Introduction
We are in the midst of a global crisis of forced displacement (Patel et al., Reference Patel, Saxena, Lund, Thornicroft, Baingana, Bolton, Chisholm, Collins, Cooper and Eaton2018). Worldwide, tens of millions of refugees and asylum seekers have been forcibly displaced by war, violence and persecution (UNHCR 2020). The trauma and stress experienced before, during and following forced migration contribute to high rates of stress- and trauma-related mental health problems (Priebe et al., Reference Priebe, Giacco and El-Nagib2016). We focus here on one particularly destructive consequence of this crisis – suicidality and, specifically, suicidal ideation, post-displacement among adults (WHO 2014; Colucci et al., Reference Colucci, Too and Minas2017; Ventevogel et al., Reference Ventevogel, Ryan, Kahi and Kane2019).
A limited number of existing studies have focused on suicidality among adult refugee populations who have the relative fortune of residing in stable resettlement contexts post-displacement, most typically in high-income countries (Vijayakumar, Reference Vijayakumar2016; Colucci et al., Reference Colucci, Too and Minas2017). In stable re-settlement post-displacement contexts, point- and lifetime-prevalence of suicidal ideation among resettled refugees (3–18.7%) (Jankovic et al., Reference Jankovic, Bremner, Bogic, Lecic-Tosevski, Ajdukovic, Franciskovic, Galeazzi, Kucukalic, Morina, Popovski, Schützwohl and Priebe2013; Ao et al., Reference Ao, Shetty, Sivilli, Blanton, Ellis, Geltman, Cochran, Taylor, Lankau and Cardozo2016; Nickerson et al., Reference Nickerson, Byrow, O'Donnell, Mau, McMahon, Pajak, Li, Hamilton, Minihan, Liu, Bryant, Berle and Liddell2019) tend to be similar or slightly elevated relative to Western Educated Industrialised Rich Democratic (WEIRD) populations and migrants (3–15.9% and 3.4–16.1%, respectively) (Nock et al., Reference Nock, Borges, Bromet, Cha, Kessler and Lee2008; Amiri, Reference Amiri2020). Similarly, population-registry studies have documented 90:100 000 suicide attempts and 11: 100 000 suicides per annum among resettled refugees in Sweden relative to 105: 100 000 suicide attempts and 20: 100 000 suicides among Swedish host population (Hollander et al., Reference Hollander, Pitman, Sjöqvist, Lewis, Magnusson, Kirkbride and Dalman2020; Amin et al., Reference Amin, Helgesson, Runeson, Tinghög, Mehlum, Qin, Holmes and Mittendorfer-Rutz2021).
In contrast, very little is known about suicidality among asylum seekers in unstable, often temporary, post-displacement or humanitarian settings, without recognised residential status, and most typically in low- or middle-income countries. This is critical, because it is in these higher-risk post-displacement contexts where ~85% of forcibly displaced people (FDP) seeking asylum currently reside (Guterres and Spiegel, Reference Guterres and Spiegel2012; UNHCR 2020). It is this increasingly common post-displacement context – characterised by a multitude of post-migration living stressors, barriers to trauma recovery and thereby elevated risk for stress- and trauma-related mental health problems – where suicidal ideation and related forms of suicidality among refugees and asylum seekers may be most likely elevated (Vijayakumar, Reference Vijayakumar2016; Nickerson et al., Reference Nickerson, Byrow, O'Donnell, Mau, McMahon, Pajak, Li, Hamilton, Minihan, Liu, Bryant, Berle and Liddell2019). Thus, epidemiologic study of suicidal ideation among asylum seekers in these high-risk post-displacement settings is urgently needed (Aichberger, Reference Aichberger, Van Bergen, Heredia Montesinos and Schouler-Ocak2014; Colucci et al., Reference Colucci, Too and Minas2017; Ventevogel et al., Reference Ventevogel, Ryan, Kahi and Kane2019; Haroz et al., Reference Haroz, Decker, Lee, Bolton, Spiegel and Ventevogel2020).
