An Evaluation of Suicide Prevention Education for People Working With Refugees and Asylum Seekers.

Background: There are concerning rates of suicidality among asylum seekers and refugees in Australia, and tailored suicide prevention initiatives are needed. Aims: We aimed to evaluate the impact of a tailored suicide prevention education program for people working with asylum seekers and refugees. Method: Attendees of the education program completed self-report questionnaires at pretraining, posttraining, and 4-6 months follow-up.
Results: Over 400 workers, volunteers, and students across Australia took part in the education program. A series of linear mixed-effects models revealed significant improvements in outcome measures from pretraining (n = 247) to posttraining (n = 231). Improvements were maintained at follow-up (n = 75). Limitations: Limitations of this research were the lack of a control group and a low follow-up response rate.
Conclusion: Findings suggest that a 2 days tailored suicide prevention education program contributes to significant improvements in workers' attitudes toward suicide prevention, and their confidence and competence in assessing and responding to suicidal distress.

International studies report elevated rates of suicidality (ideation, behavior, and deaths) among asylum seekers and refugees living in host or resettlement countries, for example, the United Kingdom, Netherlands, and United States (Ao et al., 2012; Cohen, 2008; Goosen et al., 2011; van Oostrum et al., 2011). Between 2012 and 2013, there were 846 reported incidents of self-harm in Australian immigration detention (Australian Human Rights Commission, 2013). Between 2014 and 2019, at least 27 deaths by suicide were documented among this population (boat-arrived refugees and asylum seekers) in the Australian community, immigration detention centers and regional offshore processing facilities (Border Crossing Observatory, 2019).

Refugees and asylum seekers face unique challenges placing them at increased risk of mental distress, despair, hopelessness, and suicidal ideation and behavior. Premigration risk factors such as torture and trauma history, separation from family, social isolation, loss, as well as postarrival stressors all have the potential to increase suicidality (Chen et al., 2017; Steel et al., 2009). However, Australia, like many other countries, has implemented restrictive and harsh immigration policies that are likely to compound existing suicidal distress (Posselt et al., 2020).

A recent example is Australia’s “fast track assessment” (FTA) process (Refugee Council of Australia, 2019). Under this policy, those found to be refugees would only be granted either a 3 years Temporary Protection Visa or a 5 years Safe Haven Enterprise Visa. If the person is found not to be a refugee, the decision is reviewed by a newly created Immigration Assessment Authority. Cases are reviewed by examining “the papers” of the applicant, without the individual attending a hearing or interview. The FTA caseload has left people in a state of protracted uncertainty with only temporary visa status, unable to leave Australia or work, and with limited social supports or services, and therefore at increased risk of poverty and homelessness (van Kooy, 2018). The temporary nature of their residence status and ongoing separation from family is likely to compound deepening mental deterioration (Procter et al., 2018). Combined, this prolonged and indefinite uncertainty leads to a sense of being in limbo, along with feelings of hopelessness and elevated suicidal risk (Dudley, 2003; Green & Eagar, 2010).

Suicide Prevention for Refugees and Asylum Seekers

In Australia and internationally, there is a scarcity of suicide prevention strategies developed for these groups, and those that have been evaluated to date lack supporting evidence of effectiveness (Haroz et al., 2020). While some intervention strategies have been developed internationally for refugee populations in camps (Vijayakumar et al., 2017), more research is needed to examine other potential suicide prevention strategies, such as workforce suicide prevention education. Upskilling frontline workers and volunteers may help to equip them with the necessary skills to identify and support individuals at risk of suicide. Such education programs have the potential to increase knowledge and improve attitudes and confidence among workers (Brunero et al., 2008; Ferguson et al., 2018; Morriss et al., 1999).

