Interventions for burnout and well-being in homelessness staff: a systematic scoping review

Study design

This scoping review was conducted in accordance to the Joanna Briggs Institute methodology for scoping reviews,¹⁵ based on Arksey and O’Malley¹⁶ and Levac et al’s¹⁷ framework. The review is reported using the PRISMA extension for Scoping Reviews (PRISMA-ScR) flow diagram.¹⁸ The review protocol was registered on Open Science Framework (OSF) in May 2023 (S1 File).¹⁹

Research questions

This scoping review addressed the following questions:

1. What interventions have been implemented in the homelessness sector to address staff well-being and burnout?

   1. In what settings and context were these interventions carried out?

   2. What measurement tools and outcomes were used to evaluate well-being and burnout in these studies?

   3. How did the interventions change practice?

Eligibility criteria

For the purposes of this review, well-being included any intervention addressing stress, burnout, job satisfaction, compassion fatigue, secondary traumatic stress, vicarious trauma, post-traumatic stress and well-being itself. These aspects have previously been identified as part of the emotional pressures faced among homelessness staff.^13,14

The inclusion criteria followed the Population, Concept, and Context criteria (see Table 1). Studies were selected if they met the following three criteria: (1) the intervention specifically addressed burnout and well-being in homelessness staff and/or trainees; (2) full-text was available in the English language; and (3) the evaluation contained primary data. Evidence sources were left open, with no date restrictions.

Information sources and strategy

An initial search of Medline, PsychInfo, Global Health, ASSIA and CINAHL was undertaken to identify articles relating to the review. In addition, recommended search strategies from a related systematic review and scoping review were used to supplement the initial scoping searches.^13,14 An academic librarian was subsequently consulted to help refine the search terms and databases.

The final search strategy included five electronic databases: Medline, PsychInfo, Global Health, ASSIA, CINAHL. The search strategy was conducted on August 28^th, 2023 by LN in English, due to language limitations of reviewers, and adapted to each database, with no date limitations. To identify any additional studies, Google Scholar was searched using the following terms: (“burnout”) and (“homeless”) and (“staff”) and (“intervention”). The first 300 articles in the Google Scholar search to appear were included in the screening. References of the final included sources were also screened for supplementary articles; however, none were identified. An example search string from Medline and PsychInfo is shown in Table 2.

Study selection

Our search initially yielded 8,447 articles, in addition to 300 articles retrieved from Google Scholar. SR-Accelerator was used to remove any initial duplicates, with further duplicates removed by Covidence or manually by a reviewer. Search results were uploaded onto Covidence. Inclusion and exclusion criteria were completed by the primary reviewer (LN). Four reviewers (LN, EA, DA, ED) completed Title and Abstract screening independently and two reviewers (LN and EA) completed Full Text review and Data Extraction independently. Two reviewers, including the primary reviewer (LN), independently assessed the papers and identified if they met the inclusion criteria. Where there were discrepancies in study selection, a third and fourth reviewer (SM and EA) adjudicated on the final decision. Fig 1 summarises the screening process and reasons for exclusion.

• Download figure
• Open in new tab

Figure 1. PRISMSA flowchart of scoping review

Quality assessment

Quality assessment of the final included studies (n=6) was carried out by the primary reviewer (LN), using the Mixed Methods Appraisal Tool (MMAT).²¹ While papers were not excluded based on poor quality, the tool was used to the ascertain if the methodologies were appropriately carried out and to make an informed judgment of its findings.

Study characteristics

The study characteristics are summarised in Table 3. Most studies were not published in peer-reviewed journals (n=4),^21–24 with only two undergoing peer-review (n=2).^25,26 Among the non-peer-reviewed studies, three were dissertations published in an online database^21,23,24 and one was an unpublished manuscript from an institutional repository.²² No included studies were retrieved from Google Scholar. Nearly all were published in the last fifteen years (n=5).^{21–23,25,26} Most were conducted in the United States (US) (n=4)^21,23–25, with the remaining two conducted in United Kingdom (UK).^22,26 Information on participant demographics, namely age and/or gender, were missing in four of the six studies.^22,23,25,26 From the available data, most participants were female. The settings where the studies took place included specialist homelessness organisations (n=3)^21,22,26, a domestic violence (DV) shelter (n=1),²⁴ a medical home for veterans experiencing homelessness (n=1),²⁵ and a community healthcare organisation for underserved populations, including PEH (n=1).²³ Quantitative pre-experimental design was the most common study design used (n=4).^21,22,25,26 Only one study used a randomised control trial (RCT) design²⁴ and one used a mixed-methods non-experimental design.²³