Consistent with these concerns, initial studies in convenience and community samples of asylum seekers without formal visa status, even in stable, high-income post-migration re-settlement settings, have documented worrying levels of suicidal ideation. Point-prevalence suicidal ideation of 33.9% was observed among refugees in Swedish asylum accommodations and 39.5% among refugees in Australia (Leiler et al., Reference Leiler, Hollifield, Wasteson and Bjärtå2019; Nickerson et al., Reference Nickerson, Byrow, O'Donnell, Mau, McMahon, Pajak, Li, Hamilton, Minihan, Liu, Bryant, Berle and Liddell2019). Likewise, studies of convenience samples of FDPs residing in refugee camps and among internally displaced people are also concerning – point-prevalence of suicidal ideation of 27.3–29.2% was observed in refugee camps in Nigeria and Uganda, 32.8% among Afghan refugee mothers in a refugee camp in Pakistan and a startling 62% among Rohingya mothers in refugee camps in Bangladesh (Rahman and Hafeez, Reference Rahman and Hafeez2003; Ssenyonga et al., Reference Ssenyonga, Owens and Kani Olema2013; Akinyemi et al., Reference Akinyemi, Atilola and Soyannwo2015; Tay et al., Reference Tay, Riley, Islam, Welton-Mitchell, Duchesne, Waters, Varner, Moussa, Mahmudul Alam, Elshazly, Silove and Ventevogel2019).
A related set of studies in clinical samples of refugees, even in stable post-migration settings, indicate that point-prevalence rates of suicidal ideation are markedly elevated and range from 27.8% in asylum seekers in a psychiatric clinic in Switzerland, 29.2% among treatment-seeking refugee survivors of torture in the USA, to 54.1% among refugees in an outpatient clinic in Germany (Lerner et al., Reference Lerner, Bonanno, Keatley, Joscelyne and Keller2016; Belz et al., Reference Belz, Belz, Özkan and Graef-Calliess2017; Premand et al., Reference Premand, Baeriswyl-Cottin, Gex-Fabry, Hiller, Framorando, Eytan, Giannakopoulos and Bartolomei2018). These select clinical sample findings are important in that they illustrate the theorised role of stress- and trauma-related mental health problems post-displacement for suicidal ideation (Franklin et al., Reference Franklin, Ribeiro, Fox, Bentley, Kleiman, Huang, Musacchio, Jaroszewski, Chang and Nock2017).
Despite the scale and urgency of this public health crisis, only a small number of intervention studies to prevent incidence or reduce pre-existing suicidal ideation among FDPs have been tested to date. A recent systematic review documented that interventions have demonstrated no, insufficient or only partial evidence of efficacy with respect to reduction or prevention of suicidality among FDPs (Vijayakumar, Reference Vijayakumar2016; Singla et al., Reference Singla, Kohrt, Murray, Anand, Chorpita and Patel2017; Haroz et al., Reference Haroz, Decker, Lee, Bolton, Spiegel and Ventevogel2020). Furthermore, although there are a number of additional, stress- and trauma-related mental health interventions for refugees and asylum-seekers, their potential therapeutic efficacy for suicidality has yet to be tested (Barbui et al., Reference Barbui, Purgato, Abdulmalik, Acarturk, Eaton, Gastaldon, Gureje, Hanlon, Jordans, Lund, Nosè, Ostuzzi, Papola, Tedeschi, Tol, Turrini, Patel and Thornicroft2020; Tol et al., Reference Tol, Leku, Lakin, Carswell, Augustinavicius, Adaku, Au, Brown, Bryant, Garcia-Moreno, Musci, Ventevogel, White and van Ommeren2020).
