A systematic scoping review of primary health care service outreach for homeless populations

Abstract

Background

Homeless populations (HPs) have difficulties obtaining necessary medical care, and primary health care service outreach (PHSO) might be useful to bridge this gap.

Objective

Using the Centre for Evidence-Based Management Critically Appraised Topics framework, to provide systematic evidence of the usefulness of PHSO interventions for HPs.

Methods

A systematic search was conducted in 4 electronic databases: PubMed, Web of Science, CINAHL, and Cochrane (publication dates between January 1980 and November 2020). In total, 2,872 articles were identified. Primary research about PHSO for HPs in high-income countries were included. Data were extracted from eligible studies, summarized, and collated into a narrative account.

Results

Twenty-four studies that described and evaluated PHSO interventions for adults experiencing homelessness were selected in the final synthesis. Most studies had a nonrandomized design. PHSO was found to successfully address some barriers to health care access for HPs through flexible appointments in convenient locations, fostering an understanding relationship between doctor and patients, and provision of additional basic necessities and referrals. Outreach was provided for a range of health care concerns, and several solutions to engage more HPs in primary care, improve continuity of care and to decrease the running costs were identified. Outreach also helped to implement preventative measures and reduced emergency service admissions.

Conclusion

Our review adds to the evidence that PHSO likely improves health care access for HPs. Further studies over longer time periods, involving collaborations with experts with lived experience of homelessness, and utilizing randomized study designs are needed to test outreach efficacy.

Background

Life expectancy of homeless populations (HPs) is 30 years shorter than that of the average population in the UK,^1,2 and a quarter of their deaths could be prevented if a physician had been seen in time.^3–5 Rates of homelessness in the UK have been rising over the past 5 years and are predicted to rise still following the aftermath of the COVID-19 pandemic.⁶

Despite a universally free health care system in the UK, HPs are 40-fold less likely to be registered with a mainstream general practice compared with housed people.⁷ This can partially be explained by the existence of multiple barriers for HPs to access primary health care.^8–10 First, registration with a primary health care practice is challenging, as ID documents and proof of address are still requested for, although these are no longer legally mandatory.^8,10–13 Second, the rigidity of appointments (1 problem/1 appointment)^9,10 and the long waiting time to secure one¹⁰ are other barriers encountered by HPs. Third, many patient factors pose difficulties, such as the stigma that HPs receive from doctors and other patients in mainstream primary health care practices, mostly because of their health problems as well as housing and immigration status, thus impeding trust and engagement amongst this population.^9,10 General practitioners (GPs) also lack awareness of the complex health care needs of HPs, and as a result, give nontailored advice.^8–10 These barriers, in addition to life priorities that supersede health care,^8,9 lead HPs to wait for dire health situations before reaching out for help, making them more likely to utilize accident and emergency services (A&E) compared with the general population.^14–17 Following their discharge from hospital, follow-up is difficult as they have no GP for correspondence.^8–10 As a result, they are caught in a vicious cycle: their health worsens until they are forced to seek A&E again. In contrast, it has been reported that HPs recently registered with a primary health care practice are less likely to attend A&E.⁷

Different strategies have aimed to improve access to primary health care for HPs, including tailored primary care organizations (e.g. specialist GPs for HPs), multidisciplinary/integrated care, and service colocation with mental health services and social support services.^18,19 Key aspects of these primary health care models are as follows: a patient-centered, comprehensive approach, liaison with community agencies, and care management, including coordination of care.¹⁸ In the United Kingdom, the 3 main models of primary health care delivery for HP rely on (i) a mainstream primary health care service practice with a “special interest” in treating homeless patients, (ii) a specialized primary health care service for homeless patients, and (iii) specialized primary care for HPs within a hospital.²⁰

One of the suggested strategies to improve access to primary health care services for HPs is outreach.^18,21–23 Outreach involves health care professionals coming to see patients in their own environment, while respecting their needs and experiences,^24,25 as opposed to the situation where patients themselves approach GPs in a standard or specialist homeless GP setting. This approach has been suggested to overcome barriers created by mainstream services.^20,23

A UK-based study exploring opinions of HPs about outreach identified advantages, including the ability to offer human connection after loneliness on the street, capacity to overcome barriers to health care access, and ability to provide a bridge to health care services.²⁶ Conclusions from this study were however limited by the fact that only 20% of participants had experienced an outreach service. Other factors key to outreach success were identified, like location, timing, team composition, and communication styles used by outreach workers.²⁶

The aim of this review was to assess the primary health care service outreach (PHSO) strategies used for HPs and determine whether these interventions were helpful in meeting their needs. Because outreach services for HPs can be particularly demanding, at both the human resources and economic levels,²⁷ it is necessary to map available evidence in the area to identify gaps in research and inform future policy-making.

