Exploring the trauma experiences of people working in homelessness: mixed-methods systematic review

Abstract

Objectives

Homelessness is a global public health concern which extends to the health and well-being of people working in homelessness, and this mixed-methods systematic review aims to explore their experiences of trauma.

Design

The review protocol followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidance and a mixed-methods convergent segregated approach, a systematic search for qualitative and quantitative research across seven databases.

Data sources

CINAHL, MEDLINE, PsychINFO, Psychology Database, Public Health Database, Web of Science and PubMed, searched from inception until May 2023.

Eligibility criteria

A total of 22 articles met the eligibility criteria and were retained for inclusion (12 quantitative, 13 qualitative and 3 mixed methods). The methodological quality was assessed using the Mixed Methods Appraisal Tool, and confidence in the findings was assessed via GRADE-CERQual.

Data extraction and synthesis

Peer-reviewed empirical research and relevant grey literature, available in English, were considered for inclusion. Data were synthesised per the Joanna Briggs Institute guidance on mixed-methods systematic review synthesis, drawing upon the Substance Abuse and Mental Health Services Administration (SAMHSA) definition of trauma.

Results

The methodological quality of studies varied due to unclear reporting on sample representativeness and confounder control, particularly in quantitative studies. Despite this, a high prevalence of trauma experiences was evidenced, with a significant number of participants across this sector encountering traumatic ‘events’, ‘experiences’ and ‘effects’, as conceptualised by SAMHSA, stressing the urgent need for systemic-level change.

Conclusions

This review provides an important evidence base for future research, policy and practice, and discusses the need for psychologically informed practice, to prevent and alleviate trauma experiences across homelessness, along with nuanced, inclusive and consistent measurements of trauma.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• This is the first mixed-methods systematic review to apply Substance Abuse and Mental Health Services Administration’s²⁹ definition of trauma in synthesising both qualitative and quantitative literature on the trauma experiences of people working in homelessness.

• The review followed internationally recognised guidelines such as Preferred Reporting Items for Systematic Reviews and Meta-Analyses, enhancing transparency and replicability.

• The use of the Mixed Methods Appraisal Tool for quality appraisal and GRADE-CERQual for assessing confidence in qualitative findings strengthened the methodological rigour.

• Limitations include variability in outcome measures across studies, making it challenging to compare data effectively.

• Several studies had unclear reporting on key methodological aspects (eg, sample representativeness and confounder control), potentially affecting the reliability of the findings.

Introduction

Homelessness is an increasing global public health and human rights crisis, with an ecological lens capturing the accumulating and intersecting structural to individual-level factors that impact health.1,3 Factors such as poverty and adverse childhood experiences (ACEs) are often linked to homelessness,^{4 5} resulting in social exclusion, addiction, physical and psychological health complications, and premature mortality.⁶ Challenges in accessing inclusive primary care and preventive services compound this issue, leading to high emergency healthcare use and financial strain across public and third-sector services. Failing to address health and social care needs perpetuates poor health outcomes and fragments homelessness support systems.²

Working in homelessness is shaped by these socio-political-economic challenges,⁷ creating barriers for coordinated strategies and funding allocations which can result in ‘traumatising’ commissioning conditions.⁸ The high demand for support services reflects the relentless workloads and unrealistic expectations placed on frontline workers.^{9 10} These factors, coupled with inadequate social and emotional support, training, supervision, low wages, workplace discrimination and difficult team dynamics, can impact workplace well-being.710,13 The emotional burden exacerbates when supporting people who have experienced severe and multiple disadvantage, with homelessness professionals responding directly to threats, violence and overdoses,^{11 14} encountering a greater exposure to death^{15 16} and repeatedly hearing the trauma accounts of others.^{17 18}

Previous research highlights the psychological impact of working in these contexts, revealing high levels of stress, depression,¹⁹ post-traumatic stress^{20 21} and secondary traumatic stress (STS)^{22 23} among frontline workers in homelessness. Similar outcomes emerged during the pandemic,^{11 24 25} emphasising that the pandemic and increasing overdose crisis might exacerbate distress levels. Lived experience of ACEs and homelessness also exist among people working in homelessness,^{11 26} which could heighten distress.^{11 12 21}

Additional research is necessary to explore these psychological consequences, which impact workplace well-being, compassion satisfaction and quality of care,²² and increase challenges in retaining and recruiting workforces with the skills and experience required to practice within such contexts.²⁷ To address this, a meta-synthesis aggregated emotional experiences of people working in homelessness, revealing important individual, relational and contextual themes.⁹ Thus, the current literature adds valuable contributions to understanding the emotional experiences of this sector, its implications and strategies to improve contexts systemically.

However, although Peters et al⁹ highlight experiences of trauma in the context of ‘carrying the emotional burden’, an in-depth exploration of trauma experiences is needed. Trauma is different from other emotional experiences, such as burnout, having a distinct neurobiological impact on the human nervous system²⁸ and adverse effects on well-being, Substance Abuse and Mental Health Services Administration (SAMHSA).²⁹ While existing literature acknowledges the presence of trauma, both direct and indirect, in this sector, the extent of these experiences remains unclear.^{11 21} Therefore, a systematic review is required to explore, accumulate and synthesise the quantitative and qualitative evidence concerning the trauma experiences in this sector.