We focus on one such promising approach – mindfulness-based interventions (MBI) (Tol et al., Reference Tol, Leku, Lakin, Carswell, Augustinavicius, Adaku, Au, Brown, Bryant, Garcia-Moreno, Musci, Ventevogel, White and van Ommeren2020; Aizik-Reebs et al., Reference Aizik-Reebs, Yuval, Hadash, Gebremariam and Bernstein2021) and their potential to prevent the onset of, or treat and thereby reduce pre-existing levels of, suicidality among forcibly displaced populations. First, MBIs or interventions with elements of mindfulness practices show promising therapeutic efficacy for stress- and trauma-related mental health problems among refugees and asylum seekers (Tol et al., Reference Tol, Leku, Lakin, Carswell, Augustinavicius, Adaku, Au, Brown, Bryant, Garcia-Moreno, Musci, Ventevogel, White and van Ommeren2020; Aizik-Reebs et al., Reference Aizik-Reebs, Yuval, Hadash, Gebremariam and Bernstein2021). A MBI, specifically developed to promote mental health among diverse forcibly displaced populations, Mindfulness-Based Trauma Recovery for Refugees (MBTR-R), has demonstrated randomised waitlist-controlled evidence of its efficacy to significantly improve rates and symptom severity of PTSD, depression, anxiety and their multi-morbidity among traumatised East African asylum seekers (Aizik-Reebs et al., Reference Aizik-Reebs, Yuval, Hadash, Gebremariam and Bernstein2021). Second, Mindfulness-Based Cognitive Therapy in WEIRD populations has demonstrated robust depression relapse effects important to prevention of suicidality (Mann et al., Reference Mann, Apter, Bertolote, Beautrais, Currier, Haas, Hegerl, Lonnqvist, Malone, Marusic, Mehlum, Patton, Phillips, Rutz, Rihmer, Schmidtke, Shaffer, Silverman, Takahashi, Varnik, Wasserman, Yip and Hendin2005; Piet and Hougaard, Reference Piet and Hougaard2011), dissociation between depressive symptoms and suicidal cognitions (Barnhofer et al., Reference Barnhofer, Crane, Brennan, Duggan, Crane, Eames, Radford, Silverton, Fennell and Williams2015), and reduced suicidal ideation among patients with residual depressive symptoms (Forkmann et al., Reference Forkmann, Wichers, Geschwind, Peeters, van Os, Mainz and Collip2014). Third, MBIs may be well-suited to some of the implementation challenges facing refugee mental health intervention and suicide prevention efforts in unstable post-migration settings (Patel et al., Reference Patel, Saxena, Lund, Thornicroft, Baingana, Bolton, Chisholm, Collins, Cooper and Eaton2018; UNHCR 2020; Aizik-Reebs et al., Reference Aizik-Reebs, Yuval, Hadash, Gebremariam and Bernstein2021). Yet, despite the promising initial evidence of safety, efficacy and feasibility of MBIs, to the best of our knowledge, no study to date has tested the effects of a MBI to prevent incidence or reduce suicidal ideation among FDPs broadly, let alone in unstable post-migration settings specifically where risk for suicidality is most likely elevated.
Aims
In study 1, we estimated the point-prevalence and severity of, as well as candidate risk markers (e.g. post-migration stressors, trauma history exposure) and factors (e.g. trauma- and stress-related mental health symptoms) for, suicidal ideation among an East African community sample of asylum seekers without recognised residential status, residing in an unstable, urban, post-displacement setting in the Middle East (Israel) (N = 355, 31.8% female). In study 2, we, first, tested the prospective stability of suicidal ideation among those with and without current suicidal ideation; and whether candidate risk markers and factors at baseline prospectively predict suicidal ideation incidence/onset among a community sample of Eritrean asylum seekers (N = 158, 46% female). Second, we tested whether, relative to a wait-list control, a mindfulness- and compassion-based intervention tailored to FDPs (Aizik-Reebs et al., Reference Aizik-Reebs, Yuval, Hadash, Gebremariam and Bernstein2021) could help prevent the incidence or onset of suicidal ideation among asylum seekers without current suicidal ideation; as well as treat and thereby reduce suicidal ideation severity among asylum seekers with current suicidal ideation. Finally, we tested whether the expected prevention and intervention effects of the mindfulness- and compassion-based intervention for suicidal ideation onset and severity were mediated by therapeutic effects of MBTR-R on stress- and trauma-related mental health outcomes. See Supplementary Material for more details on our rationale to focus the current study on suicidal ideation.
Study 1 method
Participants
Three-hundred-fifty-five East African asylum seekers from Eritrea and Sudan (M (s.d.)age = 35.15 (8.29) years) who sought refuge in Israel were recruited from the community between August 2013 to May 2019 in Tel Aviv, Israel. The sample included 116 Sudanese participants who completed assessments in Arabic and 239 Eritrean participants who completed assessments in Tigrinya. On average, participants had lived in Israel for 4 years (M (s.d.)post-displacement time-in-Israel = 3.91 (3.39)). Currently, in Israel, less than 0.5% of the asylum requests are recognised (Orr and Ajzenstadt, Reference Orr and Ajzenstadt2020) and none of the participants had recognised refugee status. They have a ‘group protection’ status which functionally entitles them only the temporary right not to be deported.
