Safe spaces as an alternative ​to the emergency department for suicidal distress: ​exploring guests’ experiences​

Abstract

Background

Australians experiencing suicidal crises are routinely directed to emergency departments (ED), despite EDs being perceived as inappropriate and under-resourced to adequately address mental health and suicidal crises. To address this significant gap in care, non-clinical “safe spaces” have emerged as an alternative, but research on their effectiveness in Australia remains limited. This study evaluates guest satisfaction, perceived helpfulness of services, and factors associated with distress reduction in safe spaces.

Methods

This study analysed routine care data from Stride, an Australian safe space provider, for N = 906 first-time guests at three locations (Blacktown, Wollongong, and Belconnen). Guest characteristics, service factors, and perceived helpfulness were assessed. Regression models examined predictors of distress reduction and visit helpfulness.

Results

Guests reported significant distress reduction on exit (p <.001, 95% CI: 27.94–30.57). Greater reductions were associated with longer stays (p <.001), “coffee and chats” with a peer worker (p <.001), use of sensory modulation tools (p =.007), and outgoing referrals (p =.003), while bringing a support person (p <.001) was associated with smaller distress decreases. Follow-up data (n = 486, 53.6%), showed most guests found their visit helpful (97.9%, n = 476) and preferred safe spaces over EDs for future distress (81.3%, n = 395).

Conclusions

Findings suggest safe spaces effectively reduce distress and are a preferred alternative to EDs. This highlights their potential as a community-based, non-clinical option for individuals in suicidal crisis.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12913-025-12999-w.

Introduction

Australians experiencing suicidal distress are often directed to present at hospital emergency departments (ED) for treatment [1]. However, hospitals are often under-resourced to adequately address the complexity of mental health crises in addition to patients presenting for physical ailments [2]. It is not uncommon for people presenting to the ED for mental health crises to wait longer than other patients before being assessed, spend longer in the ED for treatment, and to leave the ED at their own risk and prior to treatment completion [3, 4]. For individuals experiencing suicidal distress, the clinical environment of the ED - marked by limited privacy, harsh lighting, medical equipment, high levels of noise and activity, and unfamiliar staff - can feel unsafe, overstimulating, and disempowering, potentially heightening existing distress [1, 5–7]. Similar experiences have also been noted by bereaved family members of those who had accessed the ED for distress prior to their suicide [8].

In addition to the environmental stressors, individuals presenting to the ED for suicidal distress often struggle to receive the compassionate care they require from ED staff. Many report feeling dismissed, being subjected to stigmatising attitudes, having their psychological needs deprioritised in favour of physical symptoms, and being discharged prematurely [6, 9, 10]. These negative experiences can contribute to feelings of shame and discourage future help-seeking [10–13]. Nearly 50% of individuals who sought help for a suicidal crisis in the ED were unwilling to return due to these experiences [14]. This is concerning given that a significant proportion of people who die by suicide had visited an ED in the year preceding their death [15].

There has been a significant push for alternative suicide prevention service models that meet the needs of people experiencing suicidal distress, including physical safety, emotional validation, practical assistance, and social connection [16, 17]. Safe spaces or safe havens have been proposed as one such alternative, offering community-based, non-clinical, and peer-led support to individuals experiencing suicidal distress [17]. The concept originated in Aldershot, United Kingdom, with the launch of the ‘Safe Haven Cafe’ in 2014, a drop-in service staffed by psychiatric nurses, mental health professionals, and peer supporters. The delivery of the Safe Haven Café was associated with a 33% reduction in the number of acute in-patient psychiatric admissions within the services catchment area during April and October 2014 [18]. In Australia, safe spaces now operate across all states and territories and are delivered by non-government organisations on behalf of state governments, volunteer groups and community managed organisations.

