Homelessness & Unstable Housing

Skip anything

Every question is optional. No one has to retell painful events, name a diagnosis, or choose a religion or identity to begin.

Pause or stop

Stop at any point. Nothing is sent anywhere, and nothing is shared unless the person chooses it in the sharing review.

Whose words

Record who said each thing: the person, a caregiver or surrogate, or a professional. A caregiver's needs get their own field.

Support to take part

Extra time, a quieter place, pictures, an interpreter or a trusted person can help someone say what matters. Ask what helps.

Barriers are not preferences

Cost, travel, work, housing, caring for others and digital access are written down as barriers, not as what the person wants.

Meaning, religious or not

Hope, purpose, faith, relationships and identity can be part of the conversation if the person wants. None is assumed.

• Doctor or nurse: Care and medicines that fit the person's circumstances.
• Outreach worker: Finding and staying in touch with the person where they are.
• Social worker or housing worker: Housing, benefits, ID and plans for leaving hospital.
• Peer support worker: Lived experience and trust, where such roles exist.
• Chaplain or spiritual-care provider: Meaning, distress, hope and relationships, for people of any faith or none.
• Interpreter or communication specialist: Language and communication support so the person's own words come through.
• People the person chooses: Family, friends, or a community or faith contact, only if the person wants them. Never assumed safe, available or wanted.

Roles, titles and training differ by country and service. Not every role exists everywhere.

A small set of places chosen for contrast, not a ranking or a representative survey: Multi-country, Malaysia, United States, Canada, United Kingdom, New Zealand, Brazil, South Africa, Bangladesh, Iran. Each example states what its source establishes and what it does not. Policy on paper is not proof that a service is available.

• Multi-country · Guidance · 2005

  A 2005 WHO Regional Office for Europe evidence synthesis said primary care for people experiencing homelessness requires suitable organization, policies and procedures (including for threatening behavior), convenient appointment times and multidisciplinary team capability.

  What others could learn: Access features such as convenient appointment times can be planned rather than left to chance. The same report lists discrimination, hours of business, appointment procedures and financial disincentives for practitioners among barriers, and says stigma among staff can be addressed by training.

  Conditions, limits and source

  Conditions that matter:

  The report describes two primary care models: mainstream general practices with an extended role, and specialized practices usually found only in large cities. It says social worker support appears crucial in hospital-based models and names lack of health insurance as a barrier to address.

  Cannot be inferred:

  Does not show which model works better; the report says it is uncertain how many people use each type of practice. Does not describe current services in any country. Does not cover low-income settings, which the authors put outside its scope.

  Limits:

  Published in January 2005 from literature searched in November 2003; services, insurance rules and evidence have changed since. Covers high-income countries only and excluded asylum seekers. A commissioned report whose contents are the authors' responsibility, not official WHO/Europe policy. Uses older, non-person-first wording. Does not mention palliative care.

  Source: World Health Organization Regional Office for Europe, Health Evidence Network (Wright NMJ, Tompkins CNE), How can health care systems effectively deal with the major health care needs of homeless people? (Section on primary health care, printed p. 7 (PDF p. 7); section on barriers to health care, p. 6; staff training under 'Policy considerations' in the Summary, p. 4). Where: Multi-country (WHO European Region; evidence mainly from high-income countries). Population: People experiencing homelessness in high-income countries, including people sleeping rough and people in night shelters, hostels or insecure accommodation. Published 2005-01-01; retrieved 2026-10-12.

• Malaysia · Service in operation · 2025

  In Kuala Lumpur, a group of family medicine specialists developed a street-based acute medical care outreach program to support a volunteer group's weekly food and hygiene distribution for people experiencing homelessness.

  What others could learn: Bringing care to where people already gather, such as a regular food distribution, and then walking the streets can reach people who would not attend a clinic. The team also coordinated an ambulance so a critically ill person reached hospital.