Methods

Databases and search strategy

This review was systematically conducted in accordance with the Centre for Evidence-Based Management Critically Appraised Topics framework,^28,29 which included a step-by-step guide to providing a snapshot of the best evidence by study design. To review the available reports devoted to PHSO for HPs in high-income countries, 4 databases were searched: PubMed, Web of Science, CINAHL, and Cochrane, by using a combination of keywords. The following generic search filters were applied to all databases during the search:

• Abstract/title

• Published in 1980–2020 (including 2021 in PubMed). The search was conducted on 2020 Nov 19

• Articles in English (apart from in COCHRANE, where language setting was not possible)

In the Cochrane database, the search modes Boolean/Phrase were also used. The search strategy is presented in Figure 1. The definitions and full selection criteria are presented in Table 1.

We should note that information on whether housing was inadequate and/or insecure was often omitted in original articles. Where the numbers of HPs were not stated, we included articles where the homelessness rate >50% could be inferred from the PHSO setting or from the PHSO program. We focused on high-income countries, which share a similar definition of “homelessness” and have similar means to address this issue. They also have similar increasing rates of homelessness.^32,33

Because there were relatively few studies identified in our initial search that employed comparisons (e.g. before/after PHSO implementation or outcomes in 2 locations where PHSO was implemented in one location but not the other), we included articles despite them not being comparative studies.

Studies were included that evaluated the PHSO in some way and categorized outcomes, as was done in a similar study.³⁴ In brief, our primary outcomes were measures of unscheduled health care usage, such as A&E attendance, or inpatient hospital admission; secondary outcomes were mortality or markers of disease control; tertiary outcomes included patient opinions about PHSO and cost-effectiveness. Studies were included if they reported any of the primary, secondary, or tertiary outcomes. Objective and subjective outcomes, as reported by the patients and not the health care providers, were included.

Screening process

Each title and abstract were independently screened by 3 reviewers. To minimize risk of bias, 2 reviewers screened every paper (V.K., S.M.) and the third screen was split up between 5 reviewers (B.S., R.L., S.S., S.P., E.S.). Conflicts were discussed by the whole team. Next, full-text articles were read by at least 2 reviewers to determine eligibility. The articles were selected if they matched inclusion criteria, and data extraction was conducted. D.M.R. supervised the investigation and provided feedback while ensuring rigor using validated methods and standard definitions.

Results

Study selection

The literature search identified 4,156 unique entries, which represented 2,872 unique articles after duplicates were removed. After the selection process, 24 articles matched the inclusion criteria and were included in this review. This is demonstrated in the PRISMA Flow diagram³⁵ in Figure 2.

Study characteristics

All the selected articles, published between 2006 and 2020, described and evaluated PHSO for HPs. Sample sizes ranged from 27 to 353 participants. Study characteristics are presented in Supplementary Table 1. The study designs were recorded as stated in the papers, while the design tree on Centre for Evidence based Medicine³⁶ was utilized to resolve any that were unstated.

We focused on 7 key topics: (i) reliance on PHSO; (ii) positive factors associated with PHSO; (iii) preventative strategies, diagnosis, and treatment; (iv) retention and engagement; (v) referrals; (vi) effect on emergency and inpatient hospital admissions; (vii) cost-effectiveness.

