Unsheltered homelessness and health: A literature review

  • Jessica Richards
    Department of Community Health Sciences, Jonathan and Karin Fielding School of Public Health, UCLA, 650 Charles E. Young Drive South, Los Angeles, CA 90095-1772
    Search for articles by this author
  • Randall Kuhn
    Corresponding author: Randall Kuhn. (303) 928-9111.
    Department of Community Health Sciences, Jonathan and Karin Fielding School of Public Health, UCLA, 650 Charles E. Young Drive South, Los Angeles, CA 90095-1772
    Search for articles by this author
Open AccessPublished:October 28, 2022DOI:

      Highlights for “Unsheltered homelessness and health: A literature review”

      • Unsheltered homelessness is rising, particularly in the Western U.S.
      • We reviewed studies that compared health outcomes by shelter status.
      • Unsheltered have higher rates of physical and mental illness, and substance use.
      • Unsheltered populations use health care less and often lack health insurance.
      • Longitudinal studies are needed to assess causes of poor health among unsheltered.



      In recent years, cities across the world have seen widespread growth in unsheltered homelessness, in which a person sleeps “in a place not meant for human habitation” such as cars, parks, sidewalks, and abandoned buildings. It is widely understood that people experiencing homelessness (PEH) have poorer health outcomes than the general population. Less is known about the health of people who are unsheltered, yet they may be exposed to greater health risks than their sheltered counterparts. The purpose of this literature review is to evaluate and summarize the evidence on unsheltered homelessness and health.


      A literature search was conducted using PubMed to identify publications on unsheltered homelessness and health. A total of 42 studies were included for review.


      Unsheltered populations experience higher rates of chronic disease, serious mental illness, and substance abuse compared to sheltered populations. Unsheltered homelessness is strongly associated with chronic homelessness that exacerbates serious mental illness and substance use, which is often co-occurring. Despite having large unmet health needs, unsheltered populations have lower health care utilization and often lack health insurance.


      Evaluating the impact of shelter status on health outcomes has important implications for the allocation of housing and health services. Longitudinal studies are needed to examine the relationship between duration of sheltered and unsheltered homelessness and health outcomes and explore mediating mechanisms that lead to poor health among unsheltered populations. In spite of these limitations, our results also suggest an urgent need to address the unique and severe challenges facing unsheltered populations and the need for intervention approaches that are sensitive to these unique disease burdens.

      Graphical Abstract



      In recent years, cities across the world have seen widespread growth in unsheltered homelessness [1]. Above and beyond the epidemiologic risks associated with homelessness itself, [2–4] unsheltered individuals may experience additional disease burdens relating to exposures such as violence, exploitation, weather, pollution and poor sanitation. Yet few studies have established the health consequences of unsheltered homelessness, much less the extent or mechanisms underlying this effect [5]. This literature review evaluates and summarizes the small but growing body of literature on health outcomes among unsheltered homeless adults, specifically in comparison to those who are sheltered.


      The past decade has seen a sizable increase in the proportion of the US homeless population who are unsheltered. The US Department of Housing and Urban Development (HUD) defines an individual as homeless if they lack a fixed, regular, and adequate nighttime residence [1]. Within this category, those who sleep in a public or private place not meant for human habitation (e.g., street, tent, or other makeshift shelter) are considered unsheltered. According to HUD's Annual Homeless Assessment Report (AHAR), the unsheltered homeless population increased by 30% from 2015-2020, even as the sheltered homeless population declined by 10%. As a result, the share unsheltered nationwide rose from 31% to 39%. Aggregate data mask wide geographic variations in the distribution of unsheltered homelessness, with higher rates in the Western US. While it is widely assumed that unsheltered homelessness results from warmer weather, Figure 1 suggests a simpler relationship whereby localities with more shelter beds will have a lower share of their homeless population be unsheltered [6]. Indeed, New York City had much higher rates of unsheltered homelessness in the 1990s, which were reduced as a result of a concerted effort to build shelters and engage clients [7].
      Figure 1
      Figure 1Relationship of homeless shelter bed inventory to unsheltered homelessness, U.S. Continuums of Care, 2020.
      Source: US Department of Housing and Urban Development Point-in-Time Count and Housing Inventory Count Data, 2020, HUD Exchange.
      Recent increases in unsheltered homelessness partially track a series of rulings by the US 9th Circuit Court of Appeals, which struck down urban camping prohibitions until shelter beds were made available to house the entire homeless population.[8]
      Countries define and measure homelessness differently making cross-country comparison of unsheltered homelessness difficult. The definition of homelessness varies by country and identification of the unsheltered homeless population is beset by limited data sources and methodological challenges [9]. In some countries the definition of homelessness may include only people who are unsheltered (e.g., Japan), unsheltered and sheltered populations (e.g., US), or be applied broadly to also include precariously housed populations (e.g., Australia). Countries with a more inclusive definition of homelessness tend to report a higher incidence of homelessness [10]. Although the unsheltered are commonly included in official definitions of homelessness, they are often not separately identifiable in national indicator data outside of the United States [11]. Unsheltered homelessness has been referred to as ‘street homelessness,’[12] ‘absolutely homeless,’[13] ‘rooflessness,’ ‘sleeping rough,’[14] or ‘long grassing’[15] and in turn, individuals have been referred to as ‘rough sleepers,’ ‘street/pavement dwellers,’ and ‘encampment residents.’ Many individuals may also live in both sheltered and unsheltered locations at different points in time, or even at the same time.[16] These differences in operationalizing homelessness will be considered when drawing comparisons across studies.
      Where possible, we focus on studies that disentangle the effects of unsheltered homelessness from confounding factors associated with unshelteredness. Compared to those who are sheltered, the unsheltered are more likely to be non-Hispanic White, male, a veteran and to have a history of history of incarceration or foster care [5,17–20]. Duration may also serve as a confounder because unsheltered homelessness is associated with prolonged and more frequent episodes of homelessness.[5,17,18,20]
      Our review pays particularly close attention to chronic health conditions affecting older adults. Recent studies have framed the long-term consequences of homelessness in terms of accelerated aging due to repeated exposure to deprivation and disease, as reflected in the early onset of geriatric conditions and surgical complication risks often occurring decades sooner than in housed older adults.[21,22] The population of older homeless adults (aged 50 years and older) is also growing in the U.S. due to ongoing cohort effects.[23]


      We organized the review following Fazel and colleagues’ 2014 review of health outcomes for the broader homeless population.[2] Results were grouped into the following health outcomes: mortality, non-communicable diseases, reproductive health, communicable disease, mental health, substance abuse, health services utilization and injuries.