Procedure
Participants were recruited directly from the community of asylum seekers in Israel, via flyers, local non-governmental and municipal organisations. See Supplementary Material for more details on participants and community recruitment.
Measures
Measures were translated and back-translated to Tigrinya and Arabic and psychometrically evaluated and validated for this study or in earlier research (Badri et al., Reference Badri, Crutzen and Van den Borne2012; Reebs et al., Reference Reebs, Yuval and Bernstein2017; Yuval and Bernstein, Reference Yuval and Bernstein2017; Yuval et al., Reference Yuval, Aizik-Reebs, Lurie, Demoz and Bernstein2021). All measures were pilot-tested among Sudanese (Arabic) and Eritrean (Tigrinya) asylum seekers and revised, in an iterative process, which included cognitive interviewing with translators and asylum seekers to ensure linguistic and socio-cultural meaning (Sartorius and Kuyken, Reference Sartorius and Kuyken1994; Miller and Fernando, Reference Miller and Fernando2008).
The Harvard Trauma Questionnaire (HTQ; Mollica et al., Reference Mollica, Caspi-Yavin, Bollini, Truong, Tor and Lavelle1992) was used to measure traumatic stress exposure and PTSD symptoms. The Brief Patient Health Questionnaire (PHQ-9; Spitzer et al., Reference Spitzer, Kroenke and Williams1999) was used to measure suicidal ideation and depression symptoms. See Discussion section for expanded discussion and rationale for this measurement approach to suicidal ideation in this population and post-displacement setting. The Beck Anxiety Inventory (BAI; Beck et al., Reference Beck, Epstein, Brown and Steer1988; Norman et al., Reference Norman, Cissell, Means-Christensen and Stein2006) was used to measure levels of anxiety symptoms. Using the categorical (diagnostic) symptom status for PTSD, depression and anxiety, we computed a comorbidity index (0 = no psychiatric symptomatology, 1 = uni-morbid or diagnostic symptom levels in one condition, 2 = co-morbid or diagnostic symptom levels in two conditions, 3 = multi-morbid or diagnostic symptom levels in all three conditions). Finally, the Post-Migration Living Difficulties Scale (Silove et al., Reference Silove, Sinnerbrink, Field, Manicavasagar and Steel1997) was used to measure current post-migration stressors. See Supplementary Material for detailed information on measures and scoring.
Results
First, point-prevalence of suicidal ideation was 31.0% (n = 108). Among the sub-sample endorsing current suicidal ideation (n = 108, M (s.d.) = 1.63 (0.82)), 58.3% (n = 63) reported suicidal thoughts several days/week, 20.4% (n = 22) suicidal thoughts more than half the days/week and 21.3% (n = 23) reported suicidal thoughts nearly every day. Men (28.8%, M (s.d.) = 0.45 (0.84)) and women (35.7%, M (s.d.) = 0.62 (0.96)) did not report different rates or severity of suicidal ideation (t(346) = −1.61, p = 0.11).
Second, in multiple regression analyses, post-migration stressor severity and trauma history exposure, together, explained a significant although relatively small proportion of variance in suicidal ideation severity. Whereas post-migration stressor severity accounted for unique variance in suicidal ideation severity, trauma history exposure was not uniquely associated with suicidal ideation beyond post-migration stressor severity. Findings did not differ between men and women. See Table 1.
Table 1. Linear regression of risk markers and factors predicting suicidal ideation severity (study 1 (N = 355))
See Fig. 1 for rates of suicidal ideation as a function of degree of multi-morbidity. In logistic regression analyses, participants with v. without PTSD, with v. without depression, as well as with v. without elevated anxiety symptoms, were significantly more likely to report suicidal ideation. Participants exhibiting greater levels of multi-morbidity of PTSD, depression and anxiety (0 v. 1 v. 2 v. 3 elevated syndromes) were also significantly more likely to report suicidal ideation (χ 2(2) = 107.05, p = 0.000, B (s.e.) = 1.26, OR = 3.54, 95% CI [2.66–4.72]). In linear regressions, we found that, PTSD, depression and anxiety each explained a large and significant proportion of variance in suicidal ideation severity. Likewise, degree of multi-morbidity explained a large and significant proportion of variance in suicidal ideation severity. These associations did not differ between men and women. See Table 1.
Fig. 1. Rates of suicidal ideation by level of multi-morbidity. Note: Number and percentage of participants endorsing suicidal ideation as a function of level of multi-morbidity.