Trauma definition

This review draws upon SAMHSA’s²⁹ comprehensive definition of trauma, which is widely recognised across health and social care and informs national trauma-informed care programmes, for example, the Scottish government.³⁰ Trauma is defined as “an event, series of events, or set of circumstances experienced by an individual as physically or emotionally harmful or life threatening and has lasting adverse effects on the individual’s functioning and mental, physical, social, emotional, or spiritual well-being”. Trauma encompasses three key elements: ‘event’ (actual or perceived threat to harm or survival, single or repeated), ‘experience’ (subjective meaning assigned to events) and ‘effects’ (symptoms that adversely impact functioning and well-being). This definition acknowledges the theoretical underpinnings of neuroscience, including the effects of traumatic stress associated with the brain and nervous system survival responses, such as those described by polyvagal theory.³¹

Review aims

There is limited synthesis specifically exploring the trauma experiences of people working in homelessness. This review aims to draw upon SAMHSA’s definition to explore and synthesise the relevant qualitative and quantitative literature to date. In relation to the three ‘E’s, the quantitative data answers: (1) What is the prevalence of trauma experiences? (2) What are the factors associated with trauma experiences? The qualitative data answers: (3) What are the experiences of trauma? This review considers trauma experienced in the occupational role and personal experiences of trauma (eg, ACEs). By answering these questions, this review provides insights for future research and the implementation of Trauma-Informed Care^{29 32} and Psychologically Informed Environment principles33,35 to promote the well-being of the workforce and the individuals they support.

Methods

Search strategy

A preliminary literature search was conducted to assess the available evidence in the research area, informing the development of systematic review questions and optimising search terms for sensitivity and specificity.³⁶ The systematic review, registered on PROSPERO (ID: CRD42023380413), had deviations from the original protocol in the quality assessment and synthesis methods. These modifications were made to enhance methodological robustness and more appropriately address the mixed-methods design and diversity of the included studies. The complete search terms are detailed in online supplemental table 1.

Three search strategies were employed. First, seven databases were searched from inception until May 2023, including CINAHL,¹ MEDLINE, PsychINFO, Psychology Database, Public Health Database, Web of Science and PubMed. The search strategy incorporated keywords from preliminary searches and used controlled vocabulary to enhance the detection of relevant papers. Collaboration with an Evidence Reviewer at Mersey Care NHS Foundation Trust and a Liaison Librarian at the University of Liverpool ensured search strategy accuracy. The second search strategy included hand-searching the reference lists of all included articles and using the ‘cited by’ feature on Google Scholar. This was to identify any further articles that had not emerged from the database searches. Third, a grey literature search was conducted using Google to search relevant charity and organisation websites, including Homeless Hub, Homeless Link, FEANTSA, Canadian Observatory on Homelessness, Groundswell, Shelter, SAMHSA Homelessness Resources, St Mungo’s, Crisis and Frontline Network.

Patient and public involvement

There was no patient or public involvement in this research. This systematic review was conducted as part of the Doctorate in Clinical Psychology at the University of Liverpool, under the supervision of researchers experienced in working and conducting research in homelessness (SW, Colm Gallagher, Nick Maguire).

Eligibility criteria

To be included, the papers required the following: (1) studies available in English or for which translation resources were accessible at the time of the review; (2) an exploration of trauma experiences in line with the SAMHSA definition where the ‘event’ relates to the circumstances of working in homelessness, ‘experiences’ is the meaning assigned to events and ‘effects’ relate to traumatic stress related symptoms after the ‘event’ and ‘experience’; (3) the aim of the paper is on the experiences and emotional impact of working in homelessness; and (4) the quantitative focus is on the prevalence or the factors associated with trauma as defined by SAMHSA.

Papers were excluded if they met any of the following criteria: (1) the participant sample was focused on working with people who had not directly experienced homelessness (eg, veteran or sex worker population); (2) the participant sample focused on working with people considered as refugees or seeking asylum; (3) the research was looking at the consequences of natural disasters; (4) the stated study aims were on the practical elements or service provision, rather than the ‘events’, ‘experiences’ and ‘effects’ of working in this area; (5) where there was none or minimal reference to trauma and (6) compassion fatigue was not clearly separated into burnout and STS. This criterion aims to enhance the generalisability of findings for those working with individuals experiencing homelessness, while avoiding confounding influences from specific subgroups that have unique needs and experiences which differ from the general homelessness population.

Quality assessment

Two independent reviewers (BC and Laura Middlebrook) used the Mixed Methods Appraisal Tool (MMAT)37,39 to assess the methodological quality of eligible studies. Disagreements on three papers were resolved through discussion. The MMAT is a 27-item checklist that evaluates the quality of various study types, including qualitative, quantitative (randomised, non-randomised or quantitative descriptive) and mixed methods. It covers aspects such as methodology, data analysis, sample representativeness, outcome data reliability and researchers’ interpretation of findings. Ratings are ‘Yes’, ‘No’ or ‘Cannot Tell’. While an overall score can be calculated by counting ‘Yes’ ratings, this is discouraged due to its limited detail. Detailed criteria ratings are presented for sensitivity analysis to assess study quality.³⁹

Review selection

The review search process is outlined in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram within the ‘Results’ section (see figure 1).⁴⁰ The reference management software Rayyan⁴¹ was used for the direct exportation of citations from online databases. The search results were merged, and duplicates were removed. Titles and abstracts were screened for relevance using the eligibility criteria, and a second reviewer screened the titles and abstracts against the same criteria along with using a screening tool (see online supplemental file A). Full-text articles were retrieved to determine eligibility when it was unclear from the abstract. The same criteria were used throughout the full-article review. Discrepancies were resolved through discussion with the second reviewer (Laura Middlebrook), who assessed the full-text articles against the criteria, and by consulting the wider research team.

Figure 1. PRISMA flow diagram of study selection, including searches of databases and other sources.⁴⁰ MMAT, Mixed Methods Appraisal Tool.

