Economic evaluations of preventive interventions for self-harm and suicide: a systematic review Authors: Long Khanh-Dao Le, Dai Quy Le, Phuong Hong Le, Eng Joo Tan, Amanda Neil, Karl Andriessen, Jane Pirkis, Cathrine Mihalopoulos, Lennart Reifels Published: 2026 ABSTRACT Self-harm and suicide are major public health concerns and leading causes of mortality worldwide, highlighting a pressing need for policymakers to identify and implement cost-effective interventions. This systematic review (PROSPERO registration #CRD42023460339) followed the PRISMA guidelines and aimed to synthesize the available cost-effectiveness evidence for the prevention of self-harm and suicide. Systematic searches were performed in MEDLINE, Embase, PsycINFO, CINAHL, Econlit, and ProQuest to identify full economic evaluations and return-on-investment studies on preventive interventions for self-harm and suicide published up to January 15, 2026. Methodological quality was assessed using Drummond’s 10-item checklist, and findings were synthesized narratively. A total of 69 eligible studies evaluated 22 types of interventions: three universal, five selective, five indicated, and nine multi-level. Most studies were rated as high-quality (n = 61/69) and conducted in high-income countries (HICs) (n = 63/69), primarily assessing the cost-effectiveness of universal interventions like means restriction (n = 6), selective and indicated interventions like psychotherapy (n = 30), support services (n = 15), and medication (n = 5). Evidence consistently found that interventions for self-harm and suicide prevention were generally cost-effective or cost-saving. Strong evidence supported the cost-effectiveness of several universal (e.g. awareness training), selective (e.g. psychotherapy, support services), indicated (e.g. suicide risk screening, support services, psychotherapy for adults in HICs like Australia, US, Canada), and multi-level interventions. However, more economic evaluations are needed for interventions targeting older adults and children in all countries, especially in low- and middle-income countries, where evidence is lacking. Introduction Self-harm and suicide (definitions in Table 1) have been recognized by the World Health Organization (WHO) as major public health concerns and consistently remained in the top four causes of death among individuals aged 15–29 year (WHO, 2021, 2023). These behaviors impose substantial economic burdens, with estimated average annual costs of $570 billion (2022 USD) in the US (Peterson, Haileyesus, & Stone, 2024) and $5.87 billion (2022 USD) in Australia (Kinchin & Doran, 2017), largely due to diminished quality-of-life, productivity losses, and healthcare expenditure. While comparable estimates are lacking for low- and middle-income countries (LMICs), the impact is likely substantial, with over 73% of global suicides occurring in LMICs in 2021 (WHO, 2023).Table 1.Definitions of key terms Table 1. long description.Term/ExpressionDefinitionsSelf-harmA non-fatal act where an individual causes harm to themselves, with various motives that may or may not involve the intention of ending their life but survives. A suicide attempt is a type of self-harm in which an individual harms themselves with the intention to die but survives (Pirkis et al., 2022).SuicideA fatal act that an individual initiated and carried out to deliberately end their own life, resulting in death (Pirkis et al., 2022).Suicidal ideationThoughts of suicide that vary from a general sense of life being meaningless to an intense preoccupation with ending one’s own life (Pirkis et al., 2022).Universal interventionTarget the entire population without determining who might be at risk of self-harm, suicide, or suicidal ideation (Pirkis et al., 2023).Selective interventionFocus on individuals displaying risk factors (e.g. stressful life events, family history of suicide, mental illness, impulsivity) that might lead to self-harm or suicide, even though they have not yet engaged in these behaviors (Pirkis et al., 2023).Indicated interventionTarget individuals who have already engaged in self-harming or suicidal behaviors (e.g. people with previous self-harm episodes) (Pirkis