      Search strategy

      The literature search strategy and study selection are summarized in Figure 2. A literature search was conducted (in May 2020) using PubMed to identify publications on unsheltered homelessness from 1990 through 2020. Combinations of relevant keywords including unshelter* and rough sleeper* and street homeless* were used to capture variations of unsheltered homelessness. A total of 13,415 publications were identified. Keywords for each search are listed in Appendix Table 1. Grouping, de-duplication and coding were conducted in EndNote to minimize risk of errors or lost data. The reviewer (JR) consulted with an experienced biomedical staff librarian to translate PubMed searches into EndNote smart groups (Appendix Table 2). Publications on unsheltered homelessness were identified as the intersection of publications identified using keywords for homelessness and publications identified using keywords for unshelteredness (Appendix Table 2). A total of 174 publications were identified for screening to identify studies in which a health or health utilization measure was the outcome of interest. To address the potential risk of publication bias, we conducted parallel searches of the grey literature using Google Scholar, and found no additional studies.
      Figure 2
      Figure 2Summary of search strategy and study selection.

      Study selection

      The full text for all 174 publications were located and read to determine eligibility for review. Reference lists were also searched to identify relevant publications and an additional 31 publications were screened for review. Following earlier studies, the review included only studies which estimated an association between unsheltered homelessness and a health-related outcome[24]. Studies that did not clearly meet the inclusion criteria were discussed with a second reviewer (RK) and resolved by consensus. For transparency, the list of 132 excluded studies are included in Appendix Table 3 along with the criteria for exclusion.

      Study coding

      Included studies were grouped into two tiers: 1) comparative studies of unsheltered and sheltered homelessness and 2) studies with an exclusively unsheltered sample. Tier 2 evidence was only reported if it contradicted Tier 1 evidence or if Tier 1 evidence did not exist for a health domain.
      Due to a wide range of methodologic rigor within the included studies, a scoring system was developed to evaluate study quality within each tier. Specifically, papers were scored on: 1) the rigor of the sampling strategy, 2) the use of validated health measures, and 3) efforts to control for or otherwise account for (e.g., via standardization) the role of population composition. Initially, sampling rigor was coded by the use of probability sampling of a known population. Few studies met this requirement, however. So the standard was relaxed to code studies based according to the following hierarchy: 1) probability sampling or quota sampling occurred at a mix of known venues, and/or efforts were taken to compare the resulting sample to more representative samples of the population of interest (e.g. point-in-time homeless counts), 2) convenience samples where the sample was not selected on the basis of health or health risk (i.e., certain neighborhoods of a city), and 3) convenience samples where the sample was selected on the basis of health risk (i.e., substance abuse program). Studies were independently scored by both reviewers (JR and RK) and discrepancies were resolved by consensus. Studies with a score of 1 were excluded from analysis. The inclusion criteria and scoring are listed in Appendix Table 4. In all, 42 publications were selected for review (Tier 1: 27, Tier 2: 15).


      The review included 42 studies, 13 of them Tier 1 comparative studies with quasi-representative sample design, 14 Tier 1 comparative studies with convenience samples, and 15 Tier 2 studies with unsheltered samples only (Supplementary Electronic Appendix). Results for Tier 1 studies are reported using adjusted (if available) odds/risk ratios for the unsheltered vs. sheltered comparison. For Tier 2 studies, we report unadjusted estimates for the unsheltered population.
      Figure 3 shows that more than half of the studies, including all Tier 1 representative studies, were conducted in the past decade. Nearly half of studies (20 of 42) were published since 2016, including 10 of the 13 comparative quasi-representative studies. Nearly three-quarters of the studies took place in the U.S. (29/42), including 85% of studies published in the past five years (17/20) and 85% of comparative quasi-representative studies (11/13).
      Figure 3
      Figure 3Included studies by year and quality.
      We extracted 48 unique directional associations from the 27 Tier 1 comparative studies we reviewed (Figure 4). We summarized the type of relationship supported for each finding for all health outcomes. We classified results as those supporting substantially poorer health for unsheltered (statistically significant with RR/OR >= 2.0), somewhat poorer health for unsheltered (significant with RR/OR between 1.0 and 2.0), and not statistically significant. No Tier 1 studies provided support for unsheltered individuals having better health than sheltered comparators. Across all 48 Tier 1 findings, 44% (21 out of 48) found that those who are unsheltered had much poorer health, 29% (14/48) supported moderately but significantly poorer health, and 27% (13/48) found no significant relationship. Within each of the eight health outcome groups, a majority of findings indicated significantly worse health for unsheltered vs. sheltered, though the number and quality of findings varied by outcome.
      Figure 4
      Figure 4Classification of single associations for each health outcome by type of relationship support and outcome of interest.