Study 2 method
Participants
This study was a secondary analysis of a single-site randomised control trial examining MBTR-R v. a waitlist control in a community sample of 158 (46% female) unrecognised and traumatised Eritrean asylum seekers residing in a high-risk, unstable urban post-migration setting in the Middle East (Israel). As noted, study 2 participants are a sub-sample of the study 1 sample. See Supplementary Material for more information on sampling and participants.
Procedure
Following assessment for eligibility to participate in the study through a phone screening, consent and randomisation to condition (see Consort Diagram in Supplementary Material), participants completed the pre-intervention assessments. Following the 9-week intervention or identical waitlist-control period, participants completed a post-intervention assessment. Participants randomised to MBTR-R also completed a follow-up assessment 5 weeks after the post-intervention assessment. Waitlist-control participants only completed the 9-week post-waitlist assessment – to ensure that we did not unnecessarily withhold treatment for asylum seekers in the waitlist-control condition (Gold et al., Reference Gold, Enck, Hasselmann, Friede, Hegerl, Mohr and Otte2017). See Supplementary Material for more details on the MBTR-R intervention and waitlist-control condition. See study 1 and Supplementary Material for details on measures.
Results
Aim I: prospective stability of suicidal ideation
First, among participants in the waitlist-control with suicidal ideation at baseline (n = 26), 86.4% (n = 19) still endorsed suicidal ideation at post-waitlist assessment. Second, among participants in the waitlist control without suicidal ideation at baseline (n = 36), 23.1% (n = 6) endorsed suicidal ideation at post-waitlist assessment. Thus, 9-week prospective stability of suicidal ideation severity among participants in the waitlist-control from baseline (M (s.d.) = 0.66 (0.87), n = 48) to post-waitlist assessment (M (s.d.) = 0.81 (0.96), n = 48) was moderate (κ = 0.537, 95% CI [0.351–0.723], p < 0.000) (Altman, Reference Altman1990).
Aim II: prospective prediction of the onset and severity of suicidal ideation
Second, we conducted multi-level models in R (‘lmerTest’; Kuznetsova et al., Reference Kuznetsova, Brockhoff and Christensen2017) to test prospective prediction of the onset and severity of suicidal ideation at 9-week post-waitlist assessment. Among participants in the waitlist-control without suicidal ideation at baseline (n = 35), post-migration stressors and trauma history exposure, together, did not prospectively predict suicidal ideation onset and severity at post-waitlist assessment (β = 0.54, s.e. = 0.53, t = 1.03, p = 0.31). However, PTSD (β = 0.64, s.e. = 0.16, t = 4.03, p < 0.001), depression (β = 0.08, s.e. = 0.02, t = 4.72, p < 0.001) and anxiety (β = 0.43, s.e. = 0.12, t = 3.55, p < 0.001), each prospectively predicted suicidal ideation onset and severity at post-waitlist assessment. Likewise, degree of multi-morbidity at baseline prospectively predicted suicidal ideation onset and severity at post-waitlist assessment (β = 0.39, s.e. = 0.10, t = 3.96, p < 0.001).
Aim III: prevention effects of MBTR-R on suicidal ideation
Third, we conducted multi-level models in R to prospectively predict the onset and level of suicidal ideation severity at post-intervention assessment and 5-week follow-up. Among all participants without suicidal ideation at baseline (n = 109), MBTR-R, relative to waitlist-control, prevented the onset and severity of suicidal ideation at post-intervention (model R 2 = 0.19, β = −0.39, s.e. = 0.18, t = −2.10, p = 0.03) and follow-up (model R 2 = 0.23, β = −0.21, s.e. = 0.09, t = −2.37, p = 0.02). Specifically, whereas 23.1% of waitlist-controls endorsed suicidal ideation at post-waitlist assessment, 15.6% of MBTR-R participants endorsed suicidal ideation at post-intervention assessment and 9.8% at follow-up. The observed preventive effect of MBTR-R on suicidality onset and severity did not differ between men and women.