et al., 2023).Multi-level interventionA combination of universal, selective, and indicated interventions delivered together (Hegerl, Heinz, O’Connor, & Reich, 2021).A two-column table listing definitions for self-harm, suicide, suicidal ideation, and four types of intervention strategies. See long description. Table 1.Definitions of key terms Table 1. long description.Term/ExpressionDefinitionsSelf-harmA non-fatal act where an individual causes harm to themselves, with various motives that may or may not involve the intention of ending their life but survives. A suicide attempt is a type of self-harm in which an individual harms themselves with the intention to die but survives (Pirkis et al., 2022).SuicideA fatal act that an individual initiated and carried out to deliberately end their own life, resulting in death (Pirkis et al., 2022).Suicidal ideationThoughts of suicide that vary from a general sense of life being meaningless to an intense preoccupation with ending one’s own life (Pirkis et al., 2022).Universal interventionTarget the entire population without determining who might be at risk of self-harm, suicide, or suicidal ideation (Pirkis et al., 2023).Selective interventionFocus on individuals displaying risk factors (e.g. stressful life events, family history of suicide, mental illness, impulsivity) that might lead to self-harm or suicide, even though they have not yet engaged in these behaviors (Pirkis et al., 2023).Indicated interventionTarget individuals who have already engaged in self-harming or suicidal behaviors (e.g. people with previous self-harm episodes) (Pirkis et al., 2023).Multi-level interventionA combination of universal, selective, and indicated interventions delivered together (Hegerl, Heinz, O’Connor, & Reich, 2021).A two-column table listing definitions for self-harm, suicide, suicidal ideation, and four types of intervention strategies. See long description. Definitions of key terms Table 1. long description. A range of interventions (definitions in Table 1) have been shown to reduce self-harm and suicide, including universal interventions (e.g. restricting access to lethal means), selective interventions (e.g. medications like clozapine for people with mental illness), indicated interventions (e.g. psychological therapies for people with previous suicidal behaviors), and multi-level interventions (e.g. a combination of universal, selective, and indicated interventions) (Hegerl et al., 2021; Laflamme et al., 2022; Pirkis et al., 2024; Zalsman et al., 2016). However, evidence on the cost-effectiveness or value-for-money of these interventions still remains unclear, underscoring the need for reviewing economic evaluations to guide resource allocation (Kernick, 2003). Earlier systematic reviews of economic evidence of self-harm and suicide preventive interventions tended to situate within interventions for mental illness and mental health promotion (Feldman et al., 2021; Ha et al., 2022; Le et al., 2021; Mihalopoulos & Chatterton, 2015; Mihalopoulos et al., 2011; Zechmeister et al., 2008). These reviews largely focused on interventions in high-income countries (HICs) and primarily covered educational and training interventions (Feldman et al., 2021; Mihalopoulos, Vos, Pirkis, & Carter, 2011; Zechmeister et al., 2008). To address these limitations, Madsen and colleagues (2018) conducted the first systematic review devoted exclusively to economic evaluations of self-harm and suicide interventions across both HICs and LMICs, identifying 25 studies. However, this review was conducted a decade ago, had limited coverage of studies published between 2003 and 2016, and likely underestimated return-on-investment (ROI) studies as ROI-related search terms were not included. Since the review by Madsen and colleagues (2018), many economic evaluations of self-harm and suicide preventive interventions, including ROI studies, have been published, highlighting the need for an updated and more comprehensive synthesis. Expanding the coverage beyond the 2003–2016 period would enhance understanding of the interventions’ development history and assess whether interventions produce consistent economic outcomes across time and contexts (Drummond et al., 2009). With these motivations, this systematic review aims to provide a comprehensive and up-to-date overview of the available economic evidence on self-harm and suicide interventions without any restrictions (e.g. target population, prevention strategy) to identify those with strong cost-effectiveness potential. Methods This systematic review conformed to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Page et al., 2021). The review protocol was registered in the PROSPERO database (#CRD42023460339). Identification and selection of studies Eligible economic evaluations were identified through searches of multiple electronic databases, namely MEDLINE, Embase, PsycINFO, CINAHL, and EconLit, from inception to January 15, 2026. Reference list checks and grey literature searches on ProQuest and Google Scholar were also conducted. The search string was developed by adapting the search terms published in previous reviews (Madsen et al., 2018; Feldman et al., 2021; Le et al., 2021). Details on search strategies are provided in Supplementary Material 1. Covidence was used to ensure that each record was independently screened by at least two reviewers at each stage of the review process. Four reviewers (AT, DL, DW, and PL) screened titles and abstracts, and two (DL, PL) reviewed the full text for appropriateness based on pre-defined inclusion and exclusion criteria (see below). High proportionate agreement was observed for both abstract (93.71%) and full-text (86.01%) reviews. A senior health economist (LL) arbitrated discrepancies among the four reviewers. Inclusion and exclusion criteria The review included full economic evaluations assessing the cost-effectiveness of any preventive intervention for self-harm, suicide, and suicidal ideation, with no restrictions on prevention strategy, intervention type, target population, comparator, follow-up time, publication time, location, or setting. Eligible study designs included cost–benefit analysis (CBA), cost-minimization analysis (CMA), cost-utility analysis (CUA), cost-effectiveness analysis (CEA), or a ROI analysis that examined both the costs and benefits of at least two alternative courses of action (e.g. policies, interventions). The review excluded partial economic analyses (i.e. studies examined costs or benefits of both courses of action), studies focused on suicide gene therapy or accidental overdose, comments, editorials, opinion pieces, letters, conference abstracts, and studies written in languages other than English. Data extraction Characteristics of the included studies were extracted and summarized in a table format adapted from prior systematic reviews (Le et al., 2021; Mihalopoulos & Chatterton, 2015). To ensure comparability, all costs were converted into 2022 US dollars using the purchasing power parity approach sourced from the International Monetary Fund – World Economic Outlook database (Shemilt, James, & Marcello, 2010). If the year of pricing was not specified, it was assumed to be 2 years prior to the year of publication. If the analysis perspective was not reported or unclear, it was inferred based on the cost and outcome measures. If the study did not calculate an incremental cost-effectiveness ratio (ICER), net monetary benefit (NMB), benefit-to-cost ratio (BCR), or ROI ratio, the extracted data on costs and outcomes were used to compute these ratios. Data synthesis Interventions were first grouped into four strategies – universal, selective, indicated, and multi-level, and further classified by specific types like means restriction, medication, or psychological therapies (definitions in Supplementary Material 2). Given the substantial heterogeneity in intervention characteristics and economic evaluation methods, a meta-analysis was not conducted. Instead, a structured narrative synthesis was undertaken to evaluate and compare economic evidence across diverse interventions and contexts. Additionally, a dominance ranking framework derived from the Joanna Briggs Institute’s systematic review guidelines was employed to further categorize interventions through three color codes: red for ‘unfavored’ (higher costs and lower effectiveness), green for ‘favored’ (better outcomes at lower costs), and yellow for ‘unclear’ cost-effectiveness (e.g. more [less] effective and more [less] costly than the control) (Gomersall et al., 2015). For studies comparing multiple interventions against a control, each intervention’s cost-effectiveness was assessed separately. If conflicting cost-effectiveness outcomes appeared across different perspectives and/or outcome measures within a study, the intervention’s cost-effectiveness in that study was classified as unclear. If conflicting findings on cost-effectiveness were found among studies evaluating the same intervention type, the intervention type was concluded to have inconsistent results overall. Although there is no consensus on cost-effectiveness thresholds for commonly used outcome measures such as costs per QALY gained or DALY averted (Pichon-Riviere et al., 2023; Sun et al., 2023), the threshold of 50,000 USD per QALY gained or DALY averted has been widely adopted (Neumann, Cohen, & Weinstein, 2014), aligning with the recommendations by decision makers in Australia (Productivity Commission, 2024) and in the UK (NICE, 2013). Accordingly, this review applied the above threshold to ensure consistency in cost-effectiveness judgments while also reporting the specific thresholds within the included studies in the data extraction tables. For studies employing less frequently used cost-effectiveness measures (i.e. cost per suicide case averted), judgments were based on the thresholds specified within each study. Methodology quality assessment Drummond’s 10-item checklist was used to evaluate the methodological quality of the included studies (Drummond et al., 2015). Following prior publications (Madsen et al., 2018; Le et al., 2025), each item was scored based on whether it met (1 point), partially met (0.5 point), or did not meet (0 point) the criteria. A total score was used to classify study quality as high (8–10), average (5–7), or poor (0–4) (Le et al., 2025 5; Mihalopoulos et al., 2011). Two reviewers (PL and DL) assessed the quality of each study independently (proportionate agreement of 81.32%). Disagreements were resolved by discussion or arbitration by a third reviewer (LL). Results The systematic searches identified a total of 12,381 records. After removing duplicates and screening titles, abstracts, and full-texts, 69 studies were eligible, including 30 studies identified in previous reviews (Madsen et al., 2018; Feldman et al., 2021; Ha et al., 2022; Le et al., 2021). No additional studies were identified through reference searches of included articles. Further details are presented in the PRISMA flow diagram (Figure 1).Figure 1.PRISMA flow diagram. Figure 1. long description.A two-track flowchart showing study selection from databases and other methods, with sequential screening and exclusion steps leading to 69 studies included. See long description. PRISMA flow diagram. Figure 1. long description. Characteristics of included studies This review included 69 studies evaluating the cost-effectiveness of three types of universal, five types of selective, five types of indicated, and nine multi-level interventions for self-harm, suicide, and suicidal ideation prevention (Table 2). Most studies (n = 59) evaluated one type of intervention, with means restriction, support services, and psychotherapy were the most frequently evaluated universal, selective, and indicated interventions, respectively. Nearly all studies (n = 64) compared the interventions with no intervention or treatment-as-usual (TAU). Most studies were conducted in HICs, with only three from LMICs. The evaluation methods included CUA (n = 38), CEA (n = 37), ROI (n = 15), and CBA (n = 8). Over half of the studies (n = 39) were model-based with longer time horizons (typically 5 years to a lifetime), while the remaining were mostly trial-based with shorter time horizons (typically under 1 year). Only 19 studies targeted children and adolescents, and none focused on older adults.Table 2.Characteristics of included studies Table 2. long description.Lead author (year) countryDescription of interventionsDescription of comparatorsTypes of suicide targetedTarget population (age, %male)Evaluation methodStudy designPerspective1. Universal interventionsBandara (2022), AustraliaMeans restriction: Barriers installed at bridge and cliff sitesNo InterventionSuicidePeople who attempt suicide at bridges and cliffsROI, CEAModel (Decision tree)Partial societalDamerow (2020), Sri LankaMeans restriction: Shop-based gatekeeper training program (n = 535 shops): delivered at local pesticide shops in 2-hour sessions by field staff membersNo programSuicidePeople attempt suicide via pesticide poisoningCEAModel (Decision tree)Government perspectiveLaw (2011), Hong KongMeans restriction: Railway platform screen doors (PSDs) installation at the underground railway tracksNo interventionSelf-harmGeneral populationCUAQuasi-experimentalSocietal perspectiveLee (2021), Global (14 countries)Means restriction: Banning highly hazardous pesticidesNo interventionSuicideGeneral populationCUAModel (Markov)Healthcare systemWhitmer (2013), USMeans restriction: Suicide barrier on the Golden Gate BridgeNo barrierSuicidePeople who attempt suicide at Golden Gate BridgeCBA, CUARetrospective studyPartial societal perspective (assumed)Jackson-Morris ( 2024), UKAwareness training: Universal school-based training (5-h in-person sessions over 4 weeks by healthcare workers)No interventionSelf-harmAll adolescents enrolled in a school (aged 10–19)ROI, CUAModel (Markov)Partial societal perspective (assumed)Kinchin (2020), AustraliaAwareness training: Universal suicide prevention training (safeTALK) 3-hour workshop using presentations, videos, discussion, and questionsNo interventionSelf-harmStudents aged 15–16 yearsROIModel (Markov)Health and justice system & Societal perspectivesSari (2008), USAwareness training: General suicide education: 5 lessons to provide college students with suicide facts, warning signs, and seeking helpNo interventionSuicideYoung adults aged 18–24 yearsROIModel (Markov assumed)Societal perspectiveFlego (2022), AustraliaGuidelines/Policies: Mindframe – Media guidelines for journalists as a universal suicide prevention strategyNo interventionSuicideGeneral populationROI, CEA, CUAModel (Decision tree)Societal perspective2. Selective interventionsComans (2013), AustraliaSupport services: Standby Response Service: 24-hour multiple-center crisis response telephone number & in-person outreach (n = 90)Usual care: (n = 670)SuicidePeople bereaved by suicideCUAModel (Markov)Societal perspectivePil (2013), BelgiumSupport services: Chat helpline for suicide preventionTelephone helpline for suicide preventionNo interventionSelf-harm & SuicideUnique users of the suicide helpline, who called/chatted about their personal problemsCUAModel (Markov)Societal perspectiveRoss (2021), USSupport services: An active contact and follow-up (ACF) interventionNo interventionSelf-harm & SuicidePrimary care patientsCEA, CUAModel (Decision analytic)Healthcare system and Societal perspectiveSari (2008), USSupport services: Peer support group program: exploration of personal issues, build leadership & empathy, promote social skillsNo interventionSuicideYoung adults aged 18–24 years who may be at risk of suicideROIModel (Markov assumed)Societal perspectiveWilson-Barthes ( 2021), KenyaSupport services: Charitable Children’s Institutions: institutional care for orphanagesFamily-based careSuicidal ideationOrphaned and separated children and adolescents (OSCA)CEA, CUAProspective cohort study (n = 1,680)Societal perspectiveGray ( 2011), USPsychotherapy: Tailored and intensive family-centered in-home services (80–100 hours in 3 months); psychiatric treatment in mental health clinic (Pilot n = 22; At-scale n = 192)TAU (Pilot n = 22, At-scale n = 611)Suicidal ideationMentally ill teenagers (13–16 years old) at high risk for suicide within the juvenile court systemCEARCT for Pilot n = 44; Quasi-experimentation (At-scale n = 803)Societal perspective (assumed)Ross (2021), USPsychotherapy: A cognitive behavioral therapy (CBT) interventionNo interventionSelf-harm & SuicidePrimary care patientsCEA, CUAModel (Decision analytic)Healthcare system and Societal perspectiveTebbett-Mock ( 2020), USPsychotherapy: Dialectical behavior therapy (DBT): 9 weekly group training sessions for skill