Data extraction and synthesis

Three data extraction tools were developed for descriptive, quantitative and qualitative data to systematically extract key information and results. The data extracted included (1) authors and publication year; (2) location; (3) sample demographics; (4) main aims; (5) study design/method; (6) outcome measures; (7) data analysis and (8) analytic data and main findings. This mixed-methods review adopted a convergent segregated approach which initially involved the synthesis of quantitative and qualitative data independently per the Joanna Briggs Institute (JBI) guidance on mixed-methods systematic review synthesis.42,44 The review question dictated the inclusion of a range of research designs and findings. Thus, the heterogeneous nature of the extracted data meant that statistical techniques (eg, meta-analysis) were not appropriate for the synthesis.

Quantitative data were therefore narratively synthesised. Key aspects of the quantitative data across the included studies were categorised under the SAMHSA three ‘E’s definition of trauma (ie, prevalence and associated factors as detailed below in ‘Quantitative synthesis’). Textual descriptions of the included study findings were summarised and synthesised, outcome data were reviewed for similarities and differences and conclusions across the studies were drawn.

Thematic analytic methods⁴⁵ supported qualitative data synthesis, along with Thomas and Harden’s thematic synthesis⁴⁶ and JBI qualitative synthesis guidance.⁴³ A critical realist epistemological stance^{47 48} was adopted to explore how individuals make meaning of their experiences within social contexts. Acknowledgement of reflexivity is included in online supplemental file B. Data from each study were extracted comprehensively into tables to consider context and avoid bias. Data were structured around superordinate themes ‘event’, ‘experience’ and ‘effects’,²⁹ providing a theoretical structure within which to develop higher-order thematic categories. During the first stage, data were familiarised and coded line by line to capture meaning and context, with patterns identified within and across studies, aiming towards the generation of ‘data-driven’ descriptive themes. Codes and themes were organised using Microsoft Word software and reviewed by a second member of the research team (SW) for consistency. The final synthesis stage involved developing subthemes aligning with the superordinate themes. All stages were discussed with the team. Data and analysis tables were documented for transparency, increasing rigour, replicability and confidence in the review findings. A GRADE-CERQual⁴⁹ Evidence Profile and Interactive Summary of Qualitative Findings (iSoQ) table was created, enhancing the quality and reliability, and enabling confidence assessment to support informed decision-making. Each finding was assessed using the four criteria (ie, methodological limitations, coherence, adequacy and relevance) and the associated guidance.

Following independent synthesis, the quantitative and qualitative data were integrated. A configuration analysis was conducted via a meta-aggregation which involved a simultaneous comparison of the quantitative and qualitative findings. Themes were generated in line with the prevalence and associated factors explored in the quantitative studies and experiences explored in the qualitative studies to organise the evidence into a line of argument.^{42 44} No data transformation was carried out.

Results

Search results

The PRISMA guidance⁴⁰ was followed for research paper selection (see figure 1 for a flow diagram of the selection process). Electronic and hand searches identified 568 citations, with 255 remaining after removing duplicates. Following relevance screening, 192 records were excluded for not meeting eligibility criteria. A total of 61 full-text citations and reports were obtained after reviewing titles, abstracts and full-text articles. Of the 79 titles and abstracts reviewed, 74 were agreed upon for inclusion, with 5 omitted following discussion with the secondary reviewer and wider research team. Exclusion reasons are detailed in online supplemental file C. A total of 22 articles met the eligibility criteria (12 quantitative, 13 qualitative and 3 mixed methods).

Methodological quality

All papers underwent an initial screening using the first two questions of the MMAT37,39 to determine their suitability for further appraisal. Three unpublished dissertations were found to have significantly low quality and were subsequently excluded from the review, following the MMAT protocol and as depicted in figure 1 of the PRISMA diagram. A second reviewer independently assessed the quality of papers, enhancing reliability. As shown in online supplemental tables 2–5, each study was quality assessed using either one of the MMAT quantitative (ie, non-randomised or descriptive) or the qualitative categories. Mixed-methods studies were assessed via quantitative, qualitative and the mixed-methods category. Each study was quality assessed using the MMAT in either the quantitative or qualitative category, with mixed-methods studies assessed in all three categories.

In the quantitative assessments (online supplemental tables 2 and 3), most studies demonstrated well-defined variables and reliable outcome measures. However, many received ‘Cannot Tell’ ratings, particularly regarding sample representativeness, measurement appropriateness and consideration of confounding factors. For example, all 12 studies did not provide sufficient information to ensure representativeness of their samples. Seven studies did not adequately describe or justify their measurement tools, and four lacked clear strategies for controlling confounding variables. These ‘Cannot Tell’ ratings indicate that while the studies may have been methodologically sound, poor reporting hindered a comprehensive assessment of their quality. Further, three studies conducted during COVID-19 did not employ direct methods to control for the potential impact of the pandemic on results, which may have distorted results and reduced methodological quality.

Nearly all participants contributed to at least 80% of the outcome data. In the qualitative assessments (online supplemental table 4), the approaches used to answer research questions were appropriate. However, some studies lacked sufficient detail in reporting data collection methods, integration of data sources and the analysis and interpretation process. In two studies, the derivation of analysis methods was unclear. In two of the mixed-methods studies (online supplemental table 5), there were potential discrepancies and inconsistencies between qualitative and quantitative findings that were not fully reported.

Characteristics of the included studies

The following information regarding study characteristics can be found in online supplemental tables 6–8 for the quantitative, mixed methods and qualitative studies. All included articles were published between 2011 and 2023, with studies conducted in North America (n=9), the USA (n=5) and the UK (n=8). The total participant samples ranged from n=7 to n=701, covering various roles within the homelessness support sector, from frontline to management and executive positions. The majority of participants were White, female and average ages ranged from 35 to 54 years. Quantitative data were collected through cross-sectional surveys to explore trauma experiences and related factors. Constructs measured included STS (n=6), post-traumatic stress (n=6), potentially traumatic life events and experiences (n=5), and psychological distress and well-being factors associated with trauma experiences (n=5). Quantitative data analyses primarily used correlational and regression methods (n=10) and descriptive statistics (n=2). Qualitative data were collected through in-depth, semi-structured interviews (n=9) and focus groups (n=4), with various approaches to qualitative data analysis.