Reliance on PHSO

Ten of the 24 studies (41.7%) commented on the extent of HPs’ reliance on PHSO. In 5 studies,^37–41 PHSO was the main source of health care for over half of the patients (range: 50–71.7% of the total number of patients) and in 3 further studies,^42–44 smaller but still significant proportions (range: 21.4–49% of the total number of patients) relied on PHSO for their care. In one of these studies, out of the 49% of patients who considered PHSO as their main source of health care, 63.1% received care solely from this clinic.⁴² Two studies commented on other sources of health care used, which included a doctor’s office, a community health center, their GP, and a Safety Net fixed site clinic.^41,43 Patients who were homeless were less likely to report these as their usual source of health care (24.8%) compared with those who were currently housed (51.7%).⁴³

Reliance was further demonstrated during interviews with HPs using the outreach services: One patient stated that they would be dead if it was not for this service,⁴⁵ and in another study, 86% patients did not know where else to seek dental care if needed.³⁹

Additionally, 2 studies evaluating mobile PHSO clinics mentioned that these clinics helped patients transition to a mainstream GP practice.^45,46 However, one of these noted that patients who had been using the service for longer found it more difficult to move on to a mainstream GP practice,⁴⁵ suggesting that despite its positive effects on HPs access to primary care, PHSO has limitations that must be considered.

Positive factors associated with PHSO

Eight of the 24 studies (33.3%) reported positive factors associated with PHSO. These are displayed in Table 2. The most commonly described factors were a convenient location and flexible appointments,^41,43,45,47 and a good doctor–patient relationship with friendly staff.^{39,42,45–47} Colocation with other services and provision of additional basic necessities were also valued.^43,45,48 These studies suggested that health was not always the immediate priority for HPs who attended PHSO. Such services can nonetheless successfully provide additional basic necessities and overcome barriers to health care.

Preventive strategies, diagnosis, and treatment

Fourteen studies (58.3%) reported PHSO interventions that diagnosed and/or treated ill health, and 4 studies (16.7%) commented on the effectiveness of preventative strategies. Details of the studies are shown in Table 3. PHSO was found to be equally useful in covering dental health,^{38–40,44,47,49} mental health,^37,48,50,51 general medical problems,^{37,48,51–53} and infectious diseases.^48,52–55 There were statistically significant improvements in health, particularly mental health, when formerly homeless volunteers were involved in treatment.⁴⁸ The few evaluations of preventive services, including a vaccination program,⁵⁶ a dental health education program,⁴⁶ and dental advice,^39,47 suggest it is possible to achieve high rates of preventive measures and address health disparities.

Retention and engagement

Eleven of the 24 studies (45.8%) analyzed the engagement and retention of homeless people in PHSO. Details of these studies are shown in Table 4. Studies reported higher rates of outreach visits compared with that of conventional care,⁵⁷ high numbers of outreach visits in general,^37,48,58 and >50% of appointment attendance^{38,44,47,52,59} and treatment completion.^39,44 However, they also raised issues of high rates (>50%) of missed appointments^39,53 and uncompleted treatment,³⁸ as well as the suboptimal targeting of certain HPs, including young people, people with low benefits, people with more severe disease, people with mental illness, and rough sleepers.^44,48,53

Studies suggested that appointments arranged face-to-face were more likely to be attended,⁵⁹ that morning appointment attendance was better,⁴⁷ and that there was a preference for drop-in appointments,^38,39 suggesting retention could be improved by considering HPs’ preferences. In one study, PHSO did not deliver care in the outreach settings but invited patients to a fixed-site clinic the following day.³⁸ Just over half of those approached attended the clinic, suggesting that PHSO may be more effective if treatment and/or screening can also be offered in the same place. In one study,⁵³ an incentive of 20 Australian dollars was offered to the patients for taking part in the study; however, appointment attendance in this study was still low, suggesting that on its own, this method was not helpful in attracting patients. Together, these studies suggest PHSO, when tailored to HPs’ needs, is effective for engaging HPs.

Referrals

Seven out of the 24 studies (29.2%) commented on referrals made by the PHSO clinics (Table 5). Most referrals were made to specialist services and other primary health care services^{37,38,52–54,60}; however, some were also directed to emergency services,³⁷ soup kitchens, community housing and financial support programs.⁵² The effectiveness of referrals was difficult to establish, as the majority of studies did not state how many people visited other health care services after the outreach, and how many of these were due to referrals. In contrast, Roche et al. explicitly suggested that PHSO was not very helpful for referrals as under 50% of patients visiting another health care facility were referred there by the outreach clinic.³⁷ Overall, although PHSO clinics make referrals to different services, more evidence is needed to assess the effectiveness of these referrals.