  Conditions, limits and source

  Conditions that matter:

  Depends on volunteers: before the program, volunteer doctors came only sporadically, and the specialists bought medicines and equipment with their own funds. The authors say sustaining care needs volunteers, government support, NGO collaboration and corporate social responsibility funding. They report language barriers.

  Cannot be inferred:

  Does not show health effects, continuity of care or how many people were seen. Does not describe follow-up arrangements or how people store medicines. Not representative of other services in Malaysia.

  Limits:

  A brief first-person account by the clinicians who run the program, with no patient numbers, outcomes or views of the people served. The authors' statement that street-based care builds trust is their view, not a measured finding. Single city. PubMed holds no abstract; the PMC full text was used.

  Source: Malaysian Family Physician (Malek KA, Ismail IA, Ariffin F and colleagues; Universiti Teknologi MARA and Food for Gelandangan), Acute medical care for individuals experiencing homelessness in Kuala Lumpur. (Section 'Unmet health needs' (PMC full text, PMC12258502); conditions from 'The experience', 'Encountered medical issues' and 'The future'). Where: Malaysia (Kuala Lumpur city center). Population: People sleeping rough in central Kuala Lumpur. Published 2025-06-10; retrieved 2026-10-12.

• United States · Measured outcome · 2018

  Among 350 adults aged 50 or older with experience of homelessness in Oakland, California, enrolled 2013 to 2014, 61% reported a potential surrogate decision-maker, but 21.5% had discussed advance care planning and 19.0% reported documentation.

  What others could learn: Most participants named someone they trusted to decide for them, yet few had talked about or written down their wishes. Having 1 to 5 confidants was associated with higher odds of both, so asking whom the person trusts can open the conversation.

  Conditions, limits and source

  Conditions that matter:

  The authors say interventions must be customized for people with limited social networks and address the instability of homelessness, health literacy and the constraints of safety-net health care settings.

  Cannot be inferred:

  Associations are cross-sectional, not causal. Does not test an advance care planning intervention. Does not show whether wishes were honored at the end of life.

  Limits:

  One city, with enrollment from July 2013 to June 2014, so current figures may differ. Documentation was self-reported, not checked in charts. English speakers only. The authors say Oakland has a robust safety-net system and results may not generalize.

  Source: Journal of the American Geriatrics Society (Sudore RL, Kushel M and colleagues; University of California, San Francisco), Advance Care Planning for Older Homeless-Experienced Adults: Results from the Health Outcomes of People Experiencing Homelessness in Older Middle Age Study. (Abstract; Results in the PMC author manuscript (PMC6105505), which reports 4.0% had discussed ACP wishes with a health care provider). Where: United States (Oakland, California). Population: English-speaking adults aged 50 or older who were homeless or homeless-experienced; median age 59; 75.2% men; 82.1% Black (n=350). Published 2018-05-09; retrieved 2026-10-12.

• Canada · Reported experience · 2026

  In interviews with 14 patients of a Toronto palliative care outreach team for people experiencing homelessness, participants described the team's compassionate, person-centered relational care as central to their trust and engagement.

  What others could learn: Care that comes to a shelter room, and a direct line to a navigator, can remove practical barriers. Participants valued help with appointments, transport and coordination; a few wanted one consistent doctor rather than several.

  Conditions, limits and source

  Conditions that matter:

  An interprofessional team of palliative care physicians, psychiatrists, social workers and nurses working with health and social service partners. It connects patients with publicly funded home care, helps secure housing and covers health care and medication gaps.

  Cannot be inferred:

  Does not measure symptoms, survival or hospital use; the authors call for such studies. Does not represent people who never reached the program.

  Limits:

  Small sample (14) from one program in one city, recruited purposively and then by convenience; only 2 women. Feedback was overwhelmingly positive, and the authors note participants may have worried criticism would affect their care. Several authors are affiliated with the program, though interviewers were researchers not involved in patient care.

