WHAT IS ALREADY KNOWN ON THIS TOPIC

People experiencing homelessness suffer a higher burden of physical and mental health conditions than housed populations.

Limited studies suggest that regular physical activity may address many health conditions prevalent among people experiencing homelessness, although the evidence has not been systematically reviewed.

WHAT THIS STUDY ADDS

A variety of physical activity interventions have been designed and provided to engage people experiencing homelessness, including soccer, outdoor and indoor group activities, and individual activities.

The synthesis of qualitative and quantitative evidence suggests that physical activity can benefit the mental and physical health of people experiencing homelessness with positive translation of benefits to wider life.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

Group physical activity interventions seemed to be the most benefitical to people experiencing homelessness, perhaps due to its facilitation of social support and connection.

Qualitative data highlighted the pressure some participants felt in competetive tournament settings. Organisers should recognise this and consider support to ameliorate impacts of pressure experienced.

Consideration should be given to the intensity level of physical activity interventions for this population. Given the high prevalence and poor health of many people experiencing homelessness, lower threshhold activities are likely to be more inclusive for the population.

Introduction

Homelessness is an extreme form of social exclusion1 2 related to poverty in high-income countries.3 People experiencing homelessness (PEH) are defined as those who are ‘roofless’ (eg, no fixed abode) and ‘houseless’ (eg, living in hostel, shelter, temporary accommodation) in accordance with the European Federation of National Organisations Working with the Homeless.4 Prior to the COVID-19 pandemic, homelessness in the UK had increased annually since 20105 with estimates of all categories of homelessness in England standing at 280 000 people,6 of which 4266 were estimated to be sleeping on the streets.7 The Organisation for Economic Co-operation and Development (OECD) estimates that almost 2 million people are experiencing homelessness in 35 OECD countries.8

PEH have poorer health than the general population,9 10 often characterised by a tri-morbidity of mental health diagnoses, chronic physical health conditions and addiction.9 Poor health is thought to be both precipitated and exacerbated by poor living conditions, lack of resources, social exclusion, stigmatisation and difficulty accessing suitable health services.11

Physical activity is beneficial for people with disabilities and chronic health conditions, both from a physical health and a social perspective. Guidance suggests that the type and amount of physical activity should be determined by a person’s abilities and the severity of their condition or disability, which may change over time.12 PEH live with a high burden of physical deficits,13 falls and frailty,14 respiratory disease, cardiac problems, stroke and diabetes,15 which could be positively influenced by physical activity. A recent scoping review found that among PEH, overall levels of physical activity appeared to be low, though the authors recognised that across studies reviewed, physical activity levels varied.16 Low levels of physical activity could be due to limited opportunities or barriers to accessing physical activity, rather than through choice. Consequently, PEH may miss out on health gains and a reduced risk of harm that physical activity affords people with these conditions. It is important that this population has opportunities for physical activity to stabilise or reverse physical declines associated with homelessness. Given the multiple barriers PEH face accessing services, it may be important that physical activity interventions are specifically tailored to their needs to optimise reach and participation. This perspective is consistent with public engagement activities with PEH and staff who care for them, which took place prior to the commencement of this research. This research poses two research questions: what is the range of physical activity interventions provided to PEH? And, what is the evidence supporting the effectiveness of these interventions?

Aims

This review aims to summarise the available evidence for physical activity interventions intended to improve health outcomes of adults experiencing homelessness, focusing on physical activity interventions and their effectiveness in improving health outcomes.

MethodsDesign

A preliminary scoping review revealed that published literature in the field of physical activity for PEH comprised both quantitative and qualitative research. Therefore, a mixed-methods systematic review was adopted. This allowed for the findings of effectiveness (quantitative evidence) and participant experiences (qualitative evidence) to be brought together, to facilitate a broader understanding of whether and how interventions worked.17 18 This systematic review was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) 2020 guidelines19 and checklist,20 and the protocol was registered a priori in PROSPERO database (reference number: CRD42020216716).

IdentificationDefining search terms

Initial search terms were generated by reviewers (JFD, RR-W and JB), who between them, have extensive clinical and research expertise and experience in the health of PEH, physical activity and systematic review methodology. The search terms were refined and tailored for a preliminary search of MEDLINE, and used to test the proof of concept and search strategy. The search syntax (online supplemental file 1) was designed by a professional librarian in collaboration with two reviewers (JFD and RR-W).