One provider of Australia’s safe spaces is Stride who provide specialist mental health care through an ongoing process of co-creation with the local communities they service. They manage the three government-funded safe space locations that are the focus of this study including Safe Haven Wollongong, Safe Space Blacktown, and Safe Haven Belconnen. The Stride-run safe spaces are staffed by peer workers with lived experience of suicide with the aim of being accessible and inclusive to community members. The number of peer workers and the capacity of each safe space vary across locations, depending on the space’s size, funding, and staff availability. Visitors to the services are referred to as “guests” and no referral or appointment is required, as it operates on a walk-in basis. Guests may hear about the service through word of mouth, community advertising, or online, and can self-refer. Referrals can also be made by clinicians, care coordinators, emergency departments, case managers, GPs, and other mental health professionals. Each location has set operating hours, at the time of writing, these safe spaces were open between four and five days a week, typically from early afternoon until sometime between 7:30pm and 10pm. Visits are guest-led with each guest being supported and followed up based on their own preferences and recovery needs. Guests are welcome to stay for as long as they need during operating hours, as there is no set time limit. During their visit, guests have the option of utilising sensory items, a variety of private rooms and spaces, and tea and coffee facilities. Guests can also, in collaboration with a peer worker, devise a safety and/or sensory plan.

A safety plan is a personalised, structured plan designed to help individuals manage and reduce suicidal distress by identifying support systems, and steps to ensure safety. Similarly, a sensory plan involves identifying an individual’s sensory needs and how to minimise sensory overload during moments of distress. Peer workers can also co-develop safety plans with guests or facilitate referrals and engagement with other health professionals and community services. Rather than replacing mental health services, the safe spaces serve to provide greater options, linking guests with local services and helping them develop self-coping and management skills [17, 19].

Despite the growing number of safe spaces in Australia, there is limited research on whether they meet the needs of individuals experiencing suicidal distress. This study aims to fill this gap by assessing if safe spaces meet the needs of guests and to identify service characteristics that contribute to their effectiveness in reducing distress and supporting recovery.

Methods

Design

This study is a component of the Co-Creating Safe Spaces project [20] which investigates the implementation, effectiveness, and sustainability of safe space models as alternatives to the ED for individuals experiencing suicidal distress. The current retrospective study was approved by ACT Health Human Research Ethics Committee (2022.ETH.00043) on April 04, 2022, guest consent was not required. De-identified routine care data was provided by Stride for guests at three safe spaces: October 2021 to January 2024 for Safe Haven Belconnen (n = 1,406 visits), August 2020 to January 2024 for Safe Space Blacktown (n = 1,040 visits), and May 2021 to January 2024 for Safe Haven Wollongong (n = 2,236 visits). Care data for Belconnen and Blacktown was collected at three timepoints; site visit, 72-hour follow-up, and 7–10-day follow-up, while Wollongong care data was collected at two timepoints; site visit and a 72-hour follow-up. Guests were asked at the end of their visit if they would like a follow-up call. If consent for the follow-up call was given, the call was scheduled within the agreed time frame and responses collected over the phone by a peer worker.

Outcome measures

The variables collected differed slightly across sites, therefore, for consistency, the final dataset only included variables that were collected across all sites. At service entry, these variables included guest’s unique ID code, the date the visit was completed, and their reason for visiting (such as creating a safety plan, creating a sensory plan, curiosity about the space, seeking general support, experiencing psychological distress, managing future distress, or for other reasons). Information on which aspects of the space were used by guests was also recorded, including options such as using a 1:1 private room, engaging in “coffee and chat” sessions, using the creative space, general main area, outside area, quiet time, sensory modulation area, and other areas. A brief definition of these options is provided in Table 1. The completion of safety and sensory plans were recorded as binary variables (answered yes or no). Guest distress levels were measured using subjective units of distress (SUDS) [21] upon arrival and departure, with 0 indicating no distress and 100 indicating the highest level of distress. The total amount of time each guest spent in the service was also collected (options included 0–1 h, 1–1.5 h, 1.5–2 h, 2–2.5 h, 2.5–3 h, and 3 + hours).

Table 1.

Definitions of safe space engagement options

Engagement option	Definition
1:1 private room	One-on-one peer support in a private room - or a private nook for Safe Haven Belconnen which does not have separate rooms.
Coffee and Chat	Two hours of one-to-one peer support with a coffee or tea, almost always in the open communal space. A two-hour limit is imposed by some safe spaces to manage the growing number of guests using the space at any one time. Ostensibly, a guest’s coffee and chat with a peer worker could be longer or shorter than 2 h depending on the individuals needs and how busy the space is.
Creative space	Any area in the space where guests engage in a creative activity (i.e., colouring, drawing, painting etc.).
General main area	Engaging in the spaces welcome living and kitchen areas, located centrally in the safe space.
Outside area	Receiving peer support in the outdoor area, including the backyard or front porch.
Quiet time	Time that a guest elects to spend in the space on their own, often utilised in tandem with sensory modulation and creative space. This can be in any one of the rooms.
Sensory modulation area	Any area in the space where guests engage in sensory modulation activities– like fidget tools, calming music, dimmable and colour changing lights, colouring books, scented diffusers etc. - to regulate mood through touch, taste, smell, sight, sound and movement.
Other	Any engagement option that does not fit into the above definitions.