Findings of systematic review

Quantitative synthesis

A narrative synthesis of the included quantitative data was conducted in line with the SAMHSA²⁹ definition. The ‘Prevalence of Trauma Experiences’ relate to ‘experience’ and ‘effects’ concerning the meaning assigned to experiences (eg, recognising experiences as traumatic) and the exploration of traumatic stress via outcome measures. ‘Factors Associated with Trauma Experiences’ relate to ‘events’ and ‘effects’ concerning the associated threat of harm and the outcome of circumstances.

Prevalence of trauma experiences

Nine studies explored the prevalence of trauma experiences (see table 1 for reported levels across samples). Mixed results were highlighted for STS, with 0–75% of participants experiencing high to severe levels of STS. Prevalent STS symptoms reported were numbing, trouble sleeping, intrusive thoughts, trouble concentrating, increased annoyance with clients and difficulty recalling information.^{22 50} Across four samples, 23–42% reported experiencing high levels of post-traumatic stress, indicative of meeting post-traumatic stress disorder (PTSD) diagnostic thresholds.⁵¹

Table 1. Percentage of high to severe levels of traumatic stress experienced among samples.

Study number	High to severe levels of traumatic stress (%)	Construct (measure)
2	1.5	STS (ProQOL)
5	1.7	STS (ProQOL)
10	27.0	STS (ProQOL)
11	0.0	STS (ProQOL)
15	21.0	STS (ProQOL)
16	63.0	STS (ProQOL)
20	75.0	STS (ProQOL)
14	28.0	STS (STSS)
18	75.0	STS (STSS)
5	42.0	PTSS (PCL-6)
10	23.0	PTSS (PCL-5)
15	33.0	PTSS (PCL-C)
20	41.0	PTSS (PCL-C)

‘High to severe’ levels are defined by the cut-off levels in the individual studies.

PCL-C, post-traumatic stress disorder checklist for civilians; ProQOL, professional quality of life; PTSS, post-traumatic stress symptoms; STS, secondary traumatic stress; STSS, secondary traumatic stress scale.

Four studies explored potentially traumatic events. Among a homelessness support sector in Texas, the ACE⁵² prevalence was higher than that in the general population.²⁶ Approximately 81% reported experiencing at least one ACE, with 39% reporting four or more. Common ACEs included parental separation, emotional abuse, substance abuse and psychological distress in the home. Schiff et al²¹ revealed 95% of participants experienced at least one traumatic incident in their lives, with 88% experiencing three or more, including 56% at work. Lemieux‐Cumberlege et al⁵³ and Kerman et al¹¹ investigated workplace traumatic events. In Kerman et al,¹¹ 89% reported exposure to one or more workplace events, and in Lemieux‐Cumberlege et al,⁵³ 33% experienced at least three events in the past 6 months. The most reported included threats or verbal aggression (89–96%), responding to self-harm, suicide or near-fatal attempts (56–83%), overdose (61–81%), physical assault without injury (71%), threat of death or serious injury to self or someone close (46–66%), witnessing death (52%) and significant relationship breakdowns with colleagues or managers (35%).

Factors associated with trauma experiences

Across seven studies, factors associated with trauma experiences were explored. STS was positively associated with psychological distress and burnout and negatively associated with general well-being.^{19 23 53} Predictors of PTSD, STS and psychological distress included workplace trauma exposure and ‘maladaptive’ coping strategies.^{53 54} Increased direct client contact was significantly associated with greater exposure to trauma experiences.¹¹ Post-traumatic stress symptoms were positively associated with more frequent exposure to traumatic stressors, lived experience of behavioural health problems, more direct service work (ie, 76–100% of the time), age and negatively associated with organisational and peer support.^{11 20 54}

Despite similar symptoms to post-traumatic stress, STS was not found to significantly associate with access to organisational support in Lemieux‐Cumberlege and Taylor¹⁹ or managerial or peer support in Waegemakers-Schiff and Lane,²⁰ and, although organisational culture and compassion satisfaction have been found to be protective against burnout, they were not strong predictors of STS and post-traumatic stress when explored by Lemieux‐Cumberlege et al.⁵³ Participants with higher STS had less favourable attitudes towards trauma-informed practice and workplace safety.²³ While compassion satisfaction is negatively associated with STS^{19 53} and post-traumatic stress,²⁰ Schneider et al²³ found no relationship between compassion satisfaction and STS, highlighting some inconsistencies.

Qualitative synthesis

Synthesis of the included qualitative data was conducted, and subthemes were generated and organised under the SAMHSA constructs of trauma: ‘events’, ‘experience’ and ‘effects’. The theme occurrence across papers is depicted in table 2, and participant quotes are detailed in tables35. The GRADE-CERQual methodology was applied (see online supplemental table 9). In summary, among 15 findings, 6 are rated moderate confidence and 9 are rated high confidence, indicating a reasonable representation of the phenomenon of interest, that is, trauma experiences in the homelessness support sector.

Table 2. Occurrence of subthemes in the included.

	1. Aykanian et al85	4. Campbell et al 55	6. Kerman et al 11	8. Kerman et al 12	9. Lakeman 15	12. Levesque et al 27	13. Peters et al 13	14. Petrovich et al86	17. Theodorou et al 57	18. Twis et al 50	19. Valoroso et al 16	21. Wallace et al 14	22. Webb 84
Events
Death
Critical incidents
Trauma stories
Perceived threat
Systems trauma
Experiences
Recognition of trauma
Lived experience
Sense making
Systems trauma
Coping strategies
Effects
Traumatic stress symptoms
Cognitive changes
Overlooking and unknowing
Coping with adversity
Systems trauma

Theme discussed: yellow; theme not discussed: pink.