  Source: JAMA Network Open (Levesque AR, Morey T and colleagues; University of Toronto; Inner City Health Associates), Patient Perspectives on Palliative Care Outreach to Adults Living With Homelessness. (Abstract, Results; full text (PMC12809365), Results themes 'outreach and community care', 'relational care' and 'constructive feedback'). Where: Canada (Toronto, Ontario). Population: Adult patients living with homelessness and life-limiting illness, aged 20 to 70; 12 men and 2 women; 7 with malignant and 7 with nonmalignant conditions (n=14). Published 2026-01-02; retrieved 2026-10-12.

• United Kingdom · Measured outcome · 2023

  In an Edinburgh evaluation of a hospital in-reach program with structured discharge support for people experiencing homelessness, readmissions among those receiving any in-reach intervention fell 68.7% at 12 months versus the prior 12 months.

  What others could learn: Before the program, staff described people being discharged onto the streets, especially on Friday afternoons when community services closed. Starting discharge planning at admission, with housing workers in the hospital and step-down beds, can prevent that.

  Conditions, limits and source

  Conditions that matter:

  Funded by a homelessness charity. Funding paid for 1.6 to 2.6 FTE community link workers and a service manager; nurses and GPs from a specialist homelessness practice took part. Link workers and care navigators had lived experience of homelessness. The team lacked NHS record access for over a year.

  Cannot be inferred:

  Does not establish causation or compare with hospitals without the program. Does not report patients' own views or health outcomes beyond readmission.

  Limits:

  Uncontrolled pre-post design using aggregate data supplied by program staff, with no individual-level demographics. The authors say one-off admissions may inflate the reduction and COVID-19 housing measures may have affected results. Only 88 people were followed up, and no patients were interviewed.

  Source: BMC Public Health (Malden S, Doi L, Ng L, Cuthill F; University of Edinburgh), Reducing hospital readmissions amongst people experiencing homelessness: a mixed-methods evaluation of a multi-disciplinary hospital in-reach programme. (Abstract, Results; staff accounts of earlier discharges and the program components in full text (PMC10258765), Methods and Results). Where: United Kingdom (Scotland: two NHS hospitals in Edinburgh). Population: People experiencing homelessness admitted to hospital and referred to the program (768 referrals including readmissions; 88 individuals followed up); 15 staff interviewed. Published 2023-06-12; retrieved 2026-10-12.

• New Zealand · Reported experience · 2024

  Staff of a housing and support provider in urban Aotearoa New Zealand described housing key workers going 'above and beyond' to help people experiencing homelessness engage with mainstream health services in clinics, hospitals and emergency departments.

  What others could learn: Trusted non-clinical workers hold knowledge about the person that clinicians can ask for. Participants said many people do not keep phones or take calls from unknown numbers, so appointment systems that rely on phone contact can shut them out.

  Conditions, limits and source

  Conditions that matter:

  Many people supported are Māori or Pacific; participants described distrust of health services and the ongoing trauma of colonization. Key workers reported being questioned or asked to leave by hospital staff. Hospital social work support was not consistently resourced.

  Cannot be inferred:

  Does not report the views of people experiencing homelessness themselves, or measure access or end-of-life outcomes. Not representative of New Zealand.

  Limits:

  Staff views only; the people supported were interviewed for a separate paper. One provider in one urban area. Small sample (20).

  Source: BMC Palliative Care (Gott M, Williams L, Wiles J and colleagues; Te Ārai Palliative Care and End of Life Research Group, University of Auckland), The central role of housing key workers in supporting healthcare interactions for people experiencing homelessness and implications for palliative care: a qualitative study. (Abstract, Results; accounts of phones, distrust and hospital interactions in full text (PMC11613513), Results). Where: New Zealand (one urban area). Population: Staff of one support and housing provider, mostly non-health care staff: 12 key workers plus nurses, doctors, social workers and managers (n=20). Published 2024-12-02; retrieved 2026-10-12.

• Brazil · Reported experience · 2026

  In interviews with 12 people experiencing homelessness in Porto Alegre, Brazil, in late 2024, the Street Clinic (Consultório na Rua) was the service with the strongest bond, and hospital emergency services were sought in acute situations.

