (18.4)	17 (20.7)	19 (16.7)
Unsheltered homelessness (past 3 months)e	138 (70.4)	70 (85.4)	68 (59.6)
Incarcerated (past 3 months)	43 (21.9)	23 (28.0)	20 (17.5)
Opioid use (past 3 months)f,g	144 (73.5)	71 (86.6)	73 (64.0)
Cocaine use (past 3 months)g	164 (83.7)	75 (91.5)	89 (78.1)

^aIncludes 1 participant who responded “Don’t know” to the methamphetamine use item

^bIncludes 1 indeterminate

^cIncludes American Indian/Alaska Natives (n=3), Asian (n=1), Native Hawaiian/Pacific Islanders (n=4)

^dIncludes lesbian or gay (n=6), bisexual or pansexual (n=26), and ‘Don’t know’ (n=4)

^eIncludes sleeping outside or in a car

^fIncludes heroin/fentanyl and prescription pain medications not used as prescribed; 1 participant who responded “Don’t know” to pain medication use item is counted as not using opioids

^gIncludes any use, regardless of route

Eighty-two participants (41.8%) reported methamphetamine use in the past 3 months. Among those who reported methamphetamine use, 19 (23.2%) reported daily use, 51 (62.2%) reported injecting methamphetamine, and 74 (90.2%) reported smoking or snorting it.

Compared to participants who did not use methamphetamine in the past 3 months, those who did were younger (mean age 44.2 vs. 50.7 years; P < 0.001) and more likely to report unsheltered homelessness in the past 3 months (85.4% vs. 59.6%; P < 0.001). They also had higher rates of recent opioid use (86.6% vs. 64.0%; P = 0.001) and cocaine use (91.5% vs. 78.1%; P = 0.018) (Table 1).

3.2. Overall HIV risk behavior

Among those who used methamphetamine, 35.4% engaged in moderate overall HIV risk behavior and 23.2% engaged in high overall HIV risk behavior, compared to 29.0% and 11.4% among those who did not use methamphetamine. In multivariable models, methamphetamine use was not independently associated with overall HIV risk behavior (aRR for moderate-risk behavior = 1.39, 95% CI: 0.65–2.98; aRR for high-risk behavior = 2.20, 95% CI: 0.84–5.77). Older age was independently associated with a reduced likelihood of overall HIV risk behavior (aRR per year for moderate-risk behavior = 0.96, 95% CI: 0.93–1.00; aRR per year for high-risk behavior = 0.96, 95% CI: 0.91–1.00) (Table 2). Results remained robust in sensitivity analyses recategorizing methamphetamine use as “daily” vs. “less than daily,” with no substantive changes in the direction or magnitude of associations across outcomes (Online Resources 1–3).

Table 2.

Adjusted associations with overall HIV risk behavior in a cohort of homeless-experienced adults who use drugs

	Moderate-risk behaviora,b		High-risk behaviorb,c
Characteristic	aRR (95% CI)	P value	aRR (95% CI)	P value
Age (per year)	0.96 (0.93–1.00)	0.036	0.96 (0.91–01.00)	0.041
Sex
Female	ref	ref	ref	ref
Male	0.85 (0.42–1.72)	0.65	0.93 (0.37–2.32)	0.87
Race
White	ref	ref	ref	ref
Black	1.70 (0.78–3.75)	0.18	1.54 (0.57–4.16)	0.40
Other/unknown	0.48 (0.14–1.64)	0.24	0.20 (0.02–1.92)	0.16
Ethnicity
Non-Hispanic	ref	ref	ref	ref
Hispanic	1.10 (0.42–2.91)	0.84	0.70 (0.18–2.76)	0.62
Sexual Orientation
Heterosexual	ref	ref	ref	ref
Non-heterosexual	1.05 (0.44–2.52)	0.90	0.63 (0.18–2.21)	0.47
Methamphetamine use
Did not use	ref	ref	ref	ref
Used	1.39 (0.65–2.98)	0.39	2.20 (0.84–5.77)	0.11
Opioid use
Did not use	ref	ref	ref	ref
Used	1.38 (0.62–3.08)	0.43	1.53 (0.52–4.48)	0.44
Cocaine use
Did not use	ref	ref	ref	ref
Used	1.50 (0.58–3.88)	0.40	1.34 (0.39–4.63)	0.65