At the 72-hour follow-up, guests reported if they were satisfied with the current service hours, and whether they had accessed any mental health services since their visit (both answered ‘yes’ or ‘no’). Guests were also asked if they would visit safe space over the ED if they were feeling distressed or experiencing a crisis, and whether they found their visit helpful (both answered ‘yes’, ‘no’, or ‘maybe’).

Data cleaning and processing

All data were deidentified with guests assigned a unique identification code. Each row of data represented a single first visit for a unique guest. Data were exported from text files to SPSS version 28.0.1.0 for cleaning. The dataset contained multiple site visits for some guests (e.g., one guest accounted for over 400 visits at one site). To avoid biases from guests with multiple visits, only the first visit per guest was retained for analysis, ensuring that each guest was represented once in the dataset. An additional variable was also added to the final dataset to indicate the site of each entry.

Accounting for procedural differences across sites

Data collection procedures for items on reason for visiting, and spaces used differed across the three sites. Specifically, Blacktown and Belconnen allowed guests to select only one option, whereas Wollongong allowed for the selection of multiple options. To maintain consistency in the analysis across all sites, each potential selection for these variables was transformed into a separate binary variable. For example, each reason for visiting (e.g., creating a safety plan) was coded as a binary variable, where 0 indicated that the guest did not select that option, and 1 indicated that they did. Similarly, each space or tool utilized was coded as a binary variable. This approach ensured that all sites were represented comparably in the final dataset, regardless of their original data collection structure. Additionally, Wollongong collected data only at two timepoints; site visit and the 72-hour follow-up. To ensure consistency, data collected at the 7–10-day follow-up by Blacktown and Belconnen was excluded from the final dataset. Responding to the SUDS was optional at Wollongong. When guests chose not to provide a SUDS score peer workers recorded their response as ‘999’, which was subsequently treated as missing data in the analyses.

Data analysis

Data were exported to SPSS and RStudio for analysis. Descriptive statistics were used to summarise guest and service characteristics including guest engagement with the space, outgoing referrals, time spent at the service, completion of safety and sensory plans, changes in distress ratings, and guests’ perceived helpfulness of their visit. Ordered logistic regression assessed predictors of whether guests found their visit helpful, with predictors including changes in distress scores, time spent at the service, receipt of outgoing referral, whether a support person accompanied the guest, and sensory and safety plan completion. Dependent sample t-tests compared differences in guests’ reported measure of psychological distress on the SUDS before and after service usage. A linear regression model was used to identify factors associated with changes in distress scores, with fixed effects for service usage factors and site. Variables with near-zero variance were identified and removed from the final model. A sensitivity analysis was performed to assess the robustness of the findings comparing the main analysis which used multiple imputation by chained equations (MICE) for the predictors and outcomes, to an analysis with non-imputed outcomes. The results of the sensitivity analysis are reported in a supplementary table, with any differences in statistical significance noted. Post-hoc tests were used to compare changes in distress across the three service sites. Chi-square tests were conducted to identify differences in service characteristics between guests who completed follow-up surveys and those who did not.

Results

Basic service and guest characteristics

The dataset contained 906 first guest visits across the three safe space sites: Belconnen (n = 208, 23.0%), Blacktown (n = 367, 40.5%), and Wollongong (n = 331, 36.5%). Follow-up data was available for 486 (53.5%) guests, with a majority collected from Blacktown (n = 332, 68.3%), followed by Wollongong (n = 88, 18.1%), and Belconnen (n = 66, 13.6%). Details of basic guest and service characteristics across sites are reported in Table 2.

Table 2.