Interventions

All interventions varied in nature, with their respective components detailed in Table 4. Five were complex interventions,^{21,22,24–26} defined as interventions consisting of several interacting components and measuring multiple outcomes.²⁷ Three interventions involved therapy components, namely cognitive behavioural therapy (CBT),²² mindfulness,²³ and acceptance and commitment therapy.²⁶ Two of the interventions comprised of educational sessions, one of which involved a session on self-care²¹ and the other presenting a well-being toolkit.²⁵ Two of the sessions also incorporated elements of supervision in the intervention, namely feedback on secondary traumatic stress for homelessness staff²⁴ and psychologist supervision for CBT training.²² Four of the six interventions completed mainly in-person.^21,22,25,26 One intervention involved delivering a mindfulness intervention through an online platform²³ and another intervention used an anonymous postal feedback survey for homelessness staff on secondary traumatic stress symptoms.²⁴ Most interventions were evaluated over one to three months.^21,23,24,26 The longest evaluation period was over an academic year, estimated to be approximately 8-10 months, although the exact duration in months was not specified in the study.²⁵

Outcomes and measures

The outcomes and measures used to assess well-being in homelessness staff are listed in Table 4.

With regards to outcomes, four studies assessed burnout.^21,22,25,26 Two studies assessed staff beliefs on self and/or others, including service users.^22,24 Two studies evaluated general well-being of staff^21,26 and two studies evaluated coping abilities.^24,26 Two interventions used bespoke questions assessing the interventions themselves.^23,25 One study used stress and resilience as an outcome measure²⁵ and another study used PTSD symptoms as an outcome measure.²⁴

Nearly all measures used to evaluate well-being varied in the six studies. Only two studies used the same measurement tool, the Maslach Burnout Inventory (MBI).^22,25 Only one of the MBI scales were used in Moore et al.’s study.²⁵ Moreover, almost all studies incorporated non-validated measures in their evaluation.^22–26

Quality appraisal

Nearly all studies had methodological limitations, mainly owing to small sample sizes, high drop-out rates, insufficient details on the study’s recruitment and methodology, use of non-validated measures, and lack of accounting for confounding variables (S1 Table). Based on the MMAT criteria,²⁰ four of the six studies scored between 0-20% in methodological quality,^22–25 one study scored 40%,²¹ and one met all of the appraisal criteria.²⁶ Although one RCT evaluation was included,²⁴ the quality was poor and lacked rigor to draw conclusions from its findings. No power calculations were conducted in any of the studies. However, Reeve et al’s study²⁶ identified that a minimum of three participants were required for establishing an effect in single-case experimental design research, as used in their study, and highlighted that their study met this respective criterion.

Key findings and recommendations

The study’s key findings and future recommendations are shown in Table 5. The results of three studies reached statistical significance.^21,22,26 Two interventions demonstrated statistically significant improvements in burnout among homelessness staff following an in-person educational session on self-care,²¹ and following CBT training and supervision.²² The Acceptance and Commitment Therapy intervention illustrated statistically significant increased psychological flexibility in half of the participants.²⁶

In contrast, no statistically significant differences were seen in secondary traumatic stress levels following feedback²⁴ and after implementation of the well-being toolkit,²⁵ although the latter did demonstrate trends towards improvement in resilience and mindfulness over 8-10 months. A downward trend in secondary traumatic stress scores were seen following the in-person educational session on self-care, however the results were not statistically significant.²¹ Similarly, after CBT training and supervision, there was a reduction in negative beliefs about PEH, but these did not reach statistical significance.²²

One study did not carry out any statistically significant tests.²³ However, this was the only study that examined sustainability, rather than outcomes, and highlighted that time and workload were barriers to completing the mindfulness modules.²³

In nearly all studies, the most common recommendation was exploring the role of group interventions, rather than individual approaches (n=4).^21,23–25 Other recommendations included conducting randomised controlled trials to isolate the effects of the intervention (n=2),^21,22 and determining sustainability of the intervention.^22,25 No adverse events were reported in any studies.

