Introduction

On any given night in the United States (US), over 770,000 people experience homelessness, with approximately 36% more people experiencing homelessness now compared to early 2020 [1]. Encampments provide some shelter, protection, and community for people living on the street [2, 3]. Due to the criminalization of homelessness and the desire to “erase” homelessness from public view, encampment sweeps are a strategy used by local authorities to remove these areas, oftentimes discarding personal belongings forcibly and without consent [2, 3]. People who use drugs (PWUD) are more likely to be unhoused; homelessness may increase drug use as a coping mechanism [4].

Massachusetts (MA) has a homeless population of approximately 56,000 people and is one of two states with “Right to Shelter” laws [5]. Intended to support individuals experiencing homelessness in accessing emergency shelters and basic necessities, these laws are not fully implemented. Police-led sweeps persist across MA, with numerous incidents documented in 2025 alone and many more likely occurring without documentation [6,7,8]. Even with robust harm reduction and public health programs [9, 10], MA faces a high rate of overdose, with 2104 deaths in 2023 alone [11].

The recent proliferation of policy directives calls for the enforcement of prohibitions on open drug use and “urban camping” [12]. Over half of major US cities have banned camping, sitting, and lying in public [13]; these laws have health consequences (i.e., overdose) [14,15,16,17,18]. Encampment sweeps are also associated with increased violence, syringe sharing, and reduced access to healthcare [19]. Partially due to the loss of evidence-based overdose prevention supplies (e.g., naloxone) [17], sweeps increase deaths up to an estimated 16–24% among unsheltered PWUD [15]. Although some sweeps have been coupled with offers of transitional housing, it is not well understood how consistently these services are offered [20].

The aim of our study was to characterize the experiences of past-year encampment sweep and identify associations with suboptimal drug use-related health behaviors (e.g., overdose, injection frequency), broader health outcomes (e.g., anxiety and depressive symptoms), and healthcare access (e.g., difficulty accessing medical clinics).

Methods

Study Design and Data Collection

PrEP for Health is a randomized controlled trial (NCT04430257) focused on improving uptake, adherence, and persistence of HIV pre-exposure prophylaxis among PWUD [21]. Recruitment occurred from 2020 to 2025 in partnership with syringe service providers in Lawrence and Cambridge, MA. Eligible individuals were aged ≥ 18, self-reported past-month injection drug use (IDU), HIV-uninfected, and provided informed consent. Assessments were conducted in person or remotely by trained research staff. Participants were followed for up to 12 months. Participants received escalating compensation of $20 to $50 across visits. An additional survey module, completed by 155 participants (116 at baseline, 39 post-baseline), was conducted to explore experiences of encampment sweeps and loss of belongings. There was no compensation corresponding to the additional module. The Fenway Health Institutional Review Board approved all study protocols.

Measures

We asked participants, “In the past year, have you been voluntarily or involuntarily (forcibly) relocated (meaning you had to change your living location) due to an ‘encampment sweep’ or ‘clean up’ of public areas where you were living?”

We assessed age, gender identity, race/ethnicity, primary language, education, weekly earnings, recruitment location, past-month homelessness or, unstable housing [22], recent incarceration, ownership of belongings (current phone ownership, loss of phone access in the past 3 months), current measures of mental health (anxiety and depressive symptoms using PHQ-4 [23], perceived social support using Multidimensional Scale of Perceived Social Support (MSPSS) [24]), access and utilization of the healthcare system (did not feel welcome in medical offices/clinics in general, past-year difficulty accessing medical or social services, past-year treatment for infections due to IDU, current medication for opioid use disorder). We analyzed substances used in the past month (opioids, stimulants, downers/sedatives), past-month injection frequency, and past 3-month drug-related overdose. Both measures of mental health rely on previously validated measures and maintain high internal consistency (Cronbach's a > 0.80).

Notably, the recall period for incarceration history differed by assessment (3-month recall at baseline assessment, 6-month recall at subsequent assessments).

Statistical Analysis

Results are stratified by the experience of an encampment sweep (yes/no) in the past year. Continuous variables are shown with medians and interquartile ranges. Categorical variables are shown with frequencies and column percentages. Groups were compared using chi-square, Fisher’s exact, or Wilcoxon rank sum tests as applicable. P-values of ≤ 0.05 were considered statistically significant. Results restricted to those who reported homelessness or unstable housing in the past month (N = 136) are found in the Supplementary information.