Table 3. Synthesis of ‘events’ themes and direct quotes from papers.

Events	Direct quotes
Death	“I've responded to multiple hangings … cut people down.” (Paper 6, pp 6) “I lost seven clients in a month with the fentanyl crisis.” (Paper 6, pp 4) “Overdoses a lot … found clients deceased … had to give CPR.” (Paper 8, pp 11) “We care, we’re supporting that person … and [they’re] on the floor almost dying.” (Paper 21, pp 87)
Critical incidents	“… someone standing in front of you with bandaging, black eye, cut up in the face.” (Paper 8, pp 11) “… seeing people with chronic health conditions … broken bones … open wounds … people lose their teeth …” (Paper 8, pp 11) “… she used to cut quite badly … it was quite shocking to see … I found myself in an … alien situation …” (Paper 13, pp 5)
Trauma stories	“… here's my awful story … you might not be quite prepared for it.” (Paper 6, pp 4) “… really immersed in these stories … it's very challenging … maintaining boundaries … not becoming immersed and … encapsulated …” (Paper 8, pp 11) “[Colleagues] ‘sliming’ people with their own … vicarious trauma that they picked up from clients.” (Paper 6, pp 6)
Perceived threat	“… not knowing if you're going to get woken up at 3:00am … because someone's dead.” (Paper 6, pp 4) “… [Staff] have been spat in the face … pushed through a window.” (Paper 13, pp 5) “… we put ourselves in … dangerous situations. You call safety [security staff], I don’t know how long it’s going to take them to get there.” (Paper 18, pp 29) “… we’re already on pins and needles.” (Paper 18, pp 29)
Systems trauma	“… people have the highest needs … we don't have 24-h support.” (Paper 6, pp 3) “I'm overworking … said to my boss like, ‘This isn't all right’.” (Paper 8, pp 9) “Poverty pimps … capitalizing off of poverty … doing this work to get money.” (Paper 6, pp 4) “… clients using racial slurs towards staff …. accusations of discrimination between different ethnic groups in our staff.” (Paper 12, pp 73)

Table 5. Synthesis of ‘effects’ themes and direct quotes from papers.

Effects	Direct quotes
Traumatic stress symptoms	“… make you ill … people do go off on sick … it compounds itself…” (Paper 13, pp 7) “… your natural reaction … fight or flight … your heart rate goes …” (Paper 13, pp 7) “… I had my first panic attack … I would be physically sick every morning … all I could see, hear, and smell was that day.” (Paper 19, pp 221)
Cognitive changes	“I expect to see someone hanging … though that’s not how I found James. I think about all the things that could go wrong … the gruesome things.” (Paper 19, pp 221) “… catastrophise things after that … now I think of [welfare checks] more for that purpose … a death watch really.” (Paper 19, pp 224) “It's death by 1000 cuts … ‘I'm fine, that didn't affect me. I don't need to debrief’ … then it happens over and over again until … ‘I can't do this anymore.” (Paper 6, pp 6)
Overlooking and unknowing	“… they've had the runaround from a lot of people that they don't [experience vicarious trauma].” (Paper 1, pp 14) “I think we totally underestimate the effect [hearing trauma stories] has on us.” (Paper 17, pp 424) “I can’t imagine what our brain is really doing with all of this.” (Paper 18, pp 29) “… our community hitting the floor … potentially dying … probably does leave … more trauma … than we necessarily recognize or have talked about.” (Paper 21, pp 87) “… [it’s] too overlooked, like actually the trauma … that’s not addressed near enough.” (Paper 21, pp 87)
Coping with adversity	“… get home … pour four ounces of gin into a cup. Pretty much every night … racking my brain of like, ‘I know I'm kind of broken from this.” (Paper 6, pp 7) “… you can't even focus on self-care cause you're in crisis mode, you spend a lot of time trying to get your mind quiet.” (Paper 8, pp 10) “[Trying to carry on like normal] I made myself ill. I wasn’t coping at all. I was not talking to anyone.” (Paper 19, pp 222)
Systems trauma	“[staff are] not acknowledged … so, what it results in is increased sick leave. People using every inch of their … leave … instead of coming into work.” (Paper 8, pp 10) “… it’s undue stress … if we don’t deal with the chaos and the lack of structure, I don’t know how … because it seems to be affecting all of us.” (Paper 18, pp 30)

Events

Five subthemes were synthesised associated with ‘events’ that risk experiencing trauma, including ‘death’, ‘critical incidents’, ‘trauma stories’, ‘perceived threat’ and ‘systems trauma’ (see online supplemental figure 1 for thematic diagram). Eight papers discussed circumstances where participants witnessed or were responsible for responding to the death of the people experiencing homelessness. These circumstances included homicide, suicide, overdose and working in palliative care (as illustrated in the direct quotes found in table 3). At times, confronting death was reported to involve violence, injustice and indignity, and the increased responsibility in responding to overdose was noted as a particular cause of concern.^{14 15} ‘Critical incidents’, including violence, abuse, property destruction, substance misuse, self-harm and health complications, were emphasised in 11 studies. These ‘events’ were surrounded by a sense of unpredictability, causing staff to feel “unsafe” as risks could not be mitigated.¹⁴ Seven studies highlighted how a broad range of systemic factors affect potentially traumatic events by creating increased uncertainty and threat. ‘Systems Trauma’ could relate to limited resources, overworking, lack of team collaboration, discrimination and stigma, exacerbated by the COVID-19 pandemic.^{7 55} The GRADE-CERQual results in online supplemental table 9 indicated high confidence in all five findings in ‘Events’, with minimal concerns regarding methodological limitations, coherence, adequacy and relevance, due to providing descriptive data that closely align with the original findings.