^aModerate-risk behavior: 1–11 total HIV risk behavior events (condomless sex and shared needle events) in the past 3 months (averaging less than once per week), n=62

^bReference group for multinomial regression models: low-risk behavior (0 HIV risk behavior events [condomless sex and shared needle events] in the past 3 months), n=102

^cHigh-risk behavior: 12 or more total HIV risk behavior events (condomless sex and shared needle events) in the past 3 months (averaging at least once per week), n=32

3.3. Sex-related HIV risk behavior

Among those who used methamphetamine, 35.4% engaged in moderate sex-related HIV risk behavior and 15.8% engaged in high sex-related HIV risk behavior, compared to 24.6% and 11.4% among those who did not use methamphetamine. In multivariable models, methamphetamine use was not independently associated with sex-related HIV risk behavior (aRR for moderate-risk behavior = 1.78, 95% CI: 0.81–3.89; aRR for high-risk behavior = 1.45, 95% CI: 0.53–4.00). Older age was independently associated with a reduced likelihood of sex-related HIV risk behavior (aRR per year for moderate-risk behavior = 0.96, 95% CI: 0.93–1.00; aRR per year for high-risk behavior = 0.96, 95% CI: 0.91–1.00) (Table 3).

Table 3.

Adjusted associations with sex-related HIV risk behavior in a cohort of homeless-experienced adults who use drugs

	Moderate-risk behaviora,b		High-risk behaviorb,c
Characteristic	aRR (95% CI)	P value	aRR (95% CI)	P value
Age (per year)	0.96 (0.93–1.00)	0.039	0.96 (0.91–1.00)	0.06
Sex
Female	ref	ref	ref	ref
Male	0.64 (0.31–1.31)	0.22	0.86 (0.33–2.26)	0.76
Race
White	ref	ref	ref	ref
Black	2.50 (1.11–5.66)	0.027	1.66 (0.59–4.69)	0.34
Other/unknown	0.75 (0.22–2.57)	0.65	0.23 (0.02–2.38)	0.22
Ethnicity
Non-Hispanic	ref	ref	ref	ref
Hispanic	0.97 (0.36–2.61)	0.96	1.04 (0.26–4.16)	0.96
Sexual Orientation
Heterosexual	ref	ref	ref	ref
Non-heterosexual	1.17 (0.49–2.80)	0.72	0.39 (0.08–1.89)	0.24
Methamphetamine use
Did not use	ref	ref	ref	ref
Used	1.78 (0.81–3.89)	0.15	1.45 (0.53–4.00)	0.47
Opioid use
Did not use	ref	ref	ref	ref
Used	1.00 (0.44–2.29)	0.99	1.10 (0.37–3.29)	0.86
Cocaine use
Did not use	ref	ref	ref	ref
Used	2.57 (0.86–7.67)	0.09	1.73 (0.43–6.69)	0.42

^aModerate-risk behavior: 1–11 condomless sex events in the past 3 months (averaging less than once per week), n=57

^bReference group for multinomial regression models: low-risk behavior (0 condomless sex events in the past 3 months), n=113

^cHigh-risk behavior: 12 or more condomless sex events in the past 3 months (averaging at least once per week), n=26

3.4. Injection drug-related HIV risk behavior

In bivariate analyses, methamphetamine use was significantly associated with injection drug use-related HIV risk behavior (P <0.001; Table 4). Compared to participants who did not use methamphetamine, those who did were more likely to have injected drugs in the past 3 months without sharing needles (moderate-risk behavior: 42.7% vs. 16.7%) and to have injected drugs and shared needles (high-risk behavior: 30.5% vs. 4.4%). Older age was significantly associated with lower injection-related risk (P <0.001): the mean age was 50.9 years for low-risk, 45.2 years for moderate-risk, and 42.1 years for high-risk behavior. Race was also associated with injection-related risk (P <0.001): White participants were more likely than Black participants to report moderate-risk behavior (32.1% vs. 16.7%) and high-risk behavior (25.7% vs. 3.3%). Opioid use showed a similar pattern (P <0.001). Participants who used opioids were more likely report moderate-risk behavior (34.7% vs. 7.7%) and high-risk behavior (20.8% vs. 0.0%) compared to those who did not use opioids.