Searches

Search terms were refined, adapted and run in MEDLINE, EMBASE, Web of Science, CINAHL, PsycINFO, SPORTDiscus and the Cochrane Library. The searches were conducted on 17 February 2021, including literature from the previous 30 years (1991–2021) and restricted to English language only. The searches were re-run using the original search terms by a specialist librarian at Trinity College Dublin on 19 October 2022 to identify any new reports published between 21 October 2021 and 19 October 2022. All previous databases were searched, except SPORTDiscus, as it was unavailable in the institution’s library databases. Duplicates were removed at this stage. The reference lists of relevant systematic reviews and all included studies were hand-searched for reports to be added for screening. Corresponding authors of records that comprised an abstract only were contacted, where possible, to request full-text reports. Additionally, an expert reviewer suggested a study unidentified by searches, but met the inclusion criteria, so it was put forward for screening.

ScreeningTitle and abstract screening

On completion of the identification process, all report titles and abstracts were uploaded to the online systematic reviewing management system, Covidence. Two pairs of reviewers (JFD/RR-W and JFD/JB) independently performed (a) title and abstract screening and (b) full-text screening, judged against predetermined protocol criteria. In the event of disagreement, the third reviewer (JB or RR-W) was consulted for an additional opinion.

The PICO framework was used to identify inclusion criteria. For inclusion, all the following criteria were to be met:

Population

Studies that included adults who were homeless under the European Typology of Homelessness and housing Exclusion (ETHOS) criteria for homelessness,4 that is rooflessness, houselessness, living in insecure housing or living in inadequate housing. Age >18 years.

Intervention

Studies that included any physical activity intervention delivered as a stand-alone intervention or part of multimodal intervention, in any setting. Studies undertaken in high-income countries21 were included, where there is assumed consistency in health and social care infrastructure as well as in family and community support systems, which impact how homelessness is perceived and managed.22

Outcome

This mixed-methods review included quantitative studies reporting any measures demonstrating health outcomes, including but not limited to primary measures such as cardiovascular fitness and strength, and qualitative findings describing participant perceptions linking physical activity intervention to health and/or well-being outcomes.

Comparison

The presence of a comparison group was not required as an inclusion criterion.

Study types

This review considered quantitative, qualitative and mixed-methods studies.

Risk of bias assessment

In recognition of the diverse study designs included in this review, the Joanna Briggs Institute (JBI) critical appraisal tool portfolio was a key resource for judging quality and risk of bias.23 24 These tools provide a criterion-based checklist for determining presence (yes), absence (no), a lack of clarity (unclear) or a lack of applicability (not applicable) of quality in studies across a variety of methods.25 To determine the dependability and credibility of qualitative reports, their ConQual ratings were calculated.26 Although Munn et al discourage cut-off values in determining the quality level in quantitative studies, for clarity and consistency of this mixed-methods review, a pragmatic decision was made to select cut-offs of <25% (very low), <50% (low), <75% (moderate) and >75% (high). Munn et al state that if cut-offs are preferred, these thresholds are best decided by the reviewers themselves.25 A summary of the quality assessment of all reports is given in online supplemental file 2.

Protocol deviation

This review was registered on PROSPERO, registration number: CRD42020216716. Found at https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=216716. In the PROSPERO protocol, we stated we would use Cochrane and Downs and Black risk of bias tools. However, once the diversity of the final studies was identified, the review team recognised that JBI risk of bias tools were more suited to the studies within our review.

Data extraction

The following data were extracted to an excel spreadsheet: study design, inclusion criteria, participants (description, number, accommodation, age, education, employment, ethnicity, race, biological sex, mental health and physical health), intervention (setting, frequency, intensity, time, type, group or individual, presence of other non-physical activity intervention components), quantitative outcome measures and qualitative themes.

Initially, JFD carried out and collated data extraction from five reports. This was reviewed by RR-W and JB to ensure accuracy and consistency. Once all three team members agreed on the data extraction process, the remaining reports were divided among the team for completion of data extraction. Data from each report were checked for accuracy by another member of the research team. Any inconsistencies in interpretation or reporting were discussed, and consensus was reached.