Aim IV: intervention effects of MBTR-R on suicidal ideation
Fourth, we conducted multi-level models in R to test the intervention effects of MBTR-R at post-intervention assessment and 5-week follow-up. Among all participants endorsing suicidal ideation at baseline (n = 46), MBTR-R, relative to waitlist-control, was not associated with lower levels of suicidal ideation severity at post-intervention assessment (model R 2 = 0.18, β = −0.32, s.e. = 0.25, t = −1.26, p = 0.21) or follow-up (model R 2 = 0.20, β = −0.23, s.e. = 0.14, t = −1.63, p = 0.11). Specifically, whereas 86.4% of waitlist-controls still endorsed suicidal ideation at post-waitlist assessment, 82.4% of MBTR-R participants still endorsed suicidal ideation at post-intervention assessment and 80% at follow-up. Notably, there was no treatment effect of MBTR-R on suicidal ideation severity among either men or women.
Aim V: does trauma recovery mediate prevention effect of MBTR-R on suicidal ideation?
We used a multi-level accelerated boot-strapped cross-product test of mediation in R (‘mediation’; Tingley et al., Reference Tingley, Yamamoto, Hirose, Keele and Imai2014) to test whether change from pre-to-post intervention in PTSD, depression and anxiety symptom severity mediated the effect of MBTR-R, relative to wait-list, on suicidal ideation onset and severity. We used restricted maximum likelihood to account for missing observations. Analyses were conducted among the Full Case Complete Intent-To-Treat (ITT) sample (see CONSORT in SM). See Table 2 for mediation model pathways. Among participants without suicidal ideation at baseline (n = 109), change in PTSD, depression, anxiety and level of multi-morbidity – from baseline to post-intervention assessment – each significantly mediated the observed preventive effect of MBTR-R, relative to waitlist-control, on suicidal ideation onset and severity. Because of the null intervention effect of MBTR-R relative to wait-list control (Aim IV), a test of mediation was not conducted among participants endorsing suicidality at baseline.
Table 2. Indirect effect of trauma- and stress-related mental health outcomes on the preventive effects of MBTR-R on suicidal ideation (study 2 (N = 109))
Note: ACME, average causal mediation effect.
Discussion
There are growing concerns about a global public health crisis of suicidality among refugees and asylum seekers (Vijayakumar, Reference Vijayakumar2016; Haroz et al., Reference Haroz, Decker, Lee, Bolton, Spiegel and Ventevogel2020). Study of suicidality and suicidal ideation specifically, as well as its prevention and intervention, particularly in fast-growing high-risk unstable post-displacement settings wherein the large majority of FDPs currently reside is scarce, but much needed (Aichberger, Reference Aichberger, Van Bergen, Heredia Montesinos and Schouler-Ocak2014; Colucci et al., Reference Colucci, Too and Minas2017; Ventevogel et al., Reference Ventevogel, Ryan, Kahi and Kane2019). We, therefore, sought to estimate prevalence, associated risk factors and prospective stability of suicidal ideation. Moreover, we sought to test the capacity of a mindfulness- and compassion-based intervention program (MBTR-R; Aizik-Reebs et al., Reference Aizik-Reebs, Yuval, Hadash, Gebremariam and Bernstein2021) to prevent the onset, as well as reduce the severity of pre-existing suicidal ideation in a community sample of unrecognised East African asylum seekers in a high-risk unstable post-migration setting in the Middle East (Israel).
First, observed point-prevalence and severity of suicidal ideation in this general community sample of asylum seekers were high (31%) and comparable to previously reported estimates in refugee camps and even select clinical samples of refugees seeking mental health treatment (Rahman and Hafeez, Reference Rahman and Hafeez2003; Ssenyonga et al., Reference Ssenyonga, Owens and Kani Olema2013; Lerner et al., Reference Lerner, Bonanno, Keatley, Joscelyne and Keller2016; Premand et al., Reference Premand, Baeriswyl-Cottin, Gex-Fabry, Hiller, Framorando, Eytan, Giannakopoulos and Bartolomei2018). Observed levels of suicidal ideation are consistent with, and likely a function of, elevated post-migration stressors in this urban, unstable post-displacement context and related stress- and trauma-related mental health problems (Li et al., Reference Li, Liddell and Nickerson2016; Giacco et al., Reference Giacco, Laxhman and Priebe2018). Indeed, we observed a significant cross-sectional association between post-migration living difficulties, post-traumatic stress, depression, anxiety symptoms and their multi-morbidity with suicidal ideation. Notably, trauma exposure severity was not uniquely associated with suicidal ideation beyond post-migration stressor severity (Li et al., Reference Li, Liddell and Nickerson2016; Priebe et al., Reference Priebe, Giacco and El-Nagib2016). Furthermore, depression symptom severity and multi-morbidity prospectively predicted the onset and severity of suicidal ideation, consistent with similar findings linking psychopathology severity and suicidal ideation among WEIRD populations (Mann et al., Reference Mann, Apter, Bertolote, Beautrais, Currier, Haas, Hegerl, Lonnqvist, Malone, Marusic, Mehlum, Patton, Phillips, Rutz, Rihmer, Schmidtke, Shaffer, Silverman, Takahashi, Varnik, Wasserman, Yip and Hendin2005).