  What others could learn: A primary care team that goes to people on the street can become the service they feel most connected to. Participants also named nutrition, spirituality and self-care management among their health needs, with stigma and drug use as barriers.

  Conditions, limits and source

  Conditions that matter:

  The abstract does not describe Street Clinic staffing, funding or hours. The authors report that participants did not perceive drug use as a health issue and used it to cope with pain and hopelessness.

  Cannot be inferred:

  Does not measure access or health outcomes, or compare the Street Clinic with other services. Not representative of Porto Alegre or Brazil.

  Limits:

  Abstract only; the full text, publisher site and doi.org were unreachable. Small sample (12) in one city. The abstract does not say how participants were recruited or whether they were already in contact with the Street Clinic.

  Source: Revista Gaúcha de Enfermagem (Azevedo MB, Rosso LH and colleagues; Porto Alegre), At the limits of pain: perceptions of health care for people experiencing homelessness. (Abstract, Results (English abstract on PubMed; Rev Gaucha Enferm 2026;47:e20250164)). Where: Brazil (Porto Alegre, Rio Grande do Sul). Population: People experiencing homelessness (n=12). Published 2026-07-31; retrieved 2026-10-12.

• South Africa · Reported experience · 2026

  In a Tshwane, South Africa, qualitative study of student occupational therapy groups in shelters and housing settings, the authors found a supportive environment, client-centered care and understanding participants' real-life experiences positively influenced outcomes.

  What others could learn: Residents valued students who were not 'too pushy, demanding or questioning' and let them express themselves; participants appreciated their non-judgmental attitudes. Group topics came from a needs assessment of what members said they needed.

  Conditions, limits and source

  Conditions that matter:

  Delivered by final-year occupational therapy students on supervised placements. Reported challenges included language barriers, erratic attendance, limited resources and lack of continuity of the intervention.

  Cannot be inferred:

  Does not measure health or housing outcomes. Does not address medical, palliative or discharge care. Not representative of Tshwane shelters.

  Limits:

  Authors are from the university whose students delivered the groups. Only people who attended gave views, and some who started did not finish. Interviews about the early lockdown were held months later. Focus groups were held in English. Small samples.

  Source: Health SA Gesondheid (Ainsworth S, Lister HE, Janse van Rensburg MNS; University of Pretoria), Exploring occupational therapy interventions in homeless settings in Tshwane. (Abstract, Results; participant accounts in full text (PMC13058584), theme 2.2 'Through the lens of occupational therapy' and Table 2). Where: South Africa (Tshwane, Gauteng). Population: People experiencing homelessness (12 in three focus groups; 2 interviewed individually) and 14 shelter and housing practitioners; plus occupational therapy student documents. Published 2026-03-05; retrieved 2026-10-12.

• Bangladesh · Measured outcome · 2025

  In a 2022 to 2023 survey of 452 adults experiencing homelessness in Dhaka, Bangladesh, 67% of those with recent acute illness sought treatment, mainly from local drug stores; financial constraint was the main reported barrier.

  What others could learn: Ask where the person actually gets medicines: participants preferred nearby drug stores because they charge no consultation fee and stay open. A policymaker said ID cards are easily lost by people with no fixed place to keep them.

  Conditions, limits and source

  Conditions that matter:

  About 94% of participants had no health card or health scheme membership. Participants described discrimination at hospitals and not knowing where to go for care. The authors recommend outreach, mobile clinics and expanded health protection schemes.

  Cannot be inferred:

  Associations are not causal. Does not measure the quality of drug-store care or health outcomes. Not representative of Bangladesh outside Dhaka.

  Limits:

  Self-reported, cross-sectional data. The authors say they cannot ensure representativeness of all people experiencing homelessness in Dhaka, and people under 18 were excluded. The qualitative sample is small (6 people, 4 policymakers).