Strategy for mixed-methods data synthesis

The synthesis followed the JBI methodology for mixed-methods systematic reviews,27 whereby established convergent, segregated, results-based mixed-methods frameworks for systematic reviewing were employed.28 29 First, qualitative and quantitative data were meaningfully categorised by JFD and JB, respectively. Each reviewer conducted their analysis separately, independently and concurrently. JFD adopted a reflexive thematic analysis approach to synthesise the qualitative data, by extracting all qualitative results into an excel spreadsheet and following the six processes of thematic analysis, namely: familiarisation; coding; generating initial themes; reviewing and developing themes; refining, defining and naming themes; and writing up.30 Details of themes are outlined in online supplemental file 3. Due to the heterogeneity of quantitative studies, it was not possible for JB to carry out a meta-analysis. So narrative synthesis was used. Quantitative findings were then ‘qualitized’ to transform them into a qualitative, descriptive format. Next, quantitative and qualitative evidence were linked and organised to produce an overall ‘configured analysis’27 and reported as a series of tables and combined narrative synthesis.

Equality and diversity statement

Our author and librarian team consisted of three women and two men. The author team included early and mid-career researchers and clinicians across two disciplines (medicine and physiotherapy) from two countries (UK and Ireland). This research explores physical activity interventions for PEH, an under-served, often marginalised and excluded population who experience extreme socioeconomic disadvantage. This population is known to have complex and chronic health needs and is an often-overlooked group in physical activity research.

ResultsStudy selection

13 737 records were identified through searches. After the removal of duplicates (n=10 122), 3615 records were screened by title and abstract, with 3496 records excluded at this stage. 119 reports were sought for full-text review, 4 could not be found, so 115 full-text reports were reviewed. Of these, 97 records were excluded at this stage (exclusions based on: 1 duplicate, 9 population, 59 intervention, 8 non-English language, 19 insufficient data, 1 protocol only). Finally, 18 reports were included for quality checking. Two reports described different aspects of a single study. Therefore, data were extracted from 18 reports describing 17 studies. The full identification, screening and inclusion process are outlined in a PRISMA diagram (figure 1).

Figure 1

PRISMA flow diagram. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-analyses.

Quality assessment

The majority of the 11 qualitative studies were high quality, with 8 reporting at least 7 out of 10 quality criteria on the JBI checklist for qualitative studies (online supplemental file 2). One study was of very low quality,31 with only the statement of researcher positionality being clear, and all other criteria either unreported or unclear. Of the quantitative studies, the one randomised controlled trial (RCT)32 was assessed as moderate quality due to methodological limitations, for example, lack of clarity regarding blinding of assessor and whether treatment groups were concealed. The analytical cross-sectional study was of moderate quality, and in general, quasi-experimental studies were of high quality.

Description of studies

Eighteen reports, describing 17 studies, were included (table 1). Of these studies, 7 were from the USA, 5 from the UK, 3 from Denmark and 2 from Australia. The variety of designs across these studies comprised 11 qualitative and 6 quantitative reports (4 quasi-experimental, 1 RCT and 1 analytical cross-sectional). The interventions addressed varied, including soccer (n=7); group outdoor exercise (n=5); group indoor multimodal exercise (n=3) and individual multimodal interventions (n=2) (online supplemental file 4).

Table 1

Summary of included studies

Study populations

Online supplemental file 5 provides detail of each study population included in this systematic review. Across the 17 studies, 516 PEH were participants. Some studies included women only (n=5), men only (n=5) or mixed cohorts (n=7). Three qualitative studies reported staff/coaches’ perspectives (n=38). The age range of participants who were homeless was 16–65 years. It was specified in the review protocol that only studies with participants >18 years would be included. However, for pragmatic reasons, several studies33–37 were included despite containing participants from the age of 16 years. In these studies, proportions of participants <18 years were not specified, although one study38 stated that the ‘majority’ of participants were between the ages of 20 and 24 years. Descriptions of study participants’ experiences of homelessness varied but were mainly focused on: street homeless, living in hostel/shelter, transitional/social service accommodation or ‘homeless at time of intervention’. Studies that focused on Street Soccer and the Homeless World Cup invited participation from PEH and other socially excluded groups, for example, people attending unemployment offices or drug rehabilitation services. Although these studies did not define proportions of participants experiencing homelessness, for pragmatic reasons they were included, as the intervention had been specifically designed for PEH. Only one study specified exclusion criteria,32 which were based on reading ability and length of time staying in the shelter. In two studies, several participants were eligible but chose not to participate,39 40 the reasons for which was not specified. The number of study drop-outs was described in three reports/two studies,32 36 37 but the reasons for drop-out were not specified.