      Mortality and cause of death

      Mortality rates are significantly higher among those experiencing unsheltered homelessness. Compared to a sheltered homeless cohort, the standardized mortality ratio (SMR) for an unsheltered homeless cohort in Boston was nearly three (SMR=2.7; 95% CI, 2.3-3.2) times higher.[25] After grouping the unsheltered sample by mortality risk factors, the SMR for unsheltered vs. sheltered was 4.0 (95% CI, 3.0-5.2) times higher for a high-risk group and 2.2 (95% CI, 1.8-2.8) times higher for the lower-risk cohort.[26] A national study using data from the 100,000 Homes campaign found a significant but much smaller effect of unsheltered status on the odds of mortality (AOR=1.12; 95% CI, 1.05-1.19).[17]
      Three papers addressed cause of death.[25–27] Common causes of death among those who were unsheltered included chronic disease, substance use, and injuries.[25] Compared to a sheltered cohort, non-poisoning injuries (e.g., motor vehicle accidents, falls, drowning) were high (SMR=7.1; 95% CI, 4.4-11.0) and associated with high rates of chronic substance use (SMR=4.2; 95% CI, 2.5-6.7).[25] Unsheltered adults classified as high risk had substantially higher mortality rates for HIV/AIDS (SMR=122.3; 95% CI, 44.8-271.1), chronic substance use (SMR=104.2; 95% CI, 38.1-231.0) primarily caused by alcohol abuse, chronic liver disease (SMR=86.0; 95% CI, 45.0-150.0), and injuries (SMR=44.0; 95% CI, 17.8-91.6) compared to a sheltered high-risk cohort.[26]

      Non-communicable diseases and associated markers

      Unsheltered populations often experience poor adult health outcomes. A cross-sectional study of homeless women in Los Angeles found that after controlling for sociodemographic factors (e.g., age, education, ethnicity, number of times homeless and length of time homeless) unsheltered women had greater odds of fair or poor physical health (AOR=3.40; 95% CI, 2.34-4.94; p=0.001) and experiencing pain in the last 6 months (AOR=2.28; 95% CI, 1.54-3.37; p=0.001) than sheltered homeless women.[28] The association between unsheltered status and worse health status remained significant for women who used substances (AOR=3.0; 95% CI, 2.02-4.45) and women with poor mental health (AOR=2.24; 95% CI, 1.49-3.37).[28] Another study conducted in Wales found that oral health-related quality-of-life was significantly poorer among rough sleepers compared to sheltered homeless adults (p=0.004).[29] Other common self-reported medical problems among the unsheltered include orthopedic problems and arthritis.[30]
      Evidence is mixed regarding the impact of shelter status and chronic homelessness on chronic health conditions.[17,31,32] A study of chronically homeless veterans found that sheltered veterans were more likely to have a chronic health condition than unsheltered (43.4% vs. 40.8%; p=0.001), though this result did not control for confounders.[19] While a comparative study of chronically unsheltered and not chronically unsheltered adults did not find significant differences in the rates of serious medical conditions, chronically unsheltered adults were significantly more likely to report tri-morbidity (e.g., serious medical issue, lifetime mental illness, lifetime substance abuse) than not chronically unsheltered adults (OR=1.65; 95% CI, 1.11-2.45).[20] A study of air pollution-related health outcomes among people experiencing homelessness found that breathing difficulty and headaches did not vary significantly based on shelter status (shelter vs. unsheltered) or chronicity (chronic vs. non-chronic homelessness).[33]
      A number of studies focused on older homeless adults.[19,20] One study found significantly higher rates of vision impairment among unsheltered older adults (p=0.04).[21] However, mobility impairment (defined as self-reported difficulty walking) and other geriatric conditions (ADL impairment, IADL impairment, one or more falls in past 6 months, cognitive impairment, hearing impairment, urinary incontinence, or depression) did not vary significantly by living environment.[21] Among unsheltered older adults, 34% reported falling once or more in the last 6 months,[21] and spending any night unsheltered (compared to none) was significantly associated with increased odds of falling (AOR=1.42; 95% CI, 1.10-1.83).[34] Unsheltered older adults had more than twice the odds of very low food security compared to older adults who were recently homeless, staying in a temporary accommodation, or institution.[35]

      Communicable disease

      Just four papers addressed communicable disease, with just two Tier 1 studies and two conducted in the United States. One paper addressed unsheltered-sheltered differences in tuberculosis risk and treatment.[36] People living on the street had the greatest risk of tuberculosis compared to housed and other homeless groups (sheltered and transient) and the average days hospitalized and required for follow-up care was four times as high compared to sheltered homeless.[36] Two descriptive studies conducted in Ethiopia observed that 44-68% of street dwellers had one or more intestinal parasites.[37,38]

      Sexual and reproductive health

      Although only three studies examined sexual and reproductive health, results suggest that some high-risk sexual behaviors and rates of sexually transmitted diseases may be more common, particularly for women. In the Los Angeles study, unsheltered women had greater odds of having multiple sex partners (AOR=2.79; 95% CI, 1.93-4.03; p=0.001); having a sexually transmitted disease (AOR=2.10; 95% CI, 1.05-4.21; p=0.036) in the past six months.[28] They were also more likely to have experienced an unwanted pregnancy (AOR=1.53; 95% CI, 1.07-2.19; p=0.021).[28] Among unsheltered women experiencing reproductive health problems in India, most (78.5%) did not seek care. [39] A descriptive study of street dwellers in Ethiopia indicated that 39.4% had experienced sexually transmitted disease symptoms in the past year.[40]