Results

Sociodemographic and Participant Characteristics

The median age was 40 years, and 53% identified as cisgender male (Table 1). Almost three-quarters (72%) identified as White and non-Hispanic, and 88% reported being homeless or unstably housed in the last month. Median weekly earnings reported were approximately $105 (IQR $0, $350). Those who reported being homeless or unstably housed in the past month were more likely to experience an encampment sweep (p = 0.001).

Table 1 Sociodemographics, participant characteristics, and other key variables of interest among a cohort of 155 people in Massachusetts, stratified by experience of relocation or encampment sweep in the last year
Full size table

Experience of Encampment Sweep

In our sample, 38% (n = 59) experienced an encampment sweep in the past year, with the vast majority (n = 57, 97%) losing belongings during relocation(s). Among those who experienced an encampment sweep, 73% reported difficulty accessing health or social services or benefits as a result of the sweep (Table 1).

Ownership of Belongings

Seventy-nine percent of participants reported current ownership of a phone. Among that group (n = 123), 70% reported losing access to a phone or phone service in the past 3 months. This proportion differed significantly by experience of encampment sweep (81% among those who experienced a sweep versus 64% among those who did not; p = 0.04).

Current Measures of Mental Health

Median PHQ-4 scores (overall median = 7) differed significantly by experience of sweeps; those not relocated in the past year had median scores corresponding to mild symptoms of anxiety and depression, whereas those who had been relocated in the past year had median scores corresponding to severe symptom levels (p = 0.003). Participants reported very low levels of perceived social support from family, friends, and significant others across experiences of encampment sweep (MSPSS score of 9.8).

Access and Utilization of the Healthcare System

Twenty-six percent of participants reported not feeling welcome in medical offices/clinics in the past year, which differed by experience of encampment sweep (39% among those who reported an encampment sweep versus 18% among those who did not; p = 0.003).

Substance use History

Over 90% of participants reported past-month use of stimulants and/or opioids. Significantly more participants who reported using downers or sedatives had experienced a sweep (80% of those relocated versus 62% of those not relocated). Lastly, 56% reported injecting drugs more than once every day, and 21% reported a drug-related overdose in the past 3 months; results did not differ by experience of encampment sweep.

Discussion

Understanding the associations between the experience of encampment sweeps and suboptimal mental health symptoms, access to medical services, and loss of phones or phone services is vital to building robust, relevant, and thoughtful interventions for this marginalized population. In our sample, nearly two in five participants reported past-year relocation due to encampment sweeps. Most also reported losing belongings and having more difficulty accessing health or social services as a result of the sweeps. Those who had experienced a sweep were more likely to have higher levels of anxiety and depressive symptoms, not feel welcome in medical offices/clinics, be homeless or unstably housed, and report using downers or sedatives in the past month. There was little observed difference in substance use patterns, including injection frequency, by experience of encampment sweep, possibly illustrating that continuing to use substances may be a priority regardless of police presence or experience of encampment sweeps.

The high prevalence of the loss of phone or phone services (81% among those who experienced an encampment sweep) underscores how this form of relocation disrupts material stability and social supports (i.e., disrupts social connections, unable to make medical appointments). Observed low levels of social support and high levels of anxiety and depression among the full sample are discouraging and highlight a need for public health interventions to support social connection.

This study has limitations. Self-reporting stigmatized behaviors (i.e., IDU) via interviewer-administered questionnaires may introduce interviewer and/or social desirability bias, which may lead to the underreporting of stigmatized behaviors. Additionally, recent incarceration may be overestimated, as some participants reported on the past 3 months (n = 116), while others reported on the past 6 months (n = 39). Homelessness is likely underestimated, as the recall period was restricted to the past month. We did not pursue regression analyses due to the descriptive nature of this study and small sample sizes. P-values should be interpreted with caution.

Findings have implications for public health policy and practice. Although MA has implemented substantial harm reduction infrastructure and “Right to Shelter” protections, the continuation of encampment sweeps, police-led or otherwise, may undermine those same public health goals. Interventions to prioritize stable housing, access to harm reduction services, and non-punitive supports are needed to reduce the preventable harms associated with encampment sweeps.

Conclusions

Encampment sweeps do not happen in isolation; they are policy-driven disruptions with tangible and measurable preventable negative health consequences. Public health efforts must align with strategies that support, rather than punish, people experiencing homelessness and those who use drugs.