Physical activity interventions and their components

Online supplemental file 4 provides a description of all included interventions. The studies included seven soccer interventions (tournament focused (n=2), group training focused (n=3) and combining group training and tournament participation (n=2)); five group outdoor exercise (adventure training (n=3), running (n=1) and gardening (n=1)); three group indoor multimodal exercise (aerobic-based circuits (n=2) and dance (n=1)) and two individual multimodal interventions (pedometer with step goals and earn-a-bike scheme). Online supplemental file 4 also provides programming variables, including: setting; frequency; intensity; time; type and the presence of other non-physical activity components of multimodal interventions.

Soccer

Seven studies investigated the impact of soccer for PEH. These studies (eight reports) explored soccer group training (n=4),39 41–43 tournament participation (n=2)33 40 and interventions of training for and participating in tournaments (n=2).44 45 The studies involving tournaments were focused around national or international tournaments such as the Homeless World Cup or Street Soccer USA Cup.33 40 44 45

Group outdoor exercise

Five studies provided evidence of the value of group outdoor exercise. These included group outdoor adventure (n=3),34 35 46 women’s running groups (n=1)38 and women’s gardening groups (n=1).47 These studies described multimodal interventions, including outdoor adventure interventions which contained multiple activities (eg, archery, rock climbing, hiking), and all studies reported additional support, such as the provision of education, debriefing, opportunities for reflection, childcare, food or clothing.

Group indoor multimodal exercise

All group indoor multimodal exercise studies (n=3) were instructor-led interventions provided to small groups in settings such as leisure centres36 or shelter recreation rooms.31 All studies were multimodal as they combined different types of activity, for example, stretching, cardiovascular exercise, dance, aerobic circuits, strength-based exercise to music and meditation.

Individual multimodal interventions

Two studies reported interventions for individuals.32 48 One involved participants wearing a pedometer and working towards a step goal. This was provided along with an educational newsletter and fruit/vegetable snacks.32 The other study described cycle training to learn road safety and cycle maintenance, alongside earning a bicycle for individual use.48

Intervention and outcomes

Findings are described across four tables (tables 2–5). Table 2 shows all synthesised findings relating to mental health and table 3 shows all synthesised findings relating to physical health where the configured analysis identified qualitative and quantitative evidence supporting matched themes. Table 4 shows evidence that was identified in either quantitative or qualitative reports alone. For example, findings, where only quantitative data existed, were related to bone health and blood markers. Whereas qualitative evidence only was identified relating to other important aspects of physical activity, not specifically or directly health-related, such as the benefits carried into wider life, challenges of participation and addiction.

Table 2

Summary of synthesised findings relating to mental health benefits of physical activity participation

Table 3

Summary of synthesised findings relating to physical health benefits of physical activity participation

Table 4

Outcomes where quantitative only or qualitative only findings exist, no mixed-methods synthesis

Table 5

Summary of available evidence for physical activity interventions categorised by intervention type, findings and evidence quality

The impact of physical activity interventions on the mental health of PEH

There were several domains within mental health where both quantitative and qualitative evidence was synthesised, suggesting physical activity was beneficial (summarised in table 2). These included enhanced confidence, empowerment and self-esteem; resilience, coping and hope; independence, self-regulation and personal development; stress and anxiety; and mood and state of mind.