Table 4.

Bivariate associations with injection drug use-related HIV risk behavior in a cohort of homeless-experienced adults who use drugs

	Low-risk behaviora	Moderate-risk behaviorb	High-risk behaviorc	P value
Characteristic	n (%)	n (%)	n (%)
Age, mean (SD)	50.9 (11.0)	45.2 (8.7)	42.1 (7.9)	<0.001
Sex				0.67
Female	42 (57.5)	18 (24.7)	13 (17.8)
Male	70 (56.9)	36 (29.3)	17 (13.8)
Race				<0.001
White	46 (42.2)	35 (32.1)	28 (25.7)
Black	48 (80.0)	10 (16.7)	2 (3.3)
Other/unknown	18 (66.7)	9 (33.3)	0 (0.0)
Ethnicity				0.15
Non-Hispanic	90 (57.7)	39 (25.0)	27 (17.3)
Hispanic	22 (55.0)	15 (37.5)	3 (7.5)
Sexual Orientation				0.18
Heterosexual	95 (59.4)	44 (27.5)	21 (13.1)
Non-heterosexual	17 (47.2)	10 (27.8)	9 (25.0)
Methamphetamine use				<0.001
Did not use	90 (78.9)	19 (16.7)	5 (4.4)
Used	22 (26.8)	35 (42.7)	25 (30.5)
Opioid use				<0.001
Did not use	48 (92.3)	4 (7.7)	0 (0.0)
Used	64 (44.4)	50 (34.7)	30 (20.8)
Cocaine use				0.64
Did not use	21 (65.6)	7 (21.9)	7 (12.5)
Used	91 (55.5)	47 (28.7)	26 (15.9)

^aLow-risk behavior: no injection drug use in the past 3 months (n=112)

^bModerate-risk behavior: injection drug use but no shared needle use in the past 3 months (n=54)

^cHigh-risk behavior: shared needle use in the past 3 months; n=30

4. Discussion

In this prospective study of 196 homeless-experienced adults with recent drug use, we successfully recruited a marginalized population often excluded from HIV behavioral research. To our knowledge, this is one of the first studies to directly examine methamphetamine use and HIV risk behaviors within a Health Care of the Homeless (HCH) setting such as BHCHP. We found that methamphetamine use was significantly associated with increased injection-related HIV risk behaviors but not with overall or sex-related HIV risk behaviors. Older age was consistently associated with lower HIV risk behaviors, while racial differences emerged specifically in relation to injection-related HIV risks. These findings contribute to the growing body of literature on how methamphetamine use intersects with HIV vulnerability in marginalized populations and highlight opportunities to tailor harm reduction and prevention strategies in HCH settings.^{10,14,15,21,23}

Contrary to our initial hypothesis, methamphetamine use was not significantly associated with overall or sex-related HIV risk behaviors. Instead, it was strongly associated with injection drug use-related risk. Participants who used methamphetamine were nearly three times more likely to report moderate-risk injection behavior (injecting without shared needle use) and nearly seven times more likely to report high-risk injection behavior (injecting with shared needle use) as those who did not use methamphetamine. This underscores the central role of injection-related pathways in HIV transmission risk among homeless-experienced individuals who use methamphetamine. In contrast to findings from other populations—particularly men who have sex with men—where methamphetamine use is more often associated with sex-related HIV risk,^15–20 our findings suggest that injection behavior may be a more prominent mechanism of HIV exposure in this population of homeless-experienced individuals who use methamphetamine.