Effect on emergency and inpatient hospital admissions

Four studies commented on the effect of PHSO on A&E. Interviews with HPs suggested that outreach does help avoid A&E admission, although the one comparative study did not support this finding. A longitudinal study found no significant differences in the use of emergency services by patients accessing any of the following 3 models of primary care: (i) personal health assessment + brief intervention; (ii) clinic/health system orientation; and (iii) a combination of the first two.⁵⁸ A study evaluating a mobile outreach clinic found that patients attended the A&E when this outreach was not available, but when this outreach was active, several referral pathways existed between community services, including mental health, which helped to avoid visits to the emergency department.⁴⁵ Sixteen percent of patients attending a mobile outreach clinic in Dublin similarly stated that if the clinic did not exist, they would have attended A&E.⁴¹ In addition, 82.5% of patients attending a hostel-collocated clinic also said that this outreach service helped them avoid visiting the A&E.³⁷

Cost-effectiveness

Three studies commented on the financial costs of PHSO. A study comparing different models of dental services found that a nonoutreach model involving referrals was more cost effective than outreach models where appointments were made right after screening in an outreach setting.⁵⁹ One study described a mobile dental service that could be run at low cost because of the volunteer dental professionals.³⁹ Although this study found that the estimated value of services provided in the first year were more than the cost to set up the service, the larger part of the costs was associated with the dental equipment that could be used in future clinics, whereas non-reusable equipment did not contribute much to the expenses. A hepatitis C virus (HCV) testing outreach intervention was likely to cost-effectively improve treatment uptake. Moreover, when compared with the mainstream care, the outreach service lowered HCV-related care over the longer time period.⁵⁵ These cost evaluations of PHSO support the view that outreach is more costly than nonoutreach services, but that there are solutions to mitigate such costs: employing volunteers and health care students, investing in equipment, and running the service over longer periods of time.

Discussion

In our review, we sought to examine how effective PHSO approaches are in meeting the specific needs of HPs. We used standard systematic methods to identify, select, and synthesize findings from 24 studies that described and evaluated PHSO for HPs. To the best of our knowledge, there have been no previous systematic or scoping reviews on this specific topic.

Our review clearly showed that PHSO solves a range of health care concerns, partly explaining why many HPs rely on these services for health care. Testing for infectious diseases and preventative services such as vaccinations appear to gain significantly from being brought out. This is important as previous work has shown that vaccine uptake in HPs is more than 2-fold lower than in the general population.⁶¹

Despite these positive findings, we nonetheless identified important limitations of PHSO in addressing HP health care needs. Health was not the immediate priority for HPs, with basic needs (food and shelter) ranking higher, as was reported previously.⁶² We also found that certain categories of HPs, young people, rough sleepers, and people with severe illness, mental illness, or fewer benefits were more likely to be neglected, even in outreach care. This contrasts with conclusions of a study, which looked at a specialized mental health outreach program⁶³ rather than our definition of primary care, and suggested that it was the older population that was often left unreached. These are important limitations to bear in mind as retention and engagement of HPs in care are necessary for effective treatment and management of diseases, especially chronic or long-term ones.⁶⁴ In the studies reviewed by us, it was rarely stated whether further appointments were necessary following the initial one. This made the interpretation of increased visits to the clinic difficult, as these could be due to patients’ health worsening and requiring increased follow-up, rather than patient’s engagement. Tommasello et al., however, did suggest that increased visits were linked to improved health outcomes.⁴⁸ Previous studies have also reported that following a diagnosis of a health care condition, patients may pay greater attention to their health and subsequently visit the GP more regularly.^65,66 Similarly, it was difficult to interpret how effective referrals from outreach services were, as most studies did not state how many patients actually attended their referred appointments. Increased referrals may be interpreted negatively, in that the outreach service did not meet the needs of HPs.