Emerging research supports the value of tailoring mainstream mental health and suicide prevention training and guidelines to the populations of interest. One study explored the role of Mental Health First Aid among immigrant communities (Gurung et al., 2020). While this training relates to recognizing and responding to mental health warning signs, rather than being a suicide-specific program, recent evidence indicates that incorporating culturally appropriate orientation into the training results in improved recognition and response to specific mental health diagnoses, and mental health literacy, among Bhutanese refugees. In another study (Colucci et al., 2018), the Delphi methodology was used to identify specific guidelines for the general public when supporting people from immigrant or refugee backgrounds in suicide-related distress. Through expert consensus from professionals and advocates with lived experience, the study indicated that additional items should be included compared to guidelines for English-speaking countries. How this has been applied to training programs, and the impact of specific, tailored suicide prevention education programs on people who work with asylum seekers and refugees appears to remain unknown.

This research evaluated a novel suicide prevention education program specifically tailored to working with asylum seekers and refugees. We aimed to investigate the impact on volunteers’ attitudes toward suicide prevention, confidence, and competence, when working with refugees and asylum seekers experiencing suicidal states.

Method

This project was developed through a Crowdfunding campaign (https://chuffed.org/project/preventing-asylum-seeker-suicide), initiated by the research team and in partnership with the Australian Red Cross and, in particular, with two senior staff members (VM and NC). The crowdfunding campaign called upon the Australian public to donate funds in order to address the issue of suicide among refugees and asylum seekers. Enough funds were raised to develop an education program with a research component. Ethics approval was obtained from the University of South Australia Human Research Ethics Committee.

Participants and Recruitment

Participants were staff, volunteers, and students from various nongovernment organizations who provide case management, support, or counselling to refugees and asylum seekers across Australia, primarily in the community, but also in held detention. The education program was delivered to 13 different sites across Australia (June 2017 to April 2018). Some host organization invited workers from other agencies to attend; however, only individuals from the five organizations (13 sites) for which we had ethics approval were eligible to participate in the research. Prior to delivery of the education program, staff and volunteers were informed by their organization that there would be a research component to the training; attendees were then formally invited to participate at the commencement of the program (where information sheets and consent forms were provided). Those who did not wish to participate in the research could still attend the training.

Intervention Description

Development of the 2 day (15 h) education program was led by NP, in collaboration with asylum seekers with lived experience of attempted suicide and/or witnessing suicidal behavior in held detention and/or in the community, and authors MF, NC, and VM, and incorporated publicly available data pertaining to suicide deaths for this group. Team leaders from the key training organizations were also consulted to ensure the content and learning activities embedded in the program were specifically tailored to working with asylum seekers and refugees.

The program topics included: current international and national evidence concerning suicide and suicide prevention, the development of the asylum seeker suicidal mind, trauma-informed practice, risk and protective factors, ideation to action theories of suicide, cultural considerations, compassion and hope when engaging with suicidal persons, safety planning, postvention, and self-care. The program involved lecture-style presentations (approximately 3 h), interactive discussions (4.5 h), small group activities/role plays (1.5. hr), and case studies (2 h). All attendees received a take-home workbook, including the presentation slides, worksheets, and literature. The training was conducted by NP, MF, and MP.

Outcome Measures

Participants completed a questionnaire at three time points: pre-training (T1: paper); posttraining (T2: paper); and 4–6 months follow-up (T3: online through SurveyMonkey, using the participant email address provided at T2).

The World Health Organization (2010) suggests that evaluation of suicide prevention programs should measure changes in objective or subjective knowledge and skills; attitudes; self-confidence of trainees; sustainability/behavioral changes; and user feedback. To measure the impact of our education program, participants’ attitudes regarding suicide prevention, confidence, and competence were measured. The T3 survey sought information regarding the sustainability of the program and the “on-the-ground” usefulness of the training, once applied in practice. Following T3, a subset of participants opted in to in-depth, semistructured interviews to explore their use of training skills and knowledge in practice (results reported elsewhere).

The questionnaires consisted of demographic questions (T1), previous suicide prevention training and personal exposure to and experience with suicide (T1), the three scales described in the next section (T1, T2, T3), and evaluation questions (T2 and T3).