  Source: BMC Public Health (Rabbani MG, Alamgir HM and colleagues; icddr,b), Barriers to access and factors associated with health care utilization among the adult homeless population in Dhaka, Bangladesh: a mixed method cross-sectional study. (Abstract, Results; drug-store, discrimination and ID card accounts in full text (PMC12777165), Results). Where: Bangladesh (Dhaka). Population: Randomly selected adults experiencing homelessness (452 surveyed; 6 interviewed in depth) and 4 national policymakers. Published 2025-12-02; retrieved 2026-10-12.

• Iran · Reported experience · 2026

  Among 27 adults aged 60 or older living in shelters in Tehran, Iran, the authors identified five domains of perceived consequences of homelessness, including unmet chronic-care needs and structural exclusion due to loss of identity documents.

  What others could learn: Lost identity documents can block access to entitlements, so a care conversation can ask whether the person has ID and wants help replacing it. The authors say homelessness in later life undermines dignity and limits access to social entitlements.

  Conditions, limits and source

  Conditions that matter:

  The abstract does not describe health services in the shelters or how entitlements depend on documents.

  Cannot be inferred:

  Does not show how many participants lacked documents or measure effects on health care use. Not representative of Iran.

  Limits:

  Abstract only; the publisher site and doi.org were unreachable. Small sample (27) in shelters in one city; people living outside shelters are not included. The abstract gives no dates or recruitment methods and does not say which documents were lost.

  Source: Journal of Aging & Social Policy (Rashedi V, Parvaei S, Mollaei M; Tehran), Aging without Shelter: Lived Consequences of Homelessness Among Older Adults in Tehran, Iran. (Abstract (published online ahead of print; J Aging Soc Policy 2026:1-24)). Where: Iran (Tehran). Population: Older adults aged 60 or older residing in shelters (n=27). Published 2026-06-02; retrieved 2026-10-12.

• New Zealand · Reported experience · 2026

  In 2023 interviews with 16 people with lived experience of homelessness in a major city in Aotearoa New Zealand, participants' end-of-life aspirations rarely concerned medical procedures; they prioritized shelter and basic resources, including a safe place to rest.

  What others could learn: One participant said they would rather die in their supported apartment, where they felt safe. Some saw rest homes, hospital or hospice as options; some women questioned whether formal care spaces were appropriate, acceptable or accessible, and others described poor hospice experiences.

  Conditions, limits and source

  Conditions that matter:

  Participants wanted a trusted, non-judgmental supporter and careful handling of personal information, but could not name who would support them at the end of life. A minority who raised palliative care wanted assurance it would not be financially prohibitive.

  Cannot be inferred:

  Does not show where people experiencing homelessness actually die or whether supported housing makes dying in place possible. Does not represent rural areas or people not in contact with services.

  Limits:

  Small convenience sample from one partner service in one city; one interview each, in English only. All participants found end-of-life aspirations hard to discuss and a minority could not express any, so stated preferences come from part of the sample.

  Source: Palliative Care and Social Practice (Gott M, Black S, Wiles J, Williams L, Moeke-Maxwell T, Robinson J), 'I dream of the airplane, but I cannot board it': Exploring end-of-life aspirations with people who have lived experiences of homelessness. (Results, end-of-life aspirations on shelter and basic resources and on formal care spaces; Methods; Strengths and limitations (PMC13382104 full text)). Where: New Zealand (one large urban area, not named). Population: 16 adults aged 20 to 70 with past or present experience of homelessness (11 men, 5 women; 7 Māori, 4 Pacific, 5 New Zealand European); 13 disclosed a serious mental or physical health condition. Published 2026-07-19; retrieved 2026-10-12.

• The sourced examples include guidance, service in operation, measured outcome, reported experience. Each quote was matched to the source text; that is not independent expert review.
• The prompts and fictional case are original. They have not been tested with patients, caregivers or clinicians.
• Whether this workflow helps anyone is an open research question.
• Content is in English. Interface translations are not reviewed for this topic.