      Mental health

      Unsheltered homelessness is often accompanied by high rates of mental health illness, including major depression. In the Los Angeles study, unsheltered women had much greater odds of being in poor mental health (AOR=12.69; 95% CI, 6.68-24.13; p=0.001) than sheltered homeless women.[28] A study of unsheltered adults in Japan found street homelessness was significantly associated (OR=2.64; 95% CI, 1.15-6.06; p<0.05) with recent suicidal ideation after controlling for depression.[12] Additionally, in a descriptive study of unsheltered adults in Ethiopia, 41.8% wished to die, 21.7% had persistent suicidal thoughts, and 14.8% had attempted suicide in the past month.[14] But another study found that rates of lifetime major depression did not vary significantly between unsheltered and sheltered men.[41] A study of older homeless adults found that rates of depression and suicidal thoughts did not vary significantly between unsheltered and those in other living groups, but the study did not include a clearly sheltered comparison group.[21]
      In addition to major depression, schizophrenia and mood disorders are common mental health diagnoses among unsheltered populations. High prevalence rates have been found in descriptive studies across three countries. Most Brazilian unsheltered adults had a psychiatric diagnosis (98.8%), an Ethiopian study found 41.0% had psychosis, and among a small sample of rough sleepers in Dublin, 31.3 % had severe mental illness.[14,42,43] Among those with mental illness, the most common diagnoses were schizophrenia (88% in Ethiopia, 25% in Dublin, 9.6% in Brazil) and mood disorders, including major depression (32.5% in Brazil).[14,42,43]
      Chronicity may contribute to rates of mental illness among the unsheltered. Exploratory analysis indicated that adults with psychosis in Ethiopia were more likely to be older and to have longer durations of street homelessness.[14] Chronically unsheltered individuals were more likely to have lifetime mental illness (defined as either a history of psychiatric hospitalization or current mental health counseling or treatment) (OR=1.57; 95% CI, 1.19-2.08) than not chronically unsheltered persons.[20] In contrast, a recent study found mental health illness did not vary by shelter status (shelter vs. unsheltered) or chronicity (chronic vs. non-chronic homelessness).[33]

      Substance use

      Experiencing unsheltered homelessness can significantly increase the likelihood of substance use disorder and worsen existing substance use. In one study, unsheltered women were more likely to use alcohol or non-injection drugs in the past 6 months (AOR=2.95; 95% CI, 1.94-4.50) than their sheltered counterparts, but no significant differences were found regarding recent injection drug use.[28] Among opioid-using veterans in New York City, being unsheltered was a significant predictor (adjusted incidence rate ratio (AIRR)=2.08; 95% CI, 1.39-3.13) of greater engagement in opioid overdose risk behaviors after adjusting for demographics and prescription medications.[44] A study of sheltered and unsheltered tobacco users found few differences in tobacco use in the prior month, with the exception that unsheltered smokers were significantly more likely to report using large cigars (AOR=2.35; 95% CI, 1.05-5.23).[45] A study of young adults found that sheltered young adults were significantly less likely to have used alcohol (AOR=0.53; 95% CI, 0.34-0.82), marijuana (AOR=0.53; 95% CI, 0.34-0.83), and synthetic marijuana (AOR=0.46; 95% CI, 0.25-0.84) in the past month than unsheltered participants after controlling for other factors. Rates of stimulant and opioid use were also lower but were not statistically significant.[46]
      Among a cohort of out-of-treatment substance users, a composite HIV risk score (number of times injecting drugs, number of days using crack, and number of days having sex) was significantly higher for all unsheltered subgroups (Black males, Hispanic males, White males, Black females, Hispanic females) except for White females (p<0.05) than sheltered participants.[47] Another national study of substance users found unsheltered homelessness was strongly associated with frequent public drug use (AOR=17.44; 95% CI, 9.5-32.0) compared to stably housed participants after controlling for age and use of heroin or injection drugs.[48]
      Comorbid mental health and substance use disorder is also common among unsheltered populations.[14,42,43] Chronically unsheltered adults were more likely to have a dual diagnosis of mental illness and substance use (p=0.002) than not chronically unsheltered adults.[20] Additionally, veterans with a substance use disorder, alone or in combination with a mental illness, were significantly more likely to be unsheltered, though this did not apply to veterans with only a mental illness.[19] The Los Angeles study found that unsheltered women with recent substance use had much greater odds (AOR=11.09; 95% CI, 5.62-21.88) of poor mental health compared to sheltered women with recent substance use.[28]
      The most used substance among unsheltered populations is alcohol [30,49,50]. Rates of alcohol use were high, 68% of rough sleepers in London[49] and 72% of encampment residents in Los Angeles[30] reporting using alcohol in the past month. Other commonly used drugs among the unsheltered include crack cocaine, heroin, and cannabis.[30,49,50] Drugs are often used in combination with unsheltered individuals using an average of 3 or 4 drugs.[50]
      There is evidence that substance use may increase with duration of unsheltered homelessness. Among unsheltered adults in London, longer durations of unsheltered homelessness were accompanied by increased daily substance use, injection drug use, and dependency.[50] Substance use was a commonly reported cause of homelessness and 80% reported using at least one additional drug since homelessness onset.[49]


      Little evidence documents levels of injury risk for unsheltered and sheltered individuals. In the Los Angeles study, unsheltered women were more likely to have experienced victimization in the form of physical assault (AOR=2.74; 95% CI, 1.91-3.94; p=0.001) and robbery (AOR=5.37; 95% CI, 3.64-7.92; p=0.001) than sheltered women.[28] Though half of unsheltered homeless adults in Manhattan reported a history of repeated trauma, rates did not significantly vary between chronically unsheltered and not chronically unsheltered homeless adults.[20]