Guest and service characteristics across sites (N = 906)

	Belconnen (n = 208)	Blacktown (n = 367)	Wollongong (n = 331)	Total (N = 906)
	n (%)	n (%)	n (%)	N
Guest Type
Had subsequent visits	100 (29.50)	122 (35.99)	117 (34.51)	339
One-time visitor	108 (19.05)	245 (43.21)	214 (37.74)	567
Reason for visit ^
Create safety plan	0 (0)	1 (12.5)	7 (87.5)	8
Create sensory plan	0 (0)	2 (16.67)	10 (83.33)	12
Curious about the space	54 (32.53)	18 (10.84)	94 (56.62)	166
General support	30 (19.74)	31 (20.40)	91 (59.87)	152
In psychological distress	109 (16.95)	310 (48.21)	224 (34.84)	643
Manage future distress	11 (21.15)	4 (7.69)	37 (71.15)	52
Other	4 (28.57)	1 (7.14)	9 (64.29)	14
Guest engagement with the space ^
1:1 private room	110 (41.51)	155 (58.49)	*	265
Coffee and chat	*	3 (2.01)	146 (97.99)	149
Creative space	2 (9.52)	1 (4.76)	18 (85.71)	21
General main area	89 (23.30)	115 (30.10)	178 (46.60)	382
Outside area	2 (11.76)	10 (58.82)	5 (29.41)	17
Quiet time	*	0 (0)	27 (100)	27
Sensory modulation	3 (2.10)	82 (57.34)	58 (40.56)	143
Other	2 (66.67)	1 (33.33)	*	3
Safety plan completion
Yes	39 (21.08)	109 (58.92)	37 (20.00)	185
No	169 (23.46)	258 (35.80)	294 (40.78)	721
Sensory plan completion
Yes	23 (19.17)	85 (70.83)	12 (10.00)	120
No	185 (23.54)	282 (35.84)	319 (40.55)	786
Total time at the service
0–1 h	80 (34.78)	42 (18.26)	108 (47.02)	230
1–1.5 h	54 (25.23)	88 (41.12)	72 (33.64)	214
1.5–2 h	38 (19.29)	104 (52.78)	55 (27.91)	197
2–2.5 h	12 (10.81)	52 (46.85)	47 (42.34)	111
2.5–3 h	14 (20.29)	32 (46.38)	23 (33.33)	69
3 + hours	10 (11.77)	49 (57.65)	26 (30.59)	85
Was the guest provided with an outgoing referral
Yes	69 (32.28)	145 (67.76)	*	214
No	139 (38.47)	222 (61.43)	*	361
Did the guest attend with a support person
Yes	57 (38.51)	91 (61.49)	*	148
No	151 (35.37)	276 (64.61)	*	427

*Variable/option not collected at this site

^Wollongong allowed for multiple selections for this variable

The most common reasons for visiting the safe space was to address psychological distress (n = 643), seek general support (n = 152), and explore the space out of curiosity (n = 166). Chi-square analyses revealed that guests visiting for ‘general support’ were more likely to complete follow-up (χ²(1) = 10.92, p =.001), while those visiting out of curiosity were less likely to complete follow-up (χ²(1) = 32.39, p <.001). Guests utilized various aspects of the safe space, with the general main area being the most frequently used (n = 382), followed by 1:1 private room (n = 265), and using the space for coffee and chats (n = 149). Chi-square analyses showed that guests who engaged with coffee and chats (χ²(1) = 58.85, p <.001) or sensory modulation (χ²(1) = 12.43, p <.001) were more likely to complete follow-up. In contrast, those who used the general main area (χ²(1) = 22.19, p <.001) or sought quiet time (χ²(1) = 13.74, p <.001) were less likely to complete follow-up. Those who completed a safety plan (χ²(1) = 14.15, p <.001) and/or sensory plan (χ²(1) = 12.01, p <.001) during their visit were more likely to have completed follow-up questions. Most guests spent less than 2 h at the service: 230 guests stayed for under an hour, 214 for 1 to 1.5 h, and 197 for 1.5 to 2 h.

Effectiveness of safe spaces in meeting guest needs

At the 72-hour follow-up, guests were asked if they would choose to visit a safe space over the ED if they were feeling distressed or experiencing a crisis. Among these respondents, 81.3% (n = 395/486) selected “yes” and 11.5% guests (n = 56/486) selected “maybe”. When asked if they were satisfied with the safe space’ current operating hours most guests, 88.1% (n = 428/486), expressed satisfaction, while 11.9% (n = 58/486) reported dissatisfaction. The vast majority, 97.9% (n = 476/486), reported their visit as helpful. There were no significant predictors of whether guests found their visit helpful.