acquisition, daily intensive psychotherapy sessions in therapeutic groups (n = 425)TAUSelf-harmAdolescents (12–17 years) admitted to a coeducational, inpatient acute-careCEARetrospective cohort study (n = 801)Healthcare assumedZaloshnja (2003), USPsychotherapy: Suicide prevention & intervention program: mental health technician support, 2 weekly psychologist sessions, monthly consultations with psychiatristNo interventionSuicideNative American adolescents aged 15–19 years in communities with high risk of suicideCUA, CBAModel (Unclear)Societal perspectiveDazelle (2026) the USMedication: Oral iron therapy; Intravenous iron therapy (intravenous iron sucrose & ferric carboxymaltose)Do nothingSuicidePregnant women (M age = 28.5) with anemiaCEA, CUAModel (Markov)Healthcare & societal perspectivesDuggan (2003), UKMedication:ClozapineConventional neuroleptic therapySuicidePatients with severe mental illness & treatment-resistant schizophreniaCEA, ROIModel (Markov)Healthcare system (NHS assumed)Freemantle (1994), UKMedication:Selective serotonin reuptake inhibitors or newer tricyclic and related antidepressantsOlder tricyclicsSuicidePeople with depression & take antidepressantsCEAModel (Unclear)Healthcare perspective assumedGoren (2016), USMedication:Clozapine: daily dose of 600 mg, monitoring (weekly for the first 6 month and biweekly for the next 6 months) (n = 18,000)No clozapineSuicideVeterans with treatment-resistant schizophreniaCEAModel (Decision tree)Healthcare provider (Veteran Health Admin)Groessl (2018), USMedication:Pharmacogenetic test (IDGx): buccal cell swabs with results used to guide medication selectionStandard of care (SOC)SuicidePatients with treatment-naïve or inadequately controlled major depressive disorderCUA, CEA, CBAModel (Markov)Societal perspectiveHughes (2023), USHealth Policies: Psychologist prescriptive Authority: state policy granting prescriptive authority to qualified psychologistsNo interventionSelf-harm & SuicidePeople with mental health issuesCBA, CEA, CUAModel (Markov)Societal perspectivePark (2016), UKCombined intervention: Early intervention: mental health services from 10 professionals over a year: medication, psychotherapy, counseling, vocational support.Standard careSuicidePeople with first-episode psychosisCEAModel (Decision tree)Broader economic (societal perspective assumed)3. Indicated interventionsAcolin (2022), USPsychotherapy: Dialectical behavioral therapy (DBT): weekly individual & group sessions & clinical supervision for therapists for 1 year (n = 120)Cognitive behavioral therapy (CBT): a brief group-based treatment (5–12 sessions) (n = 180)Self-harm & SuicideAdults aged 18–64 with previous suicide attemptsCEA, CUAModel (Markov)Healthcare systemAlvi (2022), PakistanPsychotherapy: Culturally adapted manual-assisted problem-solving training (C-MAP) plus TAU (n = 108): 6–8 sessions by a psychologist at home or healthcare settingTAU (n = 113)Suicidal ideationPatients admitted after an episode of self-harmCUARCT (n = 221)Healthcare providerBernecker (2020), USPsychotherapy: Brief cognitive behavioral therapy (BCBT): 12 individual sessions + TAUTAUSelf-harm & SuicideSoldiers with recent suicidal crises in the US ArmyCEAModel (Decision tree)Program providerBjureberg (2025) SwedenPsychotherapy: Internet-delivered emotion regulation individual therapy for adolescents (IERITA) + TAU: 11 digital modules for adolescents & 6 modules for parents over 12 weeks with support from a dedicated therapistTAU: regular mental health services (therapy sessions biweekly)Self-harmAdolescents (M age = 15, 4% male) with self-harm in the past monthCEA, CUARCT (n = 166)Healthcare & societal perspectivesBojke (2024), UKPsychotherapy: Family therapy: monthly 1.25-hour therapy session over 6 monthsTAUSelf-harmYouths (11–17 years) with self-harm at least twiceCUAModel (Quasi-Markov)HealthcareByford (1999), UKPsychotherapy: Home-based social work intervention: assessment & 4 intensive, family-centered, home-based intervention sessions by psychiatric social workers + Routine care (n = 85)Routine care (n = 77)Suicidal ideationChildren aged <16 with deliberate self-poisoningCEARCT (n = 162)Societal perspectiveByford (2003), UKPsychotherapy: Manual-assisted cognitive behavior therapy (MACT): treatment manual and up to 7 treatment sessions with a trained therapist (n = 239)TAU (n = 241)Self-harmPatients aged 16–65 with a history of recurrent deliberate self-harmCUA, CEARCT (n = 480)Societal perspective (assumed)Cottrell (2018), UKPsychotherapy: Family therapy: monthly group supervision of 2 hours, monthly therapy session of ~1.25 hours over 6 months (n = 415)TAU (n = 417)Self-harmYouths (11–17 years) with self-harm at least twiceCUAPragmatic, multi-center RCT (n = 832) + Model (Markov)Healthcare system (NHS)Denchev (2018), USPsychotherapy: Suicide-focused (CBT) (9 sessions (bi)weekly psychotherapy).TAUSelf-harm & SuicideAdults with self-harm presenting to emergency careCEAModel (Markov)Healthcare perspective assumedGreen (2011), UKPsychotherapy: Group psychotherapy: 6 weekly psychotherapy sessions (acute phase) followed by a booster phase of weekly groups as needed + routine care (n = 183)Routine care (n = 183)Self-harm & Suicidal ideationAdolescent (12–17 years) with at least 2 past episodes of self-harm in the previous yearCEARCT (n = 366)Societal perspectiveHaddock (2019), UKPsychotherapy: Cognitive behavioral suicide prevention therapy: up to 20 sessions of 1 hour over 6 months in the community following discharge. + TAU (n = 24)TAU (n = 27)Suicidal ideationIn-patients (aged 18–65) in acute psychiatric wards with suicidal thoughts/behaviorsCUARCT (n = 51)Healthcare (NHS), social care, and patient perspectivesHuntjens (2025), NetherlandsPsychotherapy: Dialectical behavioral therapy (DBT) (n = 63): weekly 45 min individual CBT, weekly 2-hour skill training group, telephone coaching, weekly 1-hour therapist consultationCAU (n = 60): weekly 45-minute sessions with a psychotherapist or social workerSuicide ideationAutistic adults with suicidal behaviors (M age = 37.4, 52% male)CEA, CUARCT (n = 123)Healthcare & societal perspectivesHaga (2018), NorwayPsychotherapy: Dialectical behavioral therapy (DBT): 19 weekly 60-minute sessions of individual therapy, 120 min sessions of skills training, family therapy, telephone coaching, outpatient treatment (n = 39)Enhanced usual care (EUC) (n = 38)Self-harmAdolescents aged 12–18 years with repeated self-harm, borderline personality disorderCEARCT (n = 77)Healthcare perspectiveKrysinska (2024), AustraliaPsychotherapy: Cognitive behavioral/problem-solving therapyNo interventionSelf-harmAdults presented to hospitals with self-harmROI, CUAModel (Markov)Partial societal perspectiveLelie (2025), BelgiumAttempted Suicide Short Intervention Program (ASSIP): 3 sessions & one optional session after 3 week post discharge from hospitalCAU: short psychiatric evaluation, 3 outpatient sessions, 2 nights in psychiatric wardSelf-harm & SuicidePatients admitted to hospital with previous suicide attempt (M age = unknown, 39% male)CUAModel (Markov)Healthcare & societal perspectivesMartínez-Alés ( 2021), SpainPsychotherapy: Problem-solving Psychotherapy program: an outpatient appointment, weekly individual psychotherapy over 2 monthsTAUSelf-harmAdults aged 18 years or above with previous suicide attemptsCEAModel (Decision tree)Societal perspectiveMavranezouli (2024), UKPsychotherapy:(1) Cognitive behavioral therapy (CBT) + TAU: individual face-to-face psycho-education, problem solving(2) Dialectical behavioral therapy (DBT): weekly individual and group therapy sessions and support (consultation)TAU: treatment provided by community mental health.Self-harm(1) Adults who have self-harm and present at hospitals(2) Children and young people who have self-harm and present at hospitalsCUAModel (Decision analytic)Healthcare perspectivesMcCutchan (2022), USPsychotherapy: The Collaborative Assessment and Management of Suicidality (CAMS): builds a therapeutic relationship and addresses suicidal thoughts through 4–11 structured weekly sessionsEnhanced treatment as usual (ETAU)Suicidal ideationSoldiers with current suicidal ideationROIRCT (n = 148)Healthcare systemPark (2018), SwitzerlandPsychotherapy: Attempted Suicide Short Intervention Program (ASSIP): 3 60-minute face-to-face sessions, then semistructured personalized letters sent 6 times over 2 