      Health services

      Health care access and utilization

      Unsheltered homelessness has been associated with lower rates of health care utilization. Not residing on the streets was significantly associated with recent health services utilization (AOR=11.39; 95% CI, 3.58-36.24) after adjusting for socioeconomic factors and other covariates in a study based in South Korea.[51] After adjusting for sociodemographic and homelessness characteristics, women experiencing unsheltered homelessness were less likely to have seen a dentist in the past year (AOR=0.34; 95% CI, 0.21-0.53; p=0.001) and to have received a pap test (AOR=0.40; 95% CI, 0.28-0.59; p=0.001) or a TB test (AOR=0.22; 95% CI, 0.15-0.33; p=0.001) compared to their sheltered counterparts.[28] After adjustment, unsheltered individuals in England were less likely to be registered with a general practitioner (AOR=0.45; 95% CI, 0.30-0.66) compared to sheltered participants but were not less likely to utilize primary care services.[52] Unsheltered older adults had significantly lower rates of primary care services compared to older homeless adults in other residential categories.[53] Yet regarding follow up care, unsheltered clinic users were more likely (1.45 times) to return for care than those staying in a sheltered environment after controlling for other factors.[54]
      Findings for hospital-related health service use among unsheltered homeless individuals is mixed. Homeless veterans in Los Angeles with a history of sleeping unsheltered had lower odds of using inpatient services (OR=0.34; p=0.002) compared to sheltered veterans,[55] while a bivariate analysis found no difference in use of outpatient services among homeless veterans based on shelter status.[19] Another study found higher rates of outpatient services among an unsheltered cohort, but less use of emergency services and fewer hospital admissions compared to sheltered adults.[56] One study found higher rates of health services utilization among the unsheltered, including emergency department and outpatient services, but this effects was primarily explained by high levels of chronic health conditions.[5] A study of rough sleepers in England also found no association between shelter status and use of hospital care, hospital admissions, emergency services, or ambulance use after adjusting for covariates.[52] Unsheltered populations are also less likely to have health insurance.[19]
      Among current or past drug users, unsheltered women were less likely to have sought formal treatment in their lifetime than those sheltered (AOR=0.31; 95% CI, 0.21-0.47).[28] In an unadjusted analysis, unsheltered homeless persons were significantly more likely to report not receiving needed substance use treatment compared to sheltered homeless persons (61.0% vs. 45.6%; p<0.001).[57]


      In spite of the mixed quality of reviewed studies, our review suggests a consistent and strong association between unsheltered homelessness and higher levels of health risk, above and beyond the well-documented negative consequences of homelessness.[2] Unsheltered populations experience high rates of chronic disease, serious mental illness and substance abuse compared to sheltered populations. Despite having many unmet health needs, unsheltered populations have lower health care utilization and often lack health insurance. These health disadvantages manifest in significantly higher burdens of mortality.[26] Unsheltered homelessness is strongly associated with chronic homelessness, which exacerbates serious mental illness and substance use, which are often co-occurring.
      We note a number of concerns surrounding methodological quality. First, few studies addressed the critical outcomes of injuries, communicable disease and sexual and reproductive health. Second, we observed substantial variation in the definition of unsheltered and of comparison groups, and so comparisons between sheltered and unsheltered populations should be interpreted with care. A variety of measures have been used to determine shelter status. Unsheltered homelessness has been based on current living situation,[58] prior night location,[46] and having slept within a certain area.[40] Other studies incorporate duration data by gathering residential histories to identify where participants sleep most of the time. Residential time windows range from within the past one week,[45] one month,[28] three months[57] and even six months.[31] Finally, we note the variable quality of the sampling methods. Few studies used a probability sample. Only half of the comparative studies reviewed employed methods at gaining quasi-representativeness, either by sampling through multiple venues or comparing the sample to point-in-time estimates of target population composition. Some recent studies with consistently higher-quality evidence have produced more mixed results.
      Future studies must employ longitudinal designs to address causal mechanisms linking unsheltered homelessness to health via specific pathways of risk. The potential for reverse causality underscores the need for longitudinal studies to explore temporal relationships. During the pandemic, many jurisdictions have targeted vulnerable individuals for placement in shelters or hotels, potentially altering the temporal relationship between health and unshelteredness [59,60]. Yet at the same time, homeless services systems often impose rules or restrictions that cause vulnerable groups to avoid or be removed from shelters. This can include well-known restrictions such as sobriety requirements as well as more subtle barriers such a pet ownership restrictions that may exclude even those with service animals [61].
      An even more important need is to identify and address the specific social-ecological exposures that drive poor health among the unsheltered, and how these exposures interact with shelter status. Only one paper in this review attempted to isolate a causal factor in the relationships between health and sheltered/unsheltered homelessness.[41] Notably, risk factors like chronic exposure to low quality food and sleep disruption due to light and noise pollution have not been addressed at all. Future studies should leverage longitudinal data where possible and disaggregate people experiencing homelessness by shelter status to further explore mechanisms that drive poor health among unsheltered populations. A better understanding of these mechanisms would improve the ability to target street medicine and other street-based services towards impactful interventions.
      We also note that homeless individuals with multiple marginalized identities (e.g., racial/ethnic, gender, and sexual minorities) may be at heightened risk of poor health.[62,63] More research is needed to understand how social inequalities by race, gender and sexuality interact to shape health outcomes among people experiencing homelessness, especially unsheltered homelessness. This will require samples of adequate size to conduct stratified analysis and more rigorous methods to ensure representation.
      A number of interventions offer the potential to improve the welfare of unsheltered individuals while also engaging clients on a pathway to housing. Street medicine programs can deliver much-needed services and engage clients with service systems, though challenges remain including identification of high-impact service packages, coordinating care across fragmented service providers and ensuring sustainability [64]. Mobile phones are widely used among unsheltered individuals and can serve as a lifeline for emergency services and case worker outreach [65–67], but interventions are needed to improve access to connectivity and charging [68] and to develop equity-sensitive digital service delivery models [69]. Some interventions can increase the safety of unsheltered living arrangements, such as safe parking or camping areas [70]. Finally, evidence from Japan suggests that cash transfers delivered through basic income programs or benefit enrollment initiatives may hasten the transition from the streets [71]. Ultimately, however, any durable solution will require increased shelter inventory and a better understanding of the barriers to shelter entry [61].