Enhanced confidence, empowerment and self-esteem

There was high quality qualitative evidence that group running, soccer and indoor group exercise, and moderate quality qualitative evidence that group outdoor adventure and earn-a-bike enhanced confidence, empowerment and self-esteem. However, the only quantitative study to assess outcomes in this domain used the Hope scale (agency subscale), finding no significant differences between groups. One soccer player suggested:

… Football gave me confidence and took away feelings of depression as it made me more social.44

Resilience, coping and hope

There was high quality qualitative evidence that group running, and group outdoor adventure enhanced resilience, coping and hope. However, the only quantitative study to measure relevant outcomes using the Hope Scale (pathway domain) found no significant difference between intervention and control groups. A member of staff involved in delivering group outdoor adventure described changes in a participant’s ability to cope:

… when we went to Coniston, not even 10 min, we was there she wanted to come home, but she didn’t and she learned how to cope… she really enjoyed herself.35

Independence, self-regulation and personal development

Qualitative evidence suggested that group running, and soccer (both high quality) and group outdoor adventure and earn-a-bike (both moderate quality) enhanced independence, self-regulation and personal development. This was supported by moderate quality quantitative evidence that outdoor adventure improved life functioning. An outdoor adventure participant describes how it impacted them:

when I leave here, I face any challenges… in my life, then I know that I will be able to do them because I’ve become a stronger person from coming here.35

Stress and anxiety

There was high quality qualitative evidence that group running, indoor group exercise and outdoor adventure and moderate quality qualitative evidence that soccer and earn-a-bike had a positive effect on stress and anxiety. The studies that used quantitative measures to assess stress/anxiety in soccer (moderate quality) and indoor group exercise (low quality) did not conclusively support the qualitative evidence. A participant at a gym-based programme said:

I… didn’t have the confidence to go outside, I felt a lot of like anxiety and this, the gym and stuff helps me with my anxiety really well.36

Mood and state of mind

There was high quality qualitative evidence that soccer, group running and indoor group exercise and moderate quality qualitative evidence that earn-a-bike enhanced mood and state of mind. This was supported by moderate quality quantitative evidence that group outdoor adventure improved well-being.

The impact of physical activity interventions on the physical health of PEH

Changes were shown in the following physical health domains: body shape and weight loss; fitness levels; physical skills development and physical activity levels. The synthesised findings are summarised in table 3. Quantitative findings not corroborated by qualitative findings are summarised in table 4.

Body shape and weight loss

Synthesised findings showed that indoor group exercise and group running (both high quality qualitative evidence) were perceived as improving body shape and facilitating weight loss, while soccer was shown to significantly decrease weight-bearing fat mass and total fat mass (high quality quantitative evidence).

I took my measurements when I started street fit, and I took my measurements now, and I’m a lot more buff.36

Fitness levels

Synthesised findings for fitness levels showed that group running, group indoor training and earn-a-bike (all high-quality qualitative evidence) significantly improved fitness and endurance levels, a finding backed up by a high-quality quantitative study of soccer. A person who cycled with earn-a-bike described trying to increase fitness:

… after riding, you know, for an hour, two hours, and sometimes I’ll ride for four hours. You know, I really want to make sure that my body is fit.48

Physical skill development

While moderate quality qualitative evidence for group outdoor adventure was suggestive of positive changes in physical skills development, the quantitative research exploring this domain through measuring postural balance showed no significant difference between intervention and control groups. However, when comparing pre to post values in the intervention group, postural balance improved by 39% (p=0.004) in the right leg and 45% (p=0.006) in the left leg.

Physical activity levels

Synthesised findings showed that group indoor exercise and running groups (both high quality qualitative evidence) and earn-a-bike (moderate qualitative evidence) positively influenced physical activity levels. This was supported by a moderate quality quantitative pedometer and set a step count study. A woman from a running group described how since joining the group she now runs on her own:

I feel so much more body confident … I can actually run for the whole session without nearly dying. I also go out for runs on my own and I definitely think I’ve got faster.38

Bone health and cholesterol

A high quality study measured markers of bone health41 and cholesterol levels39 in PEH who played soccer. Although not all bone markers improved, increases in osteocalcin from pre-intervention to post-intervention were reported and this change was significantly different between controls and intervention groups. With regards to cholesterol markers (low-density lipoprotein-lipid (LDL)/high-density lipoprotein (HDL)) cholesterol was lowered and LDL:HDL ratios increased in the intervention group after 12 weeks of soccer—findings which were significantly different (p=0.05) from the control group.

Other considerations relevant to physical activity interventions for PEH

There were some findings relevant which described the impact of physical activity for PEH described in qualitative literature only. Themes include addiction, self-medication and medication; benefits carried into wider life and challenges to participation in physical activity when homeless (outlined in table 4).