Attitudes to Suicide Prevention Scale

The Attitudes to Suicide Prevention Scale (ASP) is a 14-item, 5-point Likert scale (strongly disagree to strongly agree) assessing participant attitudes toward the assessment and management of suicide risk (Appleby et al., 2000). Examples include: “suicide prevention is not my responsibility” and “people have the right to take their own lives.” Four items (4, 7, 9, and 14) were reverse-scored prior to analysis. A total score was generated by summing all items. Lower scores indicate more positive attitudes. The ASP possesses strong test–retest reliability and satisfactory internal consistency (Herron et al., 2001).

Confidence in Management of Suicide

Morriss et al.'s (1999) confidence scale was used to assess confidence in various aspects of suicide risk detection and management. The visual analogue scale (0 = not at all confident to 10 = extremely confident) consisted of four items related to confidence in skills to use time well, recognize risk, differentiate mild depression from suicide risk, and deal with the needs of individuals experiencing suicidal states. One item from the original scale “dealing with suicidal clients can be improved by attending a training course” was omitted. This scale has been widely used in evaluations of suicide prevention education programs (Appleby et al., 2000; Gask et al., 2006).

Competence/Skill

Competence was assessed using the Suicide Intervention Response Inventory-Version 2 (SIRI-2) – 24 items (Neimeyer & Bonnelle, 1997; Neimeyer & Pfeiffer, 1994), which allowed for the investigation of the difference between the respondents’ rating and an expert-rated “ideal” score. The SIRI-2 contains 24 client scenarios comprising brief clinical interview excerpts, each with two helper responses; participants rate these responses to client remarks on a 7-point scale (highly appropriate to highly inappropriate). An example is: Client: “I decided to call in tonight because I really feel like I might do something to myself…I’ve been thinking about suicide.” Helper A: “You say you’re suicidal, but what is it that’s really bothering you?” or Helper B: “Can you tell me more about your suicidal feelings?” Scores represent degree of variation from an ideal score (based on expert consensus); therefore, lower scores indicate greater competence and skill. This measure has strong test–retest reliability among mental health professionals (r = .92); appropriate discriminant validity from the Marlowe–Crowne Social Desirability Scale (r = −.01); and greater sensitivity to detect a training effect (i.e., pre-/posttraining changes) compared with its original form (original = no effect pre-/posttraining; SIRI-2 = significant improvement pre-/posttraining), making it relevant for use in suicide prevention training evaluation research (Neimeyer & Bonnelle, 1997).

Data Analysis

Analyses were performed using IBM Statistical Software for the Social Sciences (SPSS, 2014). Demographic, characteristic, and evaluation data were analyzed descriptively. Because we observed a large attrition rate in participation at T3, analyses (t tests and χ2 tests) to investigate potential attrition bias, comparing those who participated at all time points with those who dropped out after T2, were conducted. A series of linear mixed-effects models were conducted to assess changes in participants’ attitudes, confidence, and competence across all timepoints. For each outcome, we first examined the main effects of time (pretraining, posttraining, and follow-up), controlling for whether participants received previous suicide prevention training. Due to the possibility that the impact of the training program may differ depending on whether or not attendees had previously received suicide prevention training, a subsequent model included an interaction between time and previous training. Models used restricted maximum likelihood estimation to account for missing data. Including a random effect for organization (to account for the structure of the data, i.e., participants clustered within organizations) and random slope for time (specifying different individual patterns of change) did not improve model fit, therefore these parameters were excluded.

Results

Response Rate

Of the approximately 430 individuals who attended the training, 271 were eligible and participated in the T1 survey (92% response rate) and 231 in the T2 survey (85% response rate). There was a significant attrition rate at T3 (n = 75, 35% response rate) and only 54 responses could be linked with T1 or T2 data cases. Analyses to investigate attrition bias revealed no significant differences between T3 responders and nonresponders in age, sex, level of education, role, length of time in role, confidence levels, attitudes, or competence level.

Participant Demographics

Participant demographics are presented in Table 1.