      Unsheltered populations experience high rates of chronic disease, serious mental illness and substance abuse compared to sheltered populations. Unsheltered homelessness is strongly associated with chronic homelessness that exacerbates serious mental illness and substance use, which is often co-occurring. Rates of premature mortality are high relative to sheltered populations and older adults are particularly vulnerable due to accelerated aging while on the street. Despite having high unmet health needs, unsheltered populations have lower health care utilization and often lack health insurance. Results are consistently positive for most health conditions, but the evidence quality is mixed. Future research should include longitudinal studies that account for timing and duration of homelessness, explore specific causal mechanisms of impact, and address intersectionality with race/ethnicity, sex/gender, sexual orientation and other marginalized identities. While further research is desperately needed, our results also suggest an urgent need to address the unique and severe challenges facing unsheltered populations and the need for intervention approaches that are sensitive to these challenges.

      CRediT Author Statement

      Jessica Richards: Conceptualization, Methodology, Software, Validation, Formal analysis, Resources, Data curation, Writing – original and draft, Writing – review and editing, Visualization
      Randall Kuhn: Conceptualization, Methodology, Validation, Formal analysis, Resources, Data curation, Writing – original and draft, Writing – review and editing, Visualization, Supervision


      • 1
        Henry M, de Sousa T, Roddey C, Gayen S, Bednar T. The 2020 Annual Homeless Assessment Report (AHAR) to Congress. 2020;102.
      • 2
        Fazel S, Geddes JR, Kushel M. The health of homeless people in high-income countries: Descriptive epidemiology, health consequences, and clinical and policy recommendations. The Lancet. 2014;384:1529–40.
      • 3
        Morrison DS. Homelessness as an independent risk factor for mortality: results from a retrospective cohort study. Int J Epidemiol. 2009;38:877–83.
      • 4
        Baggett TP, Hwang SW, O'Connell JJ, Porneala BC, Stringfellow EJ, Orav EJ, et al. Mortality Among Homeless Adults in Boston: Shifts in Causes of Death Over a 15-year Period. JAMA Intern Med. 2013;173:189–95.
      • 5
        Petrovich JC, Hunt JJ, North CS, Pollio DE, Roark Murphy E. Comparing Unsheltered and Sheltered Homeless: Demographics, Health Services Use and Predictors of Health Services Use. Community Ment Health J. 2019/09/26 ed. 2020;56:271–9.
      • 6
        Corinth K, Lucas DS. When warm and cold don't mix: The implications of climate for the determinants of homelessness. J Hous Econ. 2018;41:45–56.
      • 7
        De Blasio B, Palacio H, Banks S. Turning the Tide on Homelessness in New York City. City N Y.:128.
      • 8
        O'Flaherty B. Homelessness research: A guide for economists (and friends). J Hous Econ. 2019;
      • 9
        Mosites E, Morris SB, Self J, Butler JC. Data Sources That Enumerate People Experiencing Homelessness in the United States: Opportunities and Challenges for Epidemiologic Research. Am J Epidemiol. 2021;190:2432–6.
      • 10
        OECD. Better data and policies to fight homelessness in the OECD, policy brief on affordable housing. OECD Paris; 2020;
      • 11
        Busch-Geertsema V, Benjaminsen L, Hrast MF, Pleace N. Extent and profile of homelessness in European Member States: A statistical update. FEANTSA; 2014;
      • 12
        Okamura T, Ito K, Morikawa S, Awata S. Suicidal behavior among homeless people in Japan. Soc Psychiatry Psychiatr Epidemiol. 2013/11/12 ed. 2014;49:573–82.
      • 13
        Stergiopoulos V, Dewa CS, Tanner G, Chau N, Pett M, Connelly JL. Addressing the needs of the street homeless: A collaborative approach. Int J Ment Health. 2010;39:3–15.
      • 14
        Fekadu A, Hanlon C, Gebre-Eyesus E, Agedew M, Solomon H, Teferra S, et al. Burden of mental disorders and unmet needs among street homeless people in Addis Ababa, Ethiopia. BMC Med. 2014/08/21 ed. 2014;12:138.
      • 15
        Topp L, Iversen J, Baldry E, Maher L. Housing instability among people who inject drugs: results from the Australian needle and syringe program survey. J Urban Health. 2012/06/27 ed. 2013;90:699–716.
      • 16
        Bevitt A, Chigavazira A, Herault N, Johnson G, Moschion J, Scutella R, et al. Journeys Home Research Report No. 6. Australian Government Department of Social Services; 2015.
      • 17
        Montgomery AE, Szymkowiak D, Marcus J, Howard P, Culhane DP. Homelessness, Unsheltered Status, and Risk Factors for Mortality: Findings From the 100 000 Homes Campaign. Public Health Rep. 2016;131:765–72.
      • 18
        Montgomery AE, Szymkowiak D, Culhane D. Gender Differences in Factors Associated with Unsheltered Status and Increased Risk of Premature Mortality among Individuals Experiencing Homelessness. Womens Health Issues. 2017/05/01 ed. 2017;27:256–63.
      • 19
        Byrne T, Montgomery AE, Fargo JD. Unsheltered Homelessness Among Veterans: Correlates and Profiles. Community Ment Health J. 2016;52:148–57.
      • 20
        Levitt AJ, Culhane DP, DeGenova J, O'Quinn P, Bainbridge J. Health and Social Characteristics of Homeless Adults in Manhattan Who Were Chronically or Not Chronically Unsheltered. Psychiatr Serv. American Psychiatric Publishing; 2009;60:978–81.