Distress levels

Overall changes in distress

Guest distress levels, measured using SUDS, showed significant reductions from entry to exit. The average distress level on arrival was 71.47 (SD = 20.30), which decreased to 42.22 (SD = 19.74) upon departure, indicating a significant reduction, t(780) = 43.74, p <.001, 95% CI: 27.94–30.57. There were no significant differences in the average change in distress between guests who completed follow-up and those who did not (p =.226).

Predictors of distress reduction

Significant predictors of changes in distress included the use of the “coffee and chat” area (b = −9.22, SE = 1.60, p <.001), spending more time in the service (b = −2.06, SE = 0.39, p <.001), using sensory modulation spaces or tools (b = −6.98, SE = 2.61, p =.007), and having an outgoing referral (b = −3.57, SE = 1.21, p <.003). Bringing a support person (b = 8.52, SE = 1.35, p <.001) was associated with a smaller decrease in distress. Table 3 presents the regression outcomes.

Table 3.

Outcomes of regression analyses on changes in distress

Predictor	Estimate (b)	Std. Error (SE)	p-value	CI (95%)
Use of 1:1 private room	−5.14	2.64	0.05	[−10.31, 0.03]
Use of ‘coffee and chats’	−9.22	1.60	< 0.001***	[−12.36, −6.08]
Use of general main area	−3.79	2.61	0.15	[−8.90, 1.32]
Use of sensory modulation	−6.98	2.61	0.007**	[−12.09, −1.87]
Safety plan	−1.24	1.50	0.41	[−4.18, 1.70]
Sensory plan	−2.62	1.80	0.15	[−6.15, 0.91]
Service time	−2.06	0.39	< 0.001***	[−2.82, −1.30]
Outgoing referral	−3.57	1.21	0.003**	[−5.94, −1.20]
Visited with a support person	8.52	1.35	< 0.001***	[5.88, 11.16]
Blacktown relative to Belconnen	−4.36	1.73	0.01*	[−7.74, 0.97]
Wollongong relative to Belconnen	−5.24	1.58	0.002***	[−8.34, −2.14]

Significant codes: ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05

Site comparisons

Site was also a significant predictor of greater distress reduction. Relative to Belconnen, guests at Wollongong experienced significantly greater reductions in distress (b = −5.24, SE = 1.58, p <.001), followed by guests at Blacktown (b = −4.36, SE = 1.73, p =.01). Post-hoc Tukey tests showed that Wollongong guests had significantly greater reductions in distress than Blacktown guests (p <.001). A one-way ANOVA indicated significant differences in distress scores at service entry across sites. Guests at Belconnen had significantly lower baseline distress scores (M = 62.84, SD = 25.16) compared to Blacktown (M = 76.32, SD = 14.29, p <.001) and Wollongong (M = 71.55, SD = 21.08, p <.001). Additionally, baseline distress scores at Wollongong were significantly lower than Blacktown (p =.014).

Sensitivity analysis

The sensitivity analysis, which used non-imputed data for distress change, found that time spent at the service, whether the guest visited with a support person, and the Wollongong site (relative to Belconnen) remained significant. However, use of “coffee and chats”, sensory modulation tools, having an outgoing referral, and the Blacktown site (relative to Belconnen), became non-significant in this analysis, as detailed in S1 Table.

Discussion

The current study aimed to examine if safe spaces were effective in meeting the needs of guests and to identify which factors were associated with this effectiveness. The results demonstrated a significant reduction in guests’ distress from arrival at the safe space to departure. Additionally, most guests found their visit helpful and reported that in future, they would choose safe space over the ED when experiencing distress. These findings provide evidence that safe spaces are perceived by guests to be genuine alternatives to the ED when experiencing psychological distress and a feasible and effective approach for managing this distress. The significant reduction in psychological distress suggests that safe spaces are likely to be effective at providing immediate emotional relief to guests, particularly those who engage with peer support and/or sensory modulation tools, have an outgoing referral, and spend more time utilizing the service. Reinforcing this observation, almost all guests of the service who completed the follow-up indicated it was a helpful service and spending more time at the service was associated with greater decreases in distress. In addition, most guests expressed preference for safe spaces over EDs at follow-up, further reinforcing the potential of safe spaces as a suitable option for some, in addition to traditional crisis care.