Table 1

Participant demographics and characteristics (n = 251)

	Proportion (n)	M (SD)
Note. TAFE = tertiary and further education.
Age (years) Range: 20–73	—	37.9 (10.9)
Gender		—
Female	69% (169)
Male	30.2% (74)
Other	0.8% (2)
Highest level of education		—
Postgraduate university	37.7% (93)
Undergraduate university	37.7% (93)
Undergraduate with honours	13.0% (32)
Secondary school or TAFE certificate or diploma	11.3% (28)
Professional background		—
Social worker	57.1% (140)
Human services	11.8% (29)
Counselling	8.2% (20)
Psychology	7.8% (19)
Law	2.4% (6)
Teaching/education	2.4% (6)
Advocacy	2.0% (5)
Other	8.2% (20)
Role		—
Case manager or caseworker	45.1% (111)
Support worker	16.7% (41)
Counsellor	10.2% (25)
Volunteer	10.2% (25)
Team-leader/manager/coordinator	7.7% (19)
Student	2.4% (6)
Other	7.7% (19)
Years in current role Range: 0.1–20	—	4.0 (3.3)
Attended previous suicide prevention training		—
Yes	72.8% (179)
No	23.2% (57)
Unsure	4.1% (10)
Personal experience with suicide		—
Client has died by suicide	8.8% (21)
Client has attempted suicide	44.4% (107)
Family member/friend has died by suicide	30.3% (73)
Family member/friend has attempted suicide	28.3% (68)
I personally have attempted suicide	5% (12)

Changes in Attitudes, Confidence, and Competence

Table 2 shows the descriptive statistics and results from the linear mixed models examining changes in outcomes over time. Attitudes to suicide prevention (F = 31.53, p < .001), confidence in skills to use time well (F = 132.89, p < .001), confidence in recognizing suicide risk (F = 133.69, p < .001), confidence in differentiating mild depression and suicide risk (F = 106.42, p < .001), confidence in managing client needs (F = 139.22, p < .001), and competence (F = 13.36, p < .001) all showed significant change over time. Specifically, all outcome variables showed significant improvements at posttraining relative to pretraining and these improvements were maintained at follow-up.

Table 2

Descriptive statistics and fixed effects from the linear mixed models examining change in attitudes, confidence, and competence related to suicide prevention

Outcome	Pretraining (T1)		Posttraining (T2)					Follow-up (T3)
	n	M (SD)	n	M (SD)	B (SE)	95% CI	p	n	M (SD)	B (SE)	95% CI	p
Attitudes	202	28.58 (4.78)	201	26.45 (4.75)	−2.43 (0.31)	−3.04, −1.82	<.001	53	26.72 (5.18)	−1.97 (0.65)	−3.27, −0.67	<.01
Confidence
Skills using time	247	5.49 (2.31)	235	7.57 (1.45)	2.10 (0.13)	1.84, 2.35	<.001	54	7.07 (1.52)	1.71 (0.19)	1.32, 2.09	<.001
Recognise risk	247	4.85 (2.16)	235	7.01 (1.82)	2.16 (0.13)	1.90, 2.43	<.001	54	6.35 (1.81)	1.54 (0.25)	1.04, 2.04	<.001
Differentiate risk	247	5.36 (2.31)	235	7.30 (1.74)	2.00 (0.14)	1.73, 2.27	<.001	54	7.06 (1.86)	1.61 (0.24)	1.14, 2.08	<.001
Management	247	5.11 (2.29)	235	7.17 (1.53)	2.10 (0.13)	1.85, 2.35	<.001	54	6.96 (1.65)	1.97 (0.21)	1.55, 2.39	<.001
Competence	219	59.95 (23.22)	216	57.18 (20.14)	−3.65 (0.79)	−5.21, −2.10	<.001	51	55.32 (27.58)	−5.36 (1.34)	−8.03, −2.69	<.001

Table 3 shows results from the linear mixed models examining the moderating effect of previous training on changes in outcomes over time. There was a significant interaction between previous training and time for all four confidence items: confidence in skills to use time well (F = 12.71, p < .001), confidence in recognizing suicide risk (F = 3.09, p = .049), confidence in differentiating depression and suicide risk (F = 11.96, p < .001), and confidence in managing client needs (F = 23.58, p < .001). Therefore, participants without previous training in suicide prevention showed greater improvements on all confidence variables relative to those with previous training. These differences were maintained at follow-up, with the exception of confidence in recognizing suicide risk. There was no evidence to suggest that participants with and without previous training in suicide prevention different in their improvement in attitudes to suicide prevention (F = 1.91, p = .15) or competence (F = 1.86, p = .16).

