Conference Agenda

Migration and ethnic minority status have emerged as key social determinants of mental health, yet their role in shaping suicidal risk remains insufficiently characterised. Understanding how migration-related factors intersect with clinical severity, social vulnerability, and longitudinal risk trajectories is essential for developing equitable suicide prevention strategies.

This presentation draws on a longitudinal clinical cohort of 809 individuals assessed in Madrid (Spain) following suicidal ideation or suicide attempts, of whom 200 had a migration background. Participants were classified as non-migrants, first-generation migrants, or second-generation migrants, and further characterised by ethnicity. All individuals were assessed using smartphone-based questionnaires and Ecological Momentary Assessment (EMA), enabling high-resolution measurement of symptoms, emotions, and risk markers in real-world settings. Participants were followed prospectively for six months, allowing the examination of both baseline profiles and short-term risk trajectories.

At baseline, individuals with a migration background—particularly second-generation migrants—displayed distinctive sociodemographic and clinical characteristics compared with non-migrants, including higher levels of social disadvantage, greater exposure to stressors, and significantly elevated depressive symptom severity. Migrant participants also presented with a higher accumulation of established suicide risk markers, including prior suicide attempts and indicators of psychosocial vulnerability. Ethnic minority status further contributed to heterogeneity within migrant groups, revealing important within-group differences that are often obscured when migration is treated as a single category.

Longitudinal analyses across the six-month follow-up showed that these disparities persisted over time. Migrant participants, and especially second-generation migrants, exhibited more severe clinical trajectories, with sustained depressive symptoms and persistently elevated indicators of suicide risk captured through repeated smartphone-based assessments and EMA.

These findings highlight the importance of disaggregating migrant populations by generation and ethnicity and demonstrate the value of smartphone-based and momentary methodologies for characterising dynamic suicide risk. They underscore the need for culturally sensitive, longitudinally informed prevention and follow-up strategies within mental health services.

Background: Ecological momentary assessment (EMA) offers a powerful approach for capturing short-term fluctuations in suicide risk, yet little is known about which dynamic symptom patterns best predict the timing of suicide attempts.

Methods: We conducted a 12-month prospective cohort study of psychiatric outpatients recently evaluated for suicidal ideation or attempt across four public hospitals in Madrid, Spain. Responses were aggregated to person-level features reflecting mean levels, variability (SD), and instability (RMSSD) across domains including suicidal ideation, anger, sleep, affect, and interpersonal distress. Time to suicide attempt, verified through electronic medical records, served as the primary outcome. Prediction models were fit using a Random Survival Forest with 2,000 trees, log-rank splitting, internal imputation, and out-of-bag performance metrics (CRPS, Harrell’s C, and time-dependent AUCs). Permutation importance and minimal depth statistics identified influential predictors and interactions.

Results: Of 494 participants with usable EMA and clinical data, 36 (7.3%) attempted suicide during monitoring. Short-term predictive accuracy was modest (30-day AUC = 0.67; 90-day AUC = 0.63) and declined toward chance at longer horizons. Permutation-based variable importance indicated that anger toward others was the strongest predictor of suicide attempts, with its variability (SD; VI = .024), instability (RMSSD; VI = .020), and mean level (VI = .015) showing the highest importance values. Sleep-related disturbances were also influential, with sleep interference with daily activities (mean; VI = .011) and insomnia variability (SD; VI = .011) ranking among the top predictors. Conditional minimal depth analyses further showed a robust interaction between anger variability and sleep-interference variability (depth = 0.97).

Conclusions: In this high-risk outpatient sample, dynamic features of anger and sleep disturbance were the most informative EMA-derived predictors of suicide-attempt timing, though overall predictive performance remained limited.

Background: Suicidal thoughts and behaviors represent a major public health concern, yet many individuals at risk struggle to access timely, individualized, and evidence-based support. Digital interventions (DIs) offer a promising opportunity to widen access to psychological care; however, existing tools often lack personalization and show variable engagement and retention.

Aims and methods: MoshiMoshi is an innovative project integrating clinical research and digital technology to deliver a tailored, multi-component psychological intervention specifically designed for people experiencing suicidal ideation or a history of suicide attempts. The 7-week program is delivered via a mobile app and combines psychoeducation and exercises informed by CBT, DBT, mindfulness, ACT, and IPT. All participants complete a core module on self-destructive thoughts, while additional modules—including depression and anxiety, burnout, chronic pain, relational problems, and lifestyle difficulties—are assigned through individualized screening and clinical evaluation. A hybrid format anchors the intervention in two in-person clinical sessions dedicated to safety planning and relapse prevention, with digital modules completed autonomously between meetings.

Results: Participants were recruited through multiple phases, beginning with two pilot studies assessing feasibility and acceptability among university students. In both the Pilot 1 (n=24) and Pilot 2 (n=49), satisfaction and usability ratings were high: in Pilot 1, app satisfaction reached 7.2/10 on a VAS and usability reached 68.3/100 on System Usability Scale (SUS), while Pilot 2 showed higher satisfaction (8.4/10) and usability 80.7/100 at the SUS. Participants valued daily questions and meditation exercises, although engagement fluctuated over time, and dropout rates were consistent, reflecting those commonly reported in the literature. A randomized controlled trial is currently underway to test the intervention’s efficacy compared with treatment as usual in patients.

Conclusions: By integrating clinical expertise, personalized module selection, continuous assessments, and digital delivery, MoshiMoshi aims to enhance suicide-prevention strategies through accessible, individualized, and engaging psychological support.

Background: Ecological momentary assessment (EMA) to predict self-harm and suicidal ideation is usually performed in intensive follow-ups with several questions. However, long term follow-ups with short questionnaires per day may also help to predict these behaviors in a most naturalistic situation.

Methods: Adult outpatients were recruited in different hospitals in Spain for presenting SI and/or a suicide attempt. Using active EMA through a mobile application during a 12-month follow-up we asked for suicidal ideation, non-suicidal self-injury, the desire to self-harm, emotional suffering, interpersonal needs, sleep and eating habits. EMA ask 2-4 questions daily. Mixed models were used to test the association between variables.

Results: A total of 106 participants were analyzed. Higher within person variability were found for each EMA category (ICC = .17-.45) showing that variables were more state-like. Emotional suffering (b (SE)=0.193 (0.03), t=6.21, p<0.001) and interpersonal needs (b (SE) =0.168 (0.03), t=5.96, p=0.001) across the follow-up. The desire to self-harm predicted the occurrence of NSSI episodes during the follow-up (OR=1.026, 95% CI [1.005, 1.046]). Moreover, greater passive suicidal ideation (p < .001), emotional suffering (p = .009) and interpersonal needs (p < .001) also predicted NSSI across the follow-up.

Conclusions: Our results show that using long-term and no intensive EMA we can provide insights about suicidal ideation and self-harm. Using less intensive EMA could be beneficial as it is less invasive than intensive EMA and increases the adherence and compliance.