Second, mindfulness- and compassion-based training tailored to diverse FDPs (Aizik-Reebs et al., Reference Aizik-Reebs, Yuval, Hadash, Gebremariam and Bernstein2021) successfully prevented the onset of suicidal ideation. This preventive effect was mediated by therapeutic effects of the intervention on stress- and trauma-related mental health outcomes. In contrast, intervention effects of MBTR-R to reduce current levels of suicidal ideation were not observed. Thus, MBTR-R therapeutically impacted suicidal ideation by preventing its onset but not by facilitating its remission. These initial findings are consistent with earlier empirical findings documenting the preventive effects of MBCT on depression relapse and suicidal cognition in WEIRD populations (Forkmann et al., Reference Forkmann, Wichers, Geschwind, Peeters, van Os, Mainz and Collip2014; Barnhofer et al., Reference Barnhofer, Crane, Brennan, Duggan, Crane, Eames, Radford, Silverton, Fennell and Williams2015). These are novel, albeit preliminary, findings with respect to preventive effects of MBIs on suicidal ideation in the context of trauma recovery broadly and among FDPs specifically. Likewise, findings are consistent with theory and previous findings that stress- and trauma-related mental health outcomes function as malleable causal risk factors for suicidality, and as likely important therapeutic targets for suicide prevention (Mann et al., Reference Mann, Apter, Bertolote, Beautrais, Currier, Haas, Hegerl, Lonnqvist, Malone, Marusic, Mehlum, Patton, Phillips, Rutz, Rihmer, Schmidtke, Shaffer, Silverman, Takahashi, Varnik, Wasserman, Yip and Hendin2005; Vijayakumar, Reference Vijayakumar2016). These initial, albeit promising, preventive effects of MBTR-R are noteworthy given limited study or interventions to prevent suicidal ideation among FDPs (Barbui et al., Reference Barbui, Purgato, Abdulmalik, Acarturk, Eaton, Gastaldon, Gureje, Hanlon, Jordans, Lund, Nosè, Ostuzzi, Papola, Tedeschi, Tol, Turrini, Patel and Thornicroft2020; Tol et al., Reference Tol, Leku, Lakin, Carswell, Augustinavicius, Adaku, Au, Brown, Bryant, Garcia-Moreno, Musci, Ventevogel, White and van Ommeren2020). Moreover, MBIs like MBTR-R may be particularly suitable for implementation and scaling-up in high-risk post-displacement settings. They are brief, group-based, low-cost and beneficial to participants with a range of stress-related distress and personal goals (Singla et al., Reference Singla, Kohrt, Murray, Anand, Chorpita and Patel2017); and have been adapted to a variety of populations and contexts in ways that are socio-culturally sensitive to diverse backgrounds, belief systems and languages (Hinton et al., Reference Hinton, Ojserkis, Jalal, Peou and Hofmann2013; Crane et al., Reference Crane, Brewer, Feldman, Kabat-Zinn, Santorelli, Williams and Kuyken2017).