Experiences

Five subthemes emerged across 13 studies: ‘recognition of trauma’, ‘lived experience’, ‘sense making’, ‘systems trauma’ and ‘coping strategies’ (see table 4 and thematic diagram in online supplemental figure 2). Individual and contextual factors, such as power imbalances, social and emotional support, and cultural beliefs, influence how people interpret events as traumatic.²⁹ All papers acknowledged trauma experiences exist following events encountered, and four studies acknowledged the risk of previous trauma experiences being ‘re-triggered’ (eg, associated with substance use or fleeing domestic violence), resulting in work being “psychologically taxing”.^{27 56} Immediate emotional responses and moral distress were noted as influential experiences.⁷ Moral distress is related to ‘systems trauma’ which further shape potentially traumatic experiences by causing feelings of powerlessness and ineffectiveness.^{7 12} Organisational support, such as reflective practice and supervision, can act as a buffer against further distress and influence trauma experiences. Those who felt more supported were “least affected”, and a lack of organisational support led to resentment and distress for some people.¹⁶ Feeling undervalued and unheard by organisations increased feelings of powerlessness, anger and exhaustion; theorised to shape traumatic experiences.^{7 29}

Table 4. Synthesis of ‘experiences’ themes and direct quotes from papers.

Experiences	Direct quotes
Recognition of trauma	“… layered trauma … dealing with things in a crisis mode.” (Paper 4, pp 9) “… it's that vicarious trauma … I'm traumatized by their trauma.” (Paper 8, pp 11) “… the most distressing thing that I’ve ever dealt with … seen in my life. It’ll never leave me. I’ll always be traumatised by it.” (Paper 19, pp 220) “It was horrendous … traumatised me in a way I didn’t expect.” (Paper 19, pp 220)
Lived experience	“… emotional toll that it takes on people to have their trauma reawakened … working with people who are also living their trauma.” (Paper 12, pp 70) “… if [lived experience is] fairly recent, particularly … it does bring anxiety and issues … anyone with lived experience … has struggled.” (Paper 12, pp 71) “… those can be the very tough cases … a bit like … holding a mirror up to yourself, that’s really …difficult.” (Paper 13, pp 4)
Sense making	“… heart-wrenching.” (Paper 6, pp 6) “… physically jarring.” (Paper 8, pp 11) “If you panicked … administering naloxone … you would feel responsible for that person’s death.” (Paper 21, pp 87) “To climb into bed after you've just covered somebody up … in 20 [degrees Celsius] with a tarp … it's pretty tough.” (Paper 6, pp 5) “It's very depressing as hell … although you've saved someone's life … you're sending them out back into the streets … it's very unpleasant to put it mildly.” (Paper 6, pp 6)
Systems trauma	“… makes me feel very little, like there's nothing that I can do for this person [to obtain housing] … it's frustration and sadness.” (Paper 6, pp 4) “If we ban someone … [it’s] an arrest sentence … there's nowhere [to go] … that's inhumane.” (Paper 6, pp 5) “… on social media … people bashing certain people … it literally shatters … angers me … affects my mental health … I want to fight.” (Paper 8, pp 11) “Every single week we would complain … bring up issues … try and change things … getting zero budging from management.” (Paper 8, pp 10) “At debriefing … you can allow yourself to be completely irrational and … angry … It is important … to have a safe place to do it.” (Paper 9, pp 939) “… it was an absolute battle to get any kind [of support] … that had a massive impact on me … what it does is embed the trauma …” (Paper 19, pp 224)
Coping strategies	“I compartmentalize a lot … I can leave it at work.” (Paper 6, pp 6) “… to block everything out, I’ll go sit in the car … I just sit there in silence … helps me make it through the second half of the day.” (Paper 18, pp 29) “I was googling … really graphic things … I needed to know … the process of things.” (Paper 19, pp 223) “… you have to be very resilient … if you are not … some of the stuff you deal with is going to take you to some very dark places.” (Paper 22, pp 243)

Individual-level coping strategies influenced traumatic experiences across nine studies. Self-care, social and emotional support, establishing boundaries and distance were considered essential in alleviating distress. Strategies included avoidance and suppression of unwanted experiences (eg, putting disturbing thoughts ‘‘in boxes’’). Positive mindsets (eg, seeing the “small wins”, acceptance, positively re-framing, normalising, hope and compassion) were thought to be protective.^{7 15 56 57} Some searched for deeper meanings,¹⁶ and an element of comfort in certainty was found, such as when following organisational procedures and focusing on what can be controlled.^{15 50 56}

In GRADE-CERQual assessment, ‘recognition of trauma’ has high confidence, with minor concerns in Lakeman¹⁵ due to author interpretation, and the complex intersection between grief and trauma must be considered with caution, in that it is difficult to separate the two emotional experiences. ‘Sense making’ and ‘lived experience’ themes have moderate confidence due to concerns about data adequacy and relevance. For example, in terms of adequacy, only two papers^{27 56} contributed direct quotes from participant samples which contributed to the synthesis; however, other studies^{16 57} also supported the context of this finding via author interpretation. ‘Systems trauma’ has moderate confidence, with more details in the limitations section. ‘Coping strategies’ exhibit high confidence. For additional information, refer to online supplemental table 9.