Table 3

Estimated mean changes and fixed effects from the linear mixed models examining the moderating effect of previous training on change in attitudes, confidence, and competence related to suicide prevention

Outcome	Previous training	Posttraining				Follow-up
		Changea		Group × Time interaction		Changea		Group × time interaction
		M	B (SE)	95% CI	p	M	B (SE)	95% CI	p
Note. Significant p values are highlighted in boldface. aBased on the estimated marginal means.
Attitudes	No	−3.43				−1.98
	Yes	−2.10	1.32 (0.71)	−0.08, 2.73	.06	−2.05	−0.08 (1.41)	−2.90, 2.74	.96
Confidence
Skills using time	No	3.10				2.72
	Yes	1.72	−1.38 (0.28)	−1.92, −0.83	<.000	1.31	−1.40 (0.41)	−2.22, −0.59	.001
Recognise risk	No	2.69				2.03
	Yes	1.97	−0.72 (0.30)	−1.30, −0.14	.02	1.34	−0.69 (0.54)	−1.76, 0.38	.20
Differentiate risk	No	3.03				2.61
	Yes	1.61	−1.42 (0.30)	−2.00, −0.84	<.000	1.19	−1.42 (0.50)	−2.41, −0.42	.01
Management	No	3.39				3.35
	Yes	1.63	−1.76 (0.26)	−2.29, −1.25	<.000	1.41	−1.93 (0.44)	−2.81, −1.25	<.000
Competence	No	−6.21				−6.84
	Yes	−2.78	3.43 (1.79)	−0.10, 6.95	.06	−4.90	1.93 (2.94)	−3.96, 7.82	.51

Program Feedback

The majority of participants (n = 220) completed T2 questions regarding training feedback. Of those, most either agreed or strongly agreed that the program was relevant to their role (95.9%), that the training will make a positive contribution to their professional practice (95%), and that the training held their interest (95.5%).

Most participants (n = 212) made additional comments about the most helpful aspect of the training. Participant comments related to appreciating the tailored nature of the training (to their client group), the opportunity to participate in role plays or observe others, learning intervention approaches such as safety planning, and learning about key concepts in suicide (e.g., cognitive constriction, fluctuating states, and acquired capability). Fewer participants (n = 62) commented on the least helpful aspects of the training; comments related to use of role plays, aspects of the training where the participant felt they already possessed adequate knowledge, and some felt there was too much theory presented. Other comments related to the training environment (room temperature and group size) and length of the training.

At T3, 63 of the 75 participants (84%) reported that they believed that the training helped them change the way they work with clients in suicide-related distress. Reported changes included improving the way they communicate with clients (71.2%, n = 47), asking clients more direct and specific questions about suicidal ideation (59.1%, n = 39), involving clients in co-producing safety plans (40.9%, n = 27), improving the way they communicate with other stakeholders about client safety needs (34.9%, n = 23), improving the way they recorded client interactions in files (27.3%, n = 18), and asking more direct and specific questions about homicidal ideation (19.7%, n = 13). By contrast, only three participants (4.6%) reported that they had not made any changes since the training.

Discussion

This research evaluated the impact of a 2-day suicide prevention education program on the attitudes toward suicide prevention, confidence, and competence of workers and volunteers who support refugees and asylum seekers experiencing suicidal states. The findings suggest that the program was associated with positive changes for participants on all outcome measures. Specifically, participants experienced more positive attitudes toward the assessment and management of suicide risk and suicidal individuals on completion of the training. Participants’ confidence in suicide risk detection and management also increased significantly. Finally, the results suggest that on completion of the training, participants also became more competent or skilled regarding appropriate ways to respond to individuals expressing suicidal distress. Most importantly, these observed improvements were all maintained at follow-up.