This finding is particularly important given EDs, while essential for acute medical care, may not always be suited to manage suicidal crises. As discussed in the introduction, those presenting to the ED for suicidal crises often report feeling distressed or inadequately supported after leaving the service [7]. For a large proportion of guests their main reported reason for visiting the space was out of curiosity, and many guests utilised the space by having coffee and chats with a peer worker, which was associated with a decrease in distress. Data represent guests’ first entry to the services, so it is possible that mistrust and previous poor experiences in other services contributed to endorsement of “curiosity” over some other options, such as safety planning, that may require more established trusting relationships. Nonetheless these motivations and options reflect the informal, peer-based nature of safe spaces, which is not likely to be as feasible in the context of an ED, which focuses on immediate medical intervention.

These findings have important implications for the design and operation of safe spaces. The ‘coffee and chat’ area, where guests can interact with peer workers and receive two-hours of one-on-one peer support, was associated with significant distress reduction. Site differences were observed, with the Wollongong site showing greater reductions in distress relative to Belconnen and Blacktown. This may be the result of Belconnen guests presenting with significantly lower distress levels on entry to the service relative to both Blacktown and Wollongong guests. However, this may also reflect the higher uptake of the ‘coffee and chats’ option at Wollongong, highlighting the potential value of informal, non-clinical interactions in reducing distress. These results emphasise the crucial role of peer support and the value of lived experience in fostering a sense of connection and safety in individuals experiencing distress [22]. This aligns with the broader literature on the efficacy of peer-delivered mental health services, which emphasize the unique capacity of peer workers to offer empathetic, non-clinical support [23]. Additionally, these findings suggest that informal, non-clinical interactions, such as “coffee and chats” may be more beneficial for people in suicidal distress than a structured, checklist approach to assessments. This supports reports from those with lived experience of receiving psychosocial assessments after presenting to the ED with self-harm. While Quinlivan et al. found that experiences of these assessments were mixed, individuals expressed that genuine and supportive connections with staff were beneficial to their experience, while bureaucratic, “tick-box” assessments, were reported as unhelpful [24]. Additionally, this is consistent with literature highlighting the role of compassionate care and connectedness, showing that staff connections with those experiencing suicidal distress influences the quality of care [6, 7, 24, 25].

Similarly, the findings on sensory modulation tools in reducing guests’ psychological distress provide important insights into the operation and design of safe spaces. The results show that these tools are an effective self-management option for distress. They align with existing literature from ED and inpatient settings that have shown that sensory modulation items are able to reduce the users self-reported levels of distress [26–28] Qualitative assessments indicate that sensory modulation tools may facilitate emotional regulation through distraction from environmental stressors and promote feelings of calm, control, and safety [26, 29]. However, reliance on self-reported measures introduces potential bias, and controlled studies are still needed to understand to the role of sensory modulation in reducing distress and the underpinning mechanisms. Notably, the current study extends the evidence base by demonstrating the effectiveness of sensory modulation tools in the non-clinical setting of safe spaces. These results highlight a key design feature of safe spaces that distinguish them from traditional EDs, providing guests with the autonomy to choose how to manage their psychological distress. This includes offering access to a variety of sensory modulation tools and spaces, ensuring guests have multiple options to suit their individual needs.

Having an outgoing referral was associated with a reduction in psychological distress. This suggests that safe spaces have the capacity to provide immediate relief while also facilitating access to relevant and ongoing support services. Guests who receive outgoing referrals may benefit from a sense of continuity of care; however, it is possible that guests who require referrals may have more complex needs and present with higher levels of distress. Being accompanied by a support person was associated with smaller decreases in distress from arrival to departure from the safe space, suggesting that not all support mechanisms may be equally effective for all individuals. Guests accompanied by a support person may have more complex needs, or the presence of the support person may have reduced rapport or the level of engagement between the guest and peer worker. Alternatively, it’s possible that the support person was the guests’ main motivation for attendance, and that the guest may have not have self-referred otherwise. This may have impacted the guests’ engagement with the service, given that self-referral is generally associated with greater confidence and positive experience with a mental health service [30]. This finding emphasises the importance of providing a variety of referral pathways, and the need for individualised approaches in safe spaces.

Future research should further examine how individual aspects of safe spaces contribute to guests’ wellbeing. These findings could help refine service delivery to ensure that multiple options within the environment are available to meet the diverse needs of all guests. To further explore the effectiveness of safe spaces in meeting the needs of guests, future studies could look at the long-term impacts of safe space visits on mental health outcomes including ED presentations and suicidal behaviours. How the elements of safe spaces including peer support, sensory modulation, and safety planning contribute to improvements in mental health could also be examined through in-depth qualitative analysis.