Effects

Trauma effects can vary, affecting neurobiology, cognition, emotions, behaviour and well-being either immediately or with a delayed onset, short term or long term.²⁹ 13 studies revealed five subthemes: ‘traumatic stress symptoms’, ‘cognitive changes’, ‘overlooking and unknowing’, ‘coping with adversity’ and ‘systems trauma’ (see table 5 and online supplemental figure 3 for thematic diagram). Across six studies, cumulative exposure to trauma and constant unpredictability resulted in psychological and traumatic stress, including exhaustion, burnout, anxiety, rumination, sleeping difficulties, nightmares, intrusive memories, flashbacks, hypervigilance and “adrenaline-induced fight/flight responses”.^{7 12 15 16 56} ‘Cognitive changes’ included thinking styles such as “catastrophising”, leading some to consider leaving their roles.^{7 16}

Despite evidence of traumatic stress, four studies highlighted that the effects of trauma can be overlooked and unrecognised. At times, attempting to cope with these adverse effects resulted in substance use to “soothe”⁷ and people feeling they were unable to cope effectively.¹⁶ All studies emphasised the role of systemic factors in mitigating these effects. Systemic changes, such as addressing constant pressure of long waiting lists and high absenteeism rates, were seen as necessary to help individuals cope with traumatic experiences as these pressures were found to prolong and worsen the effects.^{7 12 50}

The GRADE-CERQual assessment indicates high confidence in the ‘traumatic stress symptoms’ and ‘coping with adversity’ subthemes. These subthemes are supported by studies with robust methodologies, coherent and adequate data. The data are highly relevant to the research question. However, the methodological quality of Levesque et al²⁷ is uncertain due to limited data analysis reporting. The assessment gave moderate confidence ratings to the ‘cognitive changes’, ‘overlooking and unknowing’, and ‘systems trauma’ subthemes, with minor data adequacy and relevance concerns outlined in online supplemental table 9 and in the ‘Limitations section below.

Meta-aggregation

Independent syntheses of quantitative and qualitative data produced complementary findings. The quantitative data indicate that there is a high occurrence of trauma experiences among the sector which are associated with specific factors unique to homelessness. The synthesised themes align with and support the quantified data. Qualitative evidence provided a detailed synthesis across all three 'E’s, enhancing understanding of the trauma experiences. Concordant outcomes indicate the presence and experiences of trauma in this sector and factors that may shape these experiences.

Discussion

This mixed-methods systematic review presents an overview of the key characteristics and findings across 22 studies. By drawing comparisons and distinctions between participant experiences, a comprehensive synthesis of the relevant literature pertaining to trauma experiences in the homelessness support sector is provided. In line with the first and second review questions, the prevalence of and factors associated with trauma experiences were explored across the quantitative data. The qualitative data details an in-depth understanding of how trauma is experienced, thus answering the third review question. The following provides a discussion of the overall results, with consideration of clinical implications, study quality, limitations and future directions.

Discussion of the findings and clinical implications

Overall, the review findings indicate that people in this sector work in extremely challenging contexts that increase their risk of traumatic stress, and a significant proportion were found to encounter ‘events’, ‘experiences’ and ‘effects’ that result in trauma, as conceptualised by SAMHSA.²⁹ The accumulating ‘events’ ranged from systemic to individual-level factors. Quantitative and qualitative findings converged in evidencing that a significant amount of this sector ‘experience’ post-traumatic stress ‘effects’ that were indicative of meeting the PTSD diagnostic threshold,⁵¹ and STS levels that indicated that participants should seek further support for their psychological well-being.

These findings complement the wider literature,^{2 6} emphasising systemic factors that worsen human health and directly increase the risk of trauma experienced in this sector. Findings also extend previous research^{9 10 13} using the Job Demand-Resources theoretical framework⁵⁸ to explain how ‘job demands’ lead to work-related distress, including trauma. These results show that the ineffective mitigation of risk and constant unpredictability of threat results in a hypervigilant support system that can feel physically, emotionally or psychologically unsafe. Therefore, these findings hold substantial implications for informing Trauma-Informed Care and Psychologically Informed Environment policies and practices in homelessness to prevent and alleviate trauma. However, despite a growing awareness of the importance of situating Trauma-Informed Care and Psychologically Informed Environment principles in homelessness,^{32 33 35 59 60} limited empirical evidence currently supports their effectiveness and therefore more research is required.^{61 62}

The synthesised themes in this review detail the complexity of subjectively shaped trauma experiences, and although the ‘effects’ of trauma are being increasingly recognised, overall experiences of trauma (eg, “what happens inside of you, as a result of what happens to you”)⁶³ can be overlooked or unknown. Instead of preventing and alleviating trauma through systematic implementation of trauma-informed practice, trauma can be shaped by unsafe relational processes that leave individuals feeling undervalued, unheard and unseen, increasing feelings of powerlessness, anger, exhaustion and moral distress. This system-level oversight could undermine the current effectiveness of Trauma-Informed Care programmes, potentially making it challenging for organisations to respond appropriately. Properly implemented Trauma-Informed Care empowers staff to recognise detrimental workplace conditions and voice concerns. These concerns are listened to, trusted and acted upon, aiming for collective positive change,⁵⁹ in line with psychological safety principles.⁶⁴ However, the findings suggest that this is not happening effectively and failing to implement Trauma-Informed Care and Psychologically Informed Environments may lead to increased financial implications and worsen health outcomes due to higher absenteeism, sick leave and turnover rates.