yearsTAUSelf-harmIndividuals with suicide attempt and treated at hospitalCEARCT (n = 120)Healthcare systemPriebe (2012), UKPsychotherapy: Dialectical behavior therapy (DBT): weekly hour-long individual therapy sessions, 2-hour skills training group, and telephone coaching (n = 40)Treatment as Usual (TAU) (n = 40)Self-harmParticipants with personality disorder, ≥5 days of self-harm in the previous yearCEARCT (n = 80)Societal perspective assumedTubeuf (2019), UKPsychotherapy: Family therapy (FT): involved monthly sessions lasting 1.25 hour over 6 months, initially more frequent, totaling six to eight sessions. (n = 383)Treatment as usual (TAU) (n = 371)Self-harmAdolescents aged 11–17 years with previous self-harm episodeCUARCT (n = 754)HealthcareTyrer (2004), UKPsychotherapy: Manual-assisted cognitive behavioral therapy (MACT): a booklet, 7 treatment sessions with a trained therapist (n = 239)TAU (n = 241)Self-harmIndividuals with a previous episode of self-harm and personality disorderCEARCT (n = 480)Broad economic (societal perspective)Spijker (2012), NetherlandsPsychotherapy: Unguided Web-Based Self-help: six weekly modules on suicidal thoughts, intense emotions, negative thinking, and relapse prevention (n = 116)Waitlist, a website that provided information on suicidality (n = 120)Suicidal ideationAdults (over 18 years old) with mild to moderate suicidal thoughtsCEARCT (n = 236)Societal perspectivevanBentum (2025), NetherlandsPsychotherapy: Add-on module to reduce intrusive suicidal mental images + TAU (n = 46): 2–6 60-minute weekly sessions at mental healthcare center by trained therapistsCAU (n = 45): psychotherapy and/or pharmacotherapySuicide ideationAdult psychiatric outpatients with depressive symptoms and distressing suicidal intrusions (M age = unknown, 32% male)CUARCT (n = 91)Healthcare & societal perspectivesVasiliadis (2017), CanadaPsychotherapy: Increasing publicly funded access to psychotherapy: two visits with a GP, 8 sessions with a mental health professionalStatus quoSelf-harm & SuicideAdults aged 20–85 years with incident depression or self-harmCUA, ROIDiscrete event simulationHealthcare system and Societal perspectiveBorschmann (2013), UKSupport services: Joint crisis plans (JCP) developed collaboratively and finalized with care coordinators, included in treatment plans alongside TAU to enhance crisis preparedness and support (N = 46)Treatment as usual (N = 42)Self-harmAdults with borderline personality disorder & a self-harm record (Age M = 35.8, 19.3% male)CUARCT (N = 88)Health and social care sector perspectiveDenchev (2018), USSupport services:(1) postcards/ letters (bi)monthly, (2) telephone outreach (1–3 months)TAUSelf-harm & SuicideAdults with self-harm presenting to EDCEAModel (Markov)Healthcare perspective assumedGallien (2025), FranceSupport services: Brief contact intervention (BCI): post-hospital follow-up with structured crisis card, telephone calls, postcardsCAUSelf-harm & SuicideIndividuals admitted to hospital with suicide attempt (M age = 39, 37% male)ROI, CBARetrospective cohort study (N = 2,3146)Healthcare perspective (assumed)Le (2023), AustraliaSupport services: Way Back Support Service: 4 brief aftercare interventions: (a) follow-up telephone contacts, (b) care coordination, (c) safety planning, (d) brief therapyNo interventionSelf-harmIndividuals aged 15+ years admitted to hospital/ emergency after a self-harm episodeROI, CUAModel (Markov)Partial societal perspectiveLindkvist (2024), SwedenSupport services: Brief Admission by self-referral (BA) + TAU: up to 3 nights in the hospital, bring their own medication, supportive meeting with hospital staffsTAU: access to outpatient and acute psychiatric clinicsSelf-harmIndividuals with self-harm at risk of suicide (M age = 32, 12% Males)CUARCT (n = 117)Payer perspectiveMartínez-Alés ( 2021), SpainSupport services: Enhanced Contact intervention: one-time priority outpatient appointment, in-person, and telephone follow-upTAUSelf-harmAdults aged 18 years or above with previous suicide attemptsCEAModel (Decision tree)Societal perspectiveO’Connor (2017), UKSupport services: Volitional help sheet (VHS): a brief encouragement statement, a list of self-harm situations