The routine care data analysed for the current study lacked demographic information on guests, which presents an opportunity for future research to assess whether the effectiveness of safe spaces in reducing psychological distress is consistent across age, gender, socio-economic status, and marginalised and minority groups including culturally and linguistically diverse groups, First Nations people, and members of the LGBTQIA + community. The latter, is especially important given the goal of safe spaces to be inclusive and accessible, and with the structural inequalities faced by these communities leading to increased susceptibility to mental ill- health, distress, and suicidal crises and barriers to accessing traditional and clinical forms of care [31–34].

Limitations

While promising, this study has several limitations. Inconsistencies in data collection across sites required binary recoding of variables, potentially oversimplifying guests service interactions. Additionally, findings are based on three Stride-operated safe spaces limiting the generalizability to other safe space models. Future research should examine a wider range of safe spaces with different providers to better assess the components of a safe space that promote greater effectiveness.

Data collection was conducted by safe space staff, meaning that guests may have felt obligated to respond positively, or not respond if their experience was negative. Demographic data was also not included during collection, meaning we were unable to assess how variables, including age and gender, influenced guests’ experience, or whether safe space attendance was skewed towards certain demographic groups. Further, this study only assessed guests who attended, potentially excluding those who experienced crisis outside the services operating hours. While most guests were satisfied with availability, findings may not generalize to all experiencing suicidal distress, especially those facing access barriers.

To address these limitations, the broader research project will implement guest and community surveys to improve data quality and generalisability [20]. These guest surveys will ensure consistency across sites, capture demographic data such as age and gender, and employ confidential response methods to minimise bias. Additionally, community surveys will be distributed to gauge the awareness, acceptability, and accessibility of safe spaces across the broader community [20].

Finally, while reported guest rates of dissatisfaction were low, this limited statistical power to detect factors associated with helpfulness. Although guests showed significant reductions in psychological distress after attending safe spaces, it is necessary to acknowledge that there are many other approaches that contribute to reductions in distress including just waiting for the distress to pass. Without a controlled study we cannot be certain that changes in distress can be directly attributed to the safe space rather than the passage of time.

Conclusion

This study provides a robust and novel examination of whether safe spaces are meeting the needs of guests, and examines some potential factors associated with their effectiveness on reducing distress. Safe spaces were perceived as being helpful, and guests’ psychological distress was significantly reduced from arrival to departure from the service. Almost all guests indicated their satisfaction with the safe space they attended, and most guests endorsed them as a preferred option over attending an ED, providing support for the viability of these services as an alternative to EDs for those experiencing suicidal distress.

Abbreviations

ED

Emergency Department

SUDS

Subjective Units of Distress Scale

SPSS

Statistical Package for the Social Sciences

MICE

multiple imputation by chained equations

SD

Standard Deviation

SE

Standard Error

M

Mean

CI

Confidence Interval

LGBTQIA+

Lesbian, Gay, Bisexual, Transgender, Queer/Questioning, Intersex, Asexual

Authors’ contributions

SF, MB, and FS contributed to the conceptualization and resource provision for the study. PB, FS, and CC developed the data analysis plan. CC carried out formal analysis, and validation was conducted by CC and PB. CC was responsible for writing the original draft, and review and editing contributions were provided by PB, SF, AG, AC, HO, MB, and FS. All authors have read and approved the final manuscript.

Funding

This research was supported by an Australian Government Department of Health National Suicide Prevention Research Fund Targeted Research Grant managed by Suicide Prevention Australia.

Data availability

Data sharing is not applicable to this article as data are owned by an external agency and were provided only for the purposes of this study, so data sharing is not allowable under ethics approval. However, inquiries regarding data availability can be directed to Michelle Banfield, michelle.banfield@anu.edu.au.

Declarations

Ethics approval and consent to participate

This retrospective study was approved by the Australian Capital Territory (ACT) Health Human Research Ethics Committee (2022.ETH.00043) on April 4, 2022. The study adhered to the principles of the Declaration of Helsinki. All guest information was de-identified, and a waiver of consent for administrative data was granted by the ethics committee.

Consent for publication

Not applicable. This study does not include any identifiable personal data, images, or case reports requiring consent for publication.

Competing interests

The authors declare no competing interests.