Recommendations regarding several aspects of PHSO can be made based on our review. In particular, acting on feedback from HPs about how outreach could fit around their day and lifestyle may help retain HPs in care. Our findings suggest that morning, drop-in and face-to-face appointments are particularly preferred. Use of monetary incentives did not appear to improve retention in care substantially. The latter finding was in contrast to the conclusions of a study that examined incentives for TB prophylaxis in HPs and found that a $5 monetary incentive was more effective than a peer health adviser for completion of treatment (this study was not included in our review).⁶⁷ Ideally, PHSO should provide care in the outreach setting rather than inviting patients to a fixed site clinic on a later occasion.

A cross-sectoral approach to PHSO could be an excellent opportunity to engage HPs and additionally provide basic necessities, or at least directions to places that could offer food, shelter, or help with housing. Establishing better links with other community bodies, such as shelters and refuges, could also be particularly helpful for engaging the most vulnerable HPs.

We found that outreach services made many referrals, however reliance on referrals to other services is risky, as care of HPs becomes too fragmented before they can finally receive treatment. Lack of continuity of care also means that patients have to repeat their medical and social history, which is overwhelming for them given their complicated life circumstances.^10,45 A solution to this could be to include the necessary specialists in the outreach team. For example, outreach screening for specific infections could have infectious disease specialists in the team.

The dependence on primary health care physicians for certain tasks, for example to prescribe treatment, has been mentioned in several studies. This may not be optimal, particularly if the outreach is not physician-led or the physician only comes on certain days of the week, as this creates a missed opportunity to treat the hard-to-reach group. Thus, task shifting, where procedures could be conducted by health care practitioners other than doctors, may be important for outreach services to ensure higher flexibility in administering treatments.⁶⁸

To avoid simply creating a parallel service alongside mainstream GPs,⁶⁹ outreach should be provided by combined specialized and mainstream primary care services or employ dedicated peer workers who actively help patients transition to a mainstream GP practice when they are ready.

Our findings suggested that although PHSO may be more costly than mainstream services in terms of equipment, it may be more cost-saving in the longer term. Suggested ways of reducing cost could be through employing volunteers (including students), investing in equipment and running the outreach service over a longer time period.

Overall, in order to adequately assess the differential impact of PHSO on the health and well-being of HPs, more comparative studies, encompassing longer time periods, are needed. The effect of outreach on the use of emergency services would be a particularly useful parameter to evaluate in future studies, as HPs often use emergency departments, partly due to the inaccessibility of primary care services.^14–17 The decreased frequency of visits to emergency services would be a strong marker of outreach services efficiency, making it a useful evidence-based metric. Further qualitative studies coproduced with HPs would be helpful in providing insight into how these services can better meet their needs. We believe that many of our recommendations may be easily implemented, although their feasibility and acceptability should be best assessed through interviews with staff and HPs.

Limitations

Although we used a comprehensive search strategy that identified a wide range of articles and employed very specific inclusion criteria, our review has some limitations. Notably, no critical appraisal has been performed to identify the quality of the selected studies. Another limitation is that most of the reviewed studies did not have control or comparison groups to compare their intervention groups against. Moreover, most measures were self-reported, which resulted in response bias. However, Hwang et al. found that HPs’ recall of health outcomes and their views about facilitators to care were reliable.⁷⁰ Confounding factors, such as shelter or improved access to social workers, were likely present in outreach co-located with other services, which could have contributed to positive aspects associated with PHSO. Some of the reviewed studies did not describe their criteria for homelessness, failed to stratify their population adequately, for example in terms of gender, race, and age, or gave no definition of homelessness at all. Additionally, we should note that despite our search was limited to high-income countries, their health policies and support systems for HPs varied widely.

Conclusion

This review identified several key factors through which PHSO addressed health inequalities and inequities for HPs. Many HPs rely on these services and attend them for a variety of health issues. We propose several solutions to engage more HPs in care, improve continuity of care, and decrease the running cost of the service. Although these services demonstrated some success in preventative strategies and reducing emergency service admissions, comparative studies over a longer time period will be necessary to confirm and contrast the outcomes and efficiency of primary outreach for HPs with mainstream primary health care.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Acknowledgments

We are grateful to Professor Andrew Hayward for the discussions leading to this review.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Data availability

Data from all studies analyzed in this scoping review can be accessed via publicly available databases PubMed, Web of Science, CINAHL, and Cochrane.

References