Strengths and limitations

The strengths of this study include its systematic and in-depth approach to ensure a high-quality study was conducted. Its broad inclusion criteria, search of five databases and Google Scholar, with no time limitations, ensuring a diverse range of papers were assessed. The use of an academic librarian to assist in the search strategy helped reduce the chance of bias in the review. Furthermore, the quality assessment conducted guided the reviewers’ ability to determine whether meaningful conclusions could be drawn.

However, the study was limited to English language publications, potentially missing evidence in other languages. Due to language limitations of reviewers, it was not possible to broaden this further. Moreover, the multidimensional nature of well-being may not have been fully captured in the search strategy, potentially leaving gaps in the literature search. The reviewers attempted to address this by reviewing terms from key papers and recent similar reviews, with regular input from an academic librarian, to ensure a wide and inclusive search scope was upheld. Furthermore, the inclusion of DV shelter workers, while justified by their insecure housing status,²⁸ may have introduced some contextual differences. Nonetheless, similar themes to that of homelessness staff have been identified in the literature on DV support workers, including high work demands, burnout, compassion fatigue, and secondary traumatic stress.²⁹ Finally, grey literature searches were limited to Google Scholar, which may have omitted some relevant sources. While no studies from Google Scholar were ultimately included in this review, adopting a systematic approach to grey literature searches in future reviews would help ensure that all relevant evidence is captured.³⁰

Comparison with literature

Of the three studies that achieved statistical significance, two were therapeutic interventions, involving relational components facilitated by clinical psychologists,^22,26 suggesting a potential benefit for clinical psychologist roles in homelessness settings. A recent qualitative study similarly highlighted the value of on-site trainee clinical psychologists in homelessness settings, in terms of providing staff support and promoting psychologically-informed approaches.³¹

Notably, the most common recommendation from studies was to explore group interventions.^21,23–25 A scoping review on vicarious trauma correspondingly highlighted that group interventions foster group cohesion and support, which helps mitigate secondary traumatic stress symptoms.³² Additionally, the ‘Florence Nightingale effect’ suggests that staff who strongly identify with their organization may experience lower burnout and higher job satisfaction when they view client suffering through a lens of organizational commitment, rather than as a traumatic event.³³ Thus, enhancing organisational identification through group interventions could be a valuable approach for reducing burnout and improving well-being in the homelessness sector.

Implications for research and practice

All studies employed a wide heterogeneity of measures and outcomes to evaluate well-being, with only two studies using the same measurement tool, and several relying on non-validated tools. The lack of standardisation meant comparability across studies was not possible, and outcome accuracy and reliability were difficult to assess. To improve future research, it is essential to identify and agree upon the most validated measure(s) for assessing well-being and burnout among homelessness staff. A Delphi exercise could be an effective approach to achieving consensus on the most appropriate measures for this group.³⁴

Furthermore, most studies were generally low quality, with poorly characterised demographics and methodologies, small sample sizes, and no power calculations, limiting reliability of conclusions. The majority were also quantitatively focused with minimal qualitative insights, leaving the underlying barriers and facilitators of intervention engagement unclear.³⁵ Many studies also employed single-group designs with short follow-up periods, making it difficult to assess the full effects and sustainability of interventions. Researchers have underscored appropriate follow-up times are crucial to capture the full impact and sustainability of interventions addressing well-being.³⁶ Future research should prioritise robust study designs, with adequate power calculations and follow-up periods, to appropriately capture intervention effects. Incorporating Medical Research Council’s complex intervention guidelines will also enhance study rigor by ensuring interventions are systematically developed, evaluated, and refined. This approach would more effectively capture the complexity of intervention effects and provide more reliable evidence for improving well-being.²⁷