Accordingly, prevention efforts should take a three-pronged approach: (1) expanding harm reduction services that support safer injection practices (e.g., syringe service programs [SSPs] and supervised consumption sites), (2) increasing access to evidence-based substance use disorder treatment, and (3) ensuring availability of biomedical HIV prevention, including PrEP. SSPs have long been shown to reduce HIV transmission among people who use drugs, yet increased accessibility to these programs remains critical.³⁷ Though supervised consumption sites are less common in the U.S., studies suggest they could significantly reduce incident HIV infections in this population.^38–40 Beyond their direct health benefits, these harm reduction services also provide an important point of engagement for linkage to substance use disorder treatment and PrEP initiation.^38,41 While no medications are currently Food and Drug Administration-approved for methamphetamine use disorder, behavioral interventions such as contingency management and the Matrix Model have demonstrated reductions in both injection- and sex-related HIV risk behaviors.^42–44 Additionally, given the substantial overlap between methamphetamine and opioid use, expanding access to medications for opioid use disorder (MOUD), such as buprenorphine and methadone, is also essential to reduce injection-associated HIV risk.⁴⁵ PrEP significantly reduces HIV transmission, but adherence remains challenging in this population.^26,46 Low-threshold oral PrEP programs tailored to people experiencing homelessness have demonstrated feasibility and modest retention, such as BHCHP’s initiative in which 152 individuals were newly prescribed PrEP over 17 months and 44% remained on PrEP at 6 months.⁴⁷ Long-acting injectable PrEP may help overcome adherence barriers and has shown superior efficacy in other high-risk populations.^48,49 Collectively, these strategies underscore the need for integrated approaches that combine harm reduction, substance use disorder treatment, and PrEP access with adherence support to reduce HIV risk among homeless-experienced individuals who use methamphetamine. Additionally, although our analyses focus on behavioral exposure pathways associated with HIV acquisition, incorporating person-centered communication strategies that shift from “risk-based” to “reasons-based” framing of HIV prevention may help reduce stigma and support engagement in structurally marginalized populations.^50,51

Our findings align with prior research showing that younger individuals who use drugs tend to engage in higher levels of HIV-related risk behaviors compared to older adults.⁵² Although our ability to directly assess risk among adolescents and young adults is limited because our cohort skewed older on average, neurodevelopmental factors, particularly ongoing maturation of regions governing impulse control and risk assessment, may contribute to higher risk behavior in younger populations.⁵³ Early drug use in younger individuals is also more likely to be rooted in trauma, poverty, and unstable environments, potentially compounding risk.⁵⁴ Effective prevention strategies for this subgroup should therefore address structural determinants, incorporate trauma-informed care, and leverage youth-specific engagement strategies such as peer outreach, digital tools, or incentives that resonate with their stage of development.^55–57

We also observed notable racial differences in injection-related HIV risk. Compared to White participants, those identifying as Black were less likely to engage in high-risk injection behaviors. These findings align with prior studies and may reflect cultural or structural differences in drug use patterns, with Black individuals injecting drugs and sharing needles less frequently than their White counterparts.^58,59 Importantly, these results underscore the need for culturally responsive HIV prevention strategies that reflect the experiences of diverse populations. Further research is needed to better understand the underlying drivers of these behavioral differences and to ensure that HIV prevention interventions are equitably effective across racial groups.

This study has several limitations. First, the cross-sectional design limits our ability to draw causal inferences about the relationship between methamphetamine use and HIV risk behaviors. Second, all data were self-reported and may be subject to recall or social desirability bias, particularly regarding stigmatized behaviors such as drug use and sex. While we used standardized measures and limited the recall window to 3 months to reduce these effects, some misreporting is possible. Third, our HIV risk measures quantified the number of exposure events but did not capture the number of sex or injection partners, which could lead to misclassification of acquisition risk for individuals reporting repeated events within a single partnership. Fourth, the sample size limited the number of covariates included in regression models, potentially leading to residual confounding. It also prohibited modeling injection-related risk specifically. Fifth, although our study population reflects the demographics of individuals accessing homeless services in Boston, findings may not generalize to people experiencing homelessness that do not access such services or to other geographic regions. Additionally, Boston’s relatively strong harm reduction infrastructure, including SSPs, mobile outreach programs, and overdose prevention centers, may limit generalizability to areas with fewer such services.^60–62 Finally, because we restricted enrollment to individuals without HIV, our findings focus on HIV acquisition risk and do not address HIV transmission behaviors among people living with HIV, which represent an important complementary area for future research.

5. Conclusions

In this study of homeless-experienced adults who use drugs, methamphetamine use was common and strongly associated with injection-related HIV risk behaviors, but not with sex-related HIV risk behaviors. These findings suggest that in this population, injection practices may be the primary pathway for methamphetamine-associated risk for HIV transmission. Tailored HIV prevention strategies that prioritize safer injection practices, effective substance use disorder treatment, and expanded access to PrEP are urgently needed to address the evolving intersection of methamphetamine use and HIV risk among people experiencing homelessness.