The qualitative results highlight the strategies to protect and alleviate trauma experiences at organisational and individual levels, consistent with factors supporting post-traumatic growth. Henson et al⁶⁵ highlighted the importance of social support in processing traumatic events, including venting negative emotions and sharing experiences.^{66 67} Post-traumatic growth increased with someone who held an in-depth knowledge of the individual’s circumstances, and positive normalising effects were found between people who shared the same trauma experiences, along with mediating effects when a sense of belonging was felt.^{68 69}

Together, these findings have significant implications regarding organisational support, alongside self-care strategies. They highlight the importance of strengths-based Trauma-Informed Care and Psychologically Informed Environment-focused practices, such as reflective practice, debriefing, training and supervision to increase team trauma awareness and readiness in trauma-focused work.^{70 71} Organisational support should be evaluated on how safe, acknowledged and empowered staff members feel, with trauma experiences normalised as natural human responses to events. Monitoring ongoing effects and providing appropriate support over time is crucial. Supervision can serve as an ideal platform to initiate such discussions, acknowledging the role of vulnerability and attachment dynamics.⁷²

Recognising the neurobiological effects of traumatic stress as embodied sensory experiences beyond semantic understanding is essential.³¹ Some people might find it difficult to articulate their experience or feel unsafe discussing them, and reflection and debriefing without containing potential physiological dysregulation might risk dissociation or re-traumatisation.⁷³ Incorporating techniques from somatic interventions based on neuroscience, polyvagal theory, attachment and trauma theories, such as Trauma-Sensitive Yoga and mindfulness,74,76 may help individuals establish a sense of safety and self-regulation without relying on language.⁷⁷ This involves supporting individuals to feel safe.^{31 78 79} However, to promote safety, it is imperative to reduce traumatic events, requiring structural changes beyond individual organisations’ capabilities. Top-down investments are needed to enhance support provided and received across homelessness.

Limitations and future directions

While the MMAT is accessible and appropriate for mixed-methods systematic reviews, it has received criticism for being ‘too simple’ and ‘difficult to judge’.³⁷ Interpretation bias in the qualitative components was possible although mitigated by a second reviewer. Future improvements to the tool could offer more detailed guidance and examples for interpreting qualitative items.³⁷ Many studies used convenience or purposive sampling, potentially limiting generalisability and introducing selection bias, especially where sample representativeness was not clearly reported. Most studies controlled for confounders using regression analysis, though the lack of clear reporting on these strategies suggests a risk of bias. The cross-sectional design of most studies, combined with the unknown impact of the COVID-19 pandemic, may further affect the findings, as the pandemic introduced sector-specific pressures.²⁵ These limitations highlight the need for caution in interpreting the review’s MMAT results and point to the importance of clearer reporting in future research to improve transparency and replicability.

Several factors may explain the mixed STS levels reported across studies. First, debates remain as to whether STS and burnout can be effectively differentiated in the ProQOL.⁸⁰ Using the two-factor model (compassion fatigue and compassion satisfaction) instead of the three-factor model (STS, burnout and compassion satisfaction) might increase consistency and construct validity.⁸¹ The heterogeneity in STS rating scales may have contributed to inconsistent conclusions about STS levels, and using a consistent measure and cut-off levels could have provided less heterogeneous results. In comparing STS outcome measures, Roberts et al⁸² recommend using the STSS due to its well-defined cut-off thresholds and excellent internal consistency. Despite some studies reporting low STS levels, the variability in how trauma is understood, assessed and interpreted suggests that trauma experiences may still exist. For instance, in the participant sample of Kerman et al,⁵⁴ 1.7% reported high STS levels, while 42% reported high post-traumatic stress levels. This highlights the need for more comprehensive measurements of trauma experiences that encompass all forms of trauma as defined by SAMHSA,²⁹ while also considering the burden on respondents.

Considering the diverse roles across the homelessness support sector and the factors influencing trauma experience (eg, amount of direct working), mixed STS levels may result from the varied nature of the work. The review reveals many potentially traumatic events and subjective factors that contribute to the experience of trauma, potentially explaining the varying STS levels across different groups. In terms of diversity, the impact of identity characteristics on traumatic stress levels is unclear as most participants identified as female, White and working-aged adults in the UK. Qualitative findings also point to issues of stigma and discrimination, requiring further exploration to understand their influence on trauma experiences. This includes stigma towards lived experience, in line with the National Institute for Health and Care Excellence.⁸³ Fear of stigma might cause underreporting of traumatic stress, aligning with social desirability research in this area.²⁰ Therefore, how societal stigma and intersectionality issues impact this sector and are associated with trauma experiences is important to explore further and aligns with Trauma-Informed Care and Psychologically Informed Environment guidance.

The GRADE-CERQual assessment raised some concerns about the confidence in the findings. For example, there are adequacy concerns regarding the ‘cognitive changes’ subtheme due to limited data from only two studies,^{7 16} potentially reducing generalisability across the sector. The ‘systems trauma’ subthemes across all three ‘E’s highlight concerns regarding the coherence of the findings with the data and review question. While the data do not suggest that systemic issues directly cause trauma experiences, they do highlight their negative impact on existing trauma-related experiences, such as hypervigilance, overwhelm and stress. Thus, although relevant, the fit between the primary study data and the subthemes ‘systems trauma’ is less cogent. The systemic issues evidenced in this review are specific to homelessness support systems in the UK, the USA and Canada and may not apply to homelessness support systems significantly influenced by geo-political issues such as war, displacement and natural disasters.

Conclusion

This mixed-methods systematic review explored the prevalence, experiences and factors associated with trauma in people working in homelessness. Results evidence that a significant amount of people across this sector encounter ‘events’, ‘experiences’ and ‘effects’ associated with trauma, and all three ‘E’s were found to be influenced by socio-political-economic factors, stressing the urgent need for systemic-level change. These findings support that homelessness is an increasing global public health concern which extends to the health and well-being of people working in homelessness. Top-down investments are crucial to protect and promote the health and well-being of people working in, and those experiencing, homelessness, along with reducing broader influencing systemic issues such as health inequalities.

The papers included in the systematic review are 710,12 14.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.