      • 21
        Brown RT, Hemati K, Riley ED, Lee CT, Ponath C, Tieu L, et al. Geriatric Conditions in a Population-Based Sample of Older Homeless Adults. Gerontologist. 2016/02/28 ed. 2017;57:757–66.
      • 22
        Salem BE, Nyamathi A, Brecht M-L, Phillips LR, Mentes JC, Sarkisian C, et al. Constructing and identifying predictors of frailty among homeless adults—A latent variable structural equations model approach. Arch Gerontol Geriatr. 2014;58:248–56.
      • 23
        Culhane D, Treglia D, Byrne T, Metraux S, Kuhn R, Doran K, et al. The emerging crisis of aged homelessness: Could housing solutions be funded by avoidance of excess shelter, hospital, and nursing home costs? [Internet]. Philadelphia, PA: Actionable Intelligence for Social Policy; 2019. Available from:
      • 24
        Hwang S, Tolomiczenko G, Kouyoumdjian F, Garner R. Interventions to improve the health of the homeless: A systematic review. Am J Prev Med. 2005;29:311–311.
      • 25
        Roncarati JS, Baggett TP, O'Connell JJ, Hwang SW, Cook EF, Krieger N, et al. Mortality Among Unsheltered Homeless Adults in Boston, Massachusetts, 2000-2009. JAMA Intern Med. 2018;178:1242–8.
      • 26
        Roncarati JS, O'Connell JJ, Hwang SW, Baggett TP, Cook FE, Krieger N, et al. The Use of High-Risk Criteria to Assess Mortality Risk among Unsheltered Homeless Persons. J Health Care Poor Underserved. 2020;31:441–54.
      • 27
        O'Connell JJ, Roncarati JS, Reilly EC, Kane CA, Morrison SK, Swain SE, et al. Old and sleeping rough: elderly homeless persons on the streets of Boston. Care Manag J. 2005/07/29 ed. 2004;5:101–6.
      • 28
        Nyamathi AM, Leake B, Gelberg L. Sheltered versus nonsheltered homeless women. J Gen Intern Med. 2000;15:565–72.
      • 29
        Richards W, Keauffling J. Homeless who accessed a healthy living centre in Swansea, South Wales: an assessment of the impact of oral ill-health. Prim Dent Care. 2009/07/02 ed. 2009;16:94–8.
      • 30
        Cousineau MR. Health status of and access to health services by residents of urban encampments in Los Angeles. J Health Care Poor Underserved. 1997;8:70–82.
      • 31
        O'Toole TP, Gibbon JL, Hanusa BH, Fine MJ. Utilization of health care services among subgroups of urban homeless and housed poor. J Health Polit Policy Law. 1999/05/26 ed. 1999;24:91–114.
      • 32
        Tsai J, Kasprow WJ, Kane V, Rosenheck RA. Street Outreach and Other Forms of Engagement with Literally Homeless Veterans. J Health Care Poor Underserved. 2014;25:694–704.
      • 33
        DeMarco A, Hardenbrook R, Rose J, Mendoza D. Air pollution-related health impacts on individuals experiencing homelessness: Environmental justice and health vulnerability in Salt Lake County, Utah. Preprints [Internet]. 2020; Available from:
      • 34
        Abbs E, Brown R, Guzman D, Kaplan L, Kushel M. Risk Factors for Falls in Older Adults Experiencing Homelessness: Results from the HOPE HOME Cohort Study. J Gen Intern Med. 2020/01/23 ed. 2020;
      • 35
        Tong MS, Kaplan LM, Guzman D, Ponath C, Kushel MB. Persistent homelessness and violent victimization among older adults in the HOPE HOME Study. J Interpers Violence. 2019;088626051985053.
      • 36
        Feske ML, Teeter LD, Musser JM, Graviss EA. Counting the homeless: a previously incalculable tuberculosis risk and its social determinants. Am J Public Health. 2013/03/16 ed. 2013;103:839–48.
      • 37
        Feleke DG, Wage EK, Getachew T, Gedefie A. Intestinal parasitic infections and associated factors among street dwellers’ in Dessie town, North-East Ethiopia: a cross sectional study. BMC Res Notes. 2019/05/12 ed. 2019;12:262.
      • 38
        Moges F, Kebede Y, Kassu A, Degu G, Tiruneh M, Gedefaw M. Infection with HIV and intestinal parasites among street dwellers in Gondar city, northwest Ethiopia. Jpn J Infect Dis. 2006/12/26 ed. 2006;59:400–3.
      • 39
        Ray SK, Biswas R, Kumar S, Chatterjee T, Misra R, Lahiri SK. Reproductive health needs and care seeking behaviour of pavement dwellers of Calcutta. J Indian Med Assoc. 2001/08/02 ed. 2001;99:142–3, 145.
      • 40
        Wakgari N, Woyo T, Kebede E, Gemeda H, Gebremedhin S. Sexually transmitted disease among street dwellers in southern Ethiopia: a mixed methods study design. BMC Public Health. 2020/04/05 ed. 2020;20:434.
      • 41
        North CS, Pollio DE, Thompson SJ, Spitznagel EL, Smith EM. The association of psychiatric diagnosis with weather conditions in a large urban homeless sample. Soc Psychiatry Psychiatr Epidemiol. 1998/05/30 ed. 1998;33:206–10.
      • 42
        Heckert U, Andrade L, Alves MJ, Martins C. Lifetime prevalence of mental disorders among homeless people in a southeast city in Brazil. Eur Arch Psychiatry Clin Neurosci. 1999/08/05 ed. 1999;249:150–5.
      • 43
        Hynes F, Kilbride K, Fenton J. A survey of mental disorder in the long-term, rough sleeping, homeless population of inner Dublin - Corrigendum. Ir J Psychol Med. 2019/04/02 ed. 2019;36:81.
      • 44
        Bennett AS, Watford JA, Elliott L, Wolfson-Stofko B, Guarino H. Military veterans’ overdose risk behavior: Demographic and biopsychosocial influences. Addict Behav. 2019/09/09 ed. 2019;99:106036.
      • 45
        Baggett TP, Campbell EG, Chang Y, Rigotti NA. Other tobacco product and electronic cigarette use among homeless cigarette smokers. Addict Behav. 2016/04/30 ed. 2016;60:124–30.