The study is also limited in a number of ways. First, the study was conducted among one community sample of East African asylum seekers residing in Israel. This sampling strategy permits robust socio-cultural adaptation of MBTR-R to this population per best-practices in global mental health interventions (Kirmayer et al., Reference Kirmayer, Gomez-Carrillo and Veissière2017; Singla et al., Reference Singla, Kohrt, Murray, Anand, Chorpita and Patel2017), and may buffer against potential threats of internal validity emerging from ad-mixing of different refugee populations (Kirmayer et al., Reference Kirmayer, Gomez-Carrillo and Veissière2017; Yuval and Bernstein, Reference Yuval and Bernstein2017; Yuval et al., Reference Yuval, Aizik-Reebs, Lurie, Demoz and Bernstein2021). Yet, it is important that future work examine whether observed findings generalise to other refugee populations and contexts. Second, asylum seekers in study 2 were screened and 21 participants were excluded based on active suicidality. Thus, observed rates of suicidal ideation likely slightly underestimate the actual prevalence of suicidal ideation in this population. Moreover, by design based on pilot testing of alternative measurement methods, measurement of suicidal ideation was limited to one item of the PHQ-9 depression questionnaire. Our pilot work and previous research in this context indicated that due to social and religious stigma of suicide in this asylum-seeker population, repeated (e.g. multiple questions related to cognitions or behaviour) and more interpersonally direct inquiry (e.g. structured interviews) is likely to bias and systematically underestimate rates of suicidal ideation (Kashyap and Joscelyne, Reference Kashyap, Joscelyne, Page and Stritzke2020). Yet, we also, a priori, recognise that although this measure of suicidal ideation may be sensitive to predict future suicidal behaviour, it is not specific and may over-estimate risk of suicide for some asylum seekers (Razykov et al., Reference Razykov, Ziegelstein, Whooley and Thombs2012; Na et al., Reference Na, Yaramala, Kim, Kim, Goes, Zandi, Vande Voort, Sutor, Croarkin and Bobo2018). From public health and clinical ethics perspectives, we believe that it is far more costly to miss caseness or severity of suicidal ideation than it is to detect suicidal ideation that may not lead to suicidal behaviour (Vijayakumar, Reference Vijayakumar2016; Jobes and Joiner, Reference Jobes and Joiner2019). We also speculate that it may be important for future studies among diverse FDPs to more systematically study epidemiologic estimates of suicidal ideation and suicidal behaviour as a function of multi-method multi-modal measurement (Hopwood and Bornstein, Reference Hopwood and Bornstein2014). Third, because of uncertainty regarding residential status of this population of asylum seekers at the time of the study (Guthmann, Reference Guthmann2018) and the logistical complexity of following asylum seeker participants in unstable high-risk post-displacement context over time (Carlsson et al., Reference Carlsson, Sonne and Silove2014; Troup et al., Reference Troup, Fuhr, Woodward, Sondorp and Roberts2021), the study tested prospective stability of suicidal ideation and intervention effects of MBTR-R over a relatively short 9-week period. Due to censored prospective observations, detecting preventive effects was systematically less likely as a result, particularly in this modest sample size. Future research should test the stability of suicidal ideation as well as maintenance of observed prevention effects of MBTR-R over a longer time period (Priebe et al., Reference Priebe, Giacco and El-Nagib2016). Finally, observed effects need to be replicated, tested in larger samples and relative to more rigorous randomised active-controlled interventions (Carlsson et al., Reference Carlsson, Sonne and Silove2014).
The present findings may have a number of implications. Findings point to the potential significance and urgency of a global public health crisis of suicidality among FDPs, particularly in high-risk unstable urban post-displacement settings wherein the majority of refugees and asylum seekers currently reside worldwide. In addition, findings illustrate the potential importance of investment in research dedicated to early detection, prevention and intervention targeting suicidal ideation among high-risk FDPs (Aichberger, Reference Aichberger, Van Bergen, Heredia Montesinos and Schouler-Ocak2014; Ventevogel et al., Reference Ventevogel, Ryan, Kahi and Kane2019). Likewise, findings may inform post-displacement municipal, state and national policy-making to impact mental health and related suicidality outcomes (WHO, 2014); Priebe et al., Reference Priebe, Giacco and El-Nagib2016). Indeed, findings indicate that mindfulness- and compassion-based training tailored to diverse FDPs may help to prevent suicidal ideation through facilitating trauma recovery.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S2045796022000579
Data
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Author contribution
Amit Bernstein, Anna Aizik-Reebs and Kim Yuval designed study 1 and conducted it together with Ido Lurie and Yikealo Beyene Kesete. Anna Aizik-Reebs, Kim Yuval and Amit Bernstein developed the MBTR-R intervention programme and manual and designed and conducted study 2. All authors contributed to manuscript writing.
Conflict of interest
None.
Financial support
This work was supported by the Israeli Science Foundation (Grant number: ISF 2046/16), the Max-Planck-Gesellschaft, Mind and Life Institute Europe, Charney Foundation.
Open access