These results are significant in the context of Australia's Fifth National Mental Health and Suicide Prevention Plan (Commonwealth of Australia, 2017). Priority Area 2 of the plan identifies the importance of training, education, and capacity building for communities, particularly in terms of gatekeeper training, strengthening referral networks and crisis responses. The data suggest that this training program has been effective in engaging with and responding to some critical practice needs of key gatekeeper and support persons in the community. Our findings align with reviews in the gatekeeper literature indicating that training can have a positive impact on attitudes for responding, knowledge (competence), and self-efficacy for intervening (confidence and competence; Burnette et al., 2015; Isaac et al., 2009). Development in all three areas is important for encouraging the development of intentions to intervene, as well as actual intervention behavior (Kuhlman et al., 2017).

There was no evidence to suggest that participants with and without previous training in suicide prevention differed in their improvement in attitudes to suicide prevention or competence. A theme in the gatekeeper training literature is that training can be adapted to suit the particular context of client groups and related workforces (Isaac et al., 2009). Participants may have experienced improvements in attitudes to suicide prevention or competence irrespective of previous training in suicide prevention due to the highly tailored nature of this training. We acknowledge, however, that significant differences between groups may have been evident in a larger sample.

Based on observation and evaluation data, the training appeared to be particularly valued by and pertinent to participants because it addressed fundamental drivers for suicide and self-harming behavior in this community. This included in-depth discussion about factors including uncertainty regarding the future, living in limbo, temporary rather than permanent protection, feeling a burden to oneself and to others, and excruciating despair linked to being separated from family and loved ones.

Another key aspect of the training, according to evaluation data, was using content and case examples to help participants gain greater comprehension about how these factors can impact on the psychology and behavior of clients, and how suicidal processes can develop. This included understanding how experiences of multilayered processes of the FTA procedure could also be linked with feelings of entrapment and cognitive constriction, and what effective responses could be provided. A further factor was that the training acknowledged the specific challenges that workers face, in the particular context of asylum seeker law and practice in Australia, could lead to feelings of helplessness. From the perspective of prevention training, it was important to acknowledge and validate participant experiences here, while also affirming that positive support of clients could be achieved, and that participants could strengthen their service environments through learning and development (Isaac et al., 2009).

Strengths, Limitations, and Future Directions

A significant limitation of this research was the lack of a control group. We cannot rule out practice effects or potential response biases, due to the self-report outcome measures. However, given that the SIRI-2 measured competency using items presenting clinical scenarios and responses that essentially assess knowledge or skill, we feel confident that this scale may be less vulnerable to response biases such as social desirability bias, particularly also given its good discriminant validity from a social desirability scale. Future research would benefit from both incorporating a control group, obtaining feedback from asylum seeker and refugee individuals who have received support and intervention from trained participants, and examining the impact on the incidence of suicidal behavior and suicide rates among refugee and asylum seeker populations.

Another limitation of this research was the low response rate for T3. A likely explanation is that one of the participating organizations, from which 65.7% of the T1 sample came, experienced significant staff redundancies due to funding cuts at the time that T3 data collection was commencing. Many of these participants had provided only a work email address to receive the T3 survey link and therefore were no longer contactable. A potential confounding factor could be that some participants had previously undertaken suicide prevention training.

Regardless, this research was the first, to our knowledge, to evaluate the impact of suicide prevention education on a workforce supporting refugees and asylum seekers. This research has enabled us to advance our knowledge of the potential benefits of suicide prevention initiatives specific to addressing suicide among refugee and asylum seeker populations.

Conclusion

Findings suggest that a 2 days tailored suicide prevention education program can lead to significant improvements in workers’ confidence, attitudes, and competence – key markers associated with a worker’s ability to more effectively identify, respond to, and manage suicidal distress. Tailored suicide prevention initiatives appear to be vital for understanding and responding to the unique nature of refugee and asylum seeker suicidal distress and behavior.