      • 46
        Santa Maria DM, Narendorf SC, Cross MB. Prevalence and Correlates of Substance Use in Homeless Youth and Young Adults. J Addict Nurs. 2018/03/06 ed. 2018;29:23–31.
      • 47
        Smereck GA, Hockman EM. Prevalence of HIV infection and HIV risk behaviors associated with living place: on-the-street homeless drug users as a special target population for public health intervention. Am J Drug Alcohol Abuse. 1998/06/27 ed. 1998;24:299–319.
      • 48
        Sutter A, Curtis M, Frost T. Public drug use in eight U.S. cities: Health risks and other factors associated with place of drug use. Int J Drug Policy. 2018/12/24 ed. 2019;64:62–9.
      • 49
        Fountain J, Howes S, Marsden J, Taylor C, Strang J. Drug and Alcohol Use and the Link with Homelessness: Results from a Survey of Homeless People in London. Addict Res Theory. 2003;11:245–56.
      • 50
        Fountain J, Howes S, Strang J. Unmet drug and alcohol service needs of homeless people in London: a complex issue. Subst Use Misuse. 2003/05/16 ed. 2003;38:377–93.
      • 51
        Yoon C, Ju YS, Kim CY. Disparities in health care utilization among urban homeless in South Korea: a cross-sectional study. J Prev Med Public Health. 2011/12/07 ed. 2011;44:267–74.
      • 52
        Elwell-Sutton T, Fok J, Albanese F, Mathie H, Holland R. Factors associated with access to care and healthcare utilization in the homeless population of England. J Public Health Oxf. 2016/02/21 ed. 2017;39:26–33.
      • 53
        Lee CT, Guzman D, Ponath C, Tieu L, Riley E, Kushel M. Residential patterns in older homeless adults: Results of a cluster analysis. Soc Sci Med. 2016;153:131–40.
      • 54
        Macnee CL, Forrest LJ. Factors associated with return visits to a homeless clinic. J Health Care Poor Underserved. 1997/10/23 ed. 1997;8:437–45.
      • 55
        Wenzel SL, Bakhtiar L, Caskey NH, Hardie E, Redford C, Sadler N, et al. Homeless veterans’ utilization of medical, psychiatric, and substance abuse services. Med Care. 1995/11/01 ed. 1995;33:1132–44.
      • 56
        Douglass RL, Torres RE, Surfus P, Krinke B, Dale L. Health care needs and services utilization among sheltered and unsheltered Michigan homeless. J Health Care Poor Underserved. 1999/02/16 ed. 1999;10:5–18.
      • 57
        O'Toole TP, Freyder PJ, Gibbon JL, Hanusa BJ, Seltzer D, Fine MJ. ASAM Patient Placement Criteria treatment levels: do they correspond to care actually received by homeless substance abusing adults? J Addict Dis. 2004/04/14 ed. 2004;23:1–15.
      • 58
        Linton KF, Shafer MS. Factors associated with the health service utilization of unsheltered, chronically homeless adults. Soc Work Public Health. 2013/11/06 ed. 2014;29:73–80.
      • 59
        MacKenzie OW, Trimbur MC, Vanjani R. An Isolation Hotel for People Experiencing Homelessness. N Engl J Med. 2020;383:e41.
      • 60
        Robinson L, Schlesinger P, Keene DE. “You Have a Place to Rest Your Head in Peace”: Use of Hotels for Adults Experiencing Homelessness During the COVID-19 Pandemic. Hous Policy Debate. Routledge; 2022;0:1–16.
      • 61
        Wusinich C, Bond L, Nathanson A, Padgett DK. “If you're gonna help me, help me”: Barriers to housing among unsheltered homeless adults. Eval Program Plann. 2019/06/23 ed. 2019;76:101673.
      • 62
        Otiniano Verissimo AD, Henley N, Gee GC, Davis C, Grella C. Homelessness and discrimination among US adults: the role of intersectionality. J Soc Distress Homelessness. Taylor & Francis; 2021;0:1–15.
      • 63
        Weisz C, Quinn DM. Stigmatized identities, psychological distress, and physical health: Intersections of homelessness and race. Stigma Health. US: Educational Publishing Foundation; 2018;3:229–40.
      • 64
        Lynch KA, Harris T, Jain SH, Hochman M. The Case for Mobile “Street Medicine” for Patients Experiencing Homelessness. J Gen Intern Med [Internet]. 2022 [cited 2022 Sep 12]; Available from:
      • 65
        Eyrich-Garg KM. Mobile phone technology: A new paradigm for the prevention, treatment, and research of the non-sheltered “street” homeless? J Urban Health. 2010;87:365–80.
      • 66
        Rhoades H, Wenzel SL, Rice E, Winetrobe H, Henwood B. No digital divide? Technology use among homeless adults. J Soc Distress Homeless. Taylor & Francis; 2017;26:73–7.
      • 67
        Raven MC, Kaplan LM, Rosenberg M, Tieu L, Guzman D, Kushel M. Mobile Phone, Computer, and Internet Use Among Older Homeless Adults: Results from the HOPE HOME Cohort Study. JMIR MHealth UHealth. 2018;6:e10049.
      • 68
        Galperin H, Bar F, Nguyen H. The power divide: Mobile communication in Los Angeles’ Skid Row. Mob Media Commun. SAGE Publications; 2021;9:30–50.
      • 69
        McInnes DK, Sawh L, Petrakis BA, Rao SR, Shimada SL, Eyrich-Garg KM, et al. The potential for health-related uses of mobile phones and internet with homeless veterans: results from a multisite survey. Telemed E-Health. Mary Ann Liebert, Inc. 140 Huguenot Street, 3rd Floor New Rochelle, NY 10801 USA; 2014;20:801–9.
      • 70
        Giamarino C, Brozen M, Blumenberg E. Planning for and Against Vehicular Homelessness. J Am Plann Assoc. Routledge; 2022;0:1–13.
      • 71
        Goto H, Culhane DP, Marr MD. Why Street Homelessness Has Decreased in Japan: A Comparison of Public Assistance in Japan and the US. Eur J Homelessness Vol. 2022;16.