Addiction, self-medication and medication

Across several qualitative studies of soccer (high quality) and earn-a-bike (moderate quality), physical activity positively influencing addiction was described. One person who played football stated:

I’m drinking less and do not think I need alcohol as much now… It’s great to feel this way and football is a focus for us.44

Benefits of physical activity participation carried into wider life

Most of the qualitative studies, including soccer, running groups, earn-a-bike, outdoor adventure, gardening and dance, described benefits to wider life. Subthemes included: development of life and interpersonal skills, improved social connectedness and relationships with others, practical and functional benefits, and physical activity as a catalyst for positive healthy life change. A participant who undertook leisure centre-based group indoor training said:

I’ve noticed a massive improvement in my fitness, and it’s definitely keeping me motivated to live a healthy lifestyle, because you don’t put in all that hard work and then want to ruin it, you know what I mean?36

… and similarly, how a participant of soccer described life change:

We can go back there and show that homelessness isn’t permanent and that you can change your life through sports.33

Challenges to participation in physical activity when homeless

Qualitative evidence demonstrated the importance of acknowledging specific challenges related to physical activity PEH faced, which impacted uptake and dropout rates across a variety of interventions. Those who participated in soccer tournaments described heavy defeats impacting on self-worth.44 Women who participated in running groups described lack of funds for transport or the unpredictability of homelessness as a barrier to attending.38 There was also worry about loss of donated kit (eg, running clothes)38 and equipment (eg, bicycle)48 through theft and staff who led dance groups reported inconsistent attendance among shelter-dwellers.31

An overall summary of available evidence for physical activity interventions categorised by intervention type, findings and evidence quality is provided in table 5.

Discussion

This review identified evidence for diverse physical activity interventions for PEH. The mixed-methods methodology enabled a meaningfully configured synthesis of the breadth of available evidence. This review demonstrated positive impacts of physical activity for PEH in relation to mental and physical health outcomes with translation of benefits to wider life.

Physical activity interventions were heterogeneous, grouped into broad categories of soccer, group outdoor exercise, group indoor multimodal exercise and individualised multimodal interventions. In terms of specific sports, soccer predominated (7/17). This is unsurprising considering its global resonance.37 The mental health benefits of physical activity participation identified in our review align with research carried out in non-homeless populations, for example, the psychological state of ‘flow’ (where a person feels simultaneously cognitively efficient, motivated and happy) has been found to be increased by soccer training and running.49 However, the majority (4/7) of soccer interventions included in our review included tournament participation. While benefits to tournament participation exist, negative experiences of pressure and fear of letting down teammates were qualitatively reported. Organisers of soccer tournaments for PEH should consider support to ameliorate impacts of possible pressure experienced, which could negatively impact mental health or self-management of addiction. Moreover, our review highlights that comparing the nuances of benefits and challenges of tournament participation and training warrants further research.

Group exercise appeared to be most beneficial for PEH. It is likely that group activities facilitated social support, which is especially pertinent for PEH whose social networks are often fragmented.50 Configured qualitative and quantitative findings highlighted most evidence for mental health benefits in group outdoor exercise. Specifically, these benefits related to an improvement in mood and state of mind and increased independence, focus, personal development and ability to foster relationships. This may be related to emerging evidence for optimised benefits of outdoor exercise.51 Corroboration of qualitative and quantitative evidence indicated that PEH who participated in physical activity interventions increased their physical activity levels. There is inherent difficulty comparing types of interventions for levels of benefit, as interventions and outcome measures were heterogeneous. Many physical activity interventions included additional intervention components such as counselling, food or sports kit. Consequently, it is not known if these additional components, enhanced or diluted the effect of physical activity. Moreover, descriptions of physical activity programme variables such as dosage were often lacking, limiting judgment of interventions.

Programme intensity deserves consideration. Soccer, which predominated, is a vigorous intensity sport (10 metabolic equivalent of task (METs) for competitive soccer and 7 METs for casual soccer)52 so it is likely this high entry level may be exclusionary for some PEH. It should also be considered that some participants may be content to participate on the field while exerting minimal energy, so a diversity in intensity levels is also possible. Given the high prevalence and early manifestation of non-communicable diseases and poor general physical health in many PEH,53 specifically focused lower threshold physical activity interventions should be also considered. Some low threshold programmes were identified such as gardening and dance. People designing physical activity interventions for PEH should consider a range of abilities and likely poor physical condition, perhaps offering a choice of low threshold activity, as well as higher intensity options, depending on ability and interest.

Qualitative studies dominated the evidence base, justifying the methodological decision of a mixed-methods review. The quality of evidence of most qualitative studies was judged to be high, with perspectives of staff enhancing credibility to the understanding of intervention impact. Significant changes were reported for the outcomes of weightbearing fat mass and overall fat mass in one soccer study,41 although changes in muscle mass were not reported. Cardiovascular fitness and endurance also improved significantly in soccer studies.39 42 While these findings were in small, uncontrolled studies, the implication of even minor changes to outcomes such as cardiovascular fitness and endurance may be of importance to PEH, as this group is significantly more likely than housed people to be hospitalised due to acute trauma.54 Although not specifically explored in this population, it is likely that higher baseline fitness and strength levels may aid recovery post-hospitalisation, so multifaceted programmes addressing cardiovascular endurance and strength may be most beneficial for this population. A limitation of the evidence identified was that only one quantitative study was an RCT. While RCTs are considered the highest evidence level, this review attests to the usefulness of other study designs in this novel and emerging topic. It is acknowledged that RCTs may be especially difficult to undertake due to possible implementation barriers and complexities within this cohort. We propose that to build the evidence base, forms of controlled trials should be conducted where possible, with a view to including more randomised trials in the future. A further limitation was that feasibility outcomes such as adherence and retention rates were not well described, though challenges to participation were described in several qualitative studies. Feasibility analysis, including assessment of adherence and retention, should be included in future studies. Outcome measures employed were not consistent, for example, cardiovascular fitness was measured in three different ways: the Yo-Yo endurance test, cycle ergometry and maximal treadmill testing. The evidence base is limited in terms of the most suitable outcome measures55 to use in physical activity interventions for PEH. Future studies should explore the most suitable outcome measures with a view to improving consistency in their use to enable future evidence syntheses.

Strengths of this review were its mixed-methods design and the global spread of identified research including studies from the UK and Europe, North America and Australia. However, only high-income countries were included, as low-income and middle-income settings were considered to have different structural influences on homelessness. So a limitation of this systematic review is that the translation of findings to other settings is not known. With regards to descriptions of exclusion criteria, included studies appeared to be pragmatic with minimal reporting of these. For example, no studies listed addiction status or gender diversity as pre-specified barriers to inclusion. Notably, most studies described the outcomes of ‘real world’ established programmes for PEH. In these cases, study eligibility criteria were dependent on those who engaged with the specific programme in the first instance, the eligibility criteria for which were not described, and were most likely self-selection. Only a small number of drop-outs were reported and there was minimal detail about their characteristics. We recognise that a level of stability in addiction and overall socioeconomic status is required to enable engagement in any type of physical activity intervention. Consequently, conclusions drawn from our review may not be applicable to the full diversity of PEH.

A final strength was that the review team capitalised on expertise in inclusion health, physical activity interventions and evidence synthesis with input from expert medical librarians. Studies were quality assessed using a consistent ‘family’ of critical analysis tools from JBI.

Conclusion

This mixed-methods systematic review demonstrates the value in exploring literature across a wide variety of methodological domains to gain insights into the existence and impact of a variety of physical activity interventions for PEH. To confidently inform policy, more research in this topic is required, however, from a practice and research perspective, our results provide initial justification for the inclusion of this typically under-represented group in targeted physical activity interventions with benefits to multiple aspects of physical and mental health, and positive translation into wider life demonstrated. Future research should include larger-scale high quality quantitative research to provide more robust evidence regarding objective impact.

Data availability statement

Data are available upon reasonable request.

Ethics statementsPatient consent for publication

Not applicable.

Ethics approval

Not applicable.

Acknowledgments

The authors would like to thank Jacqui Smith, clinical librarian at UCL for sharing her extensive knowledge and supporting the team with their protocol design, searching strategy and carrying out the searches. Thanks also to and David Mockler, librarian, Trinity College Dublin, Dr Cliona Ni Cheallaigh and Professor Andrew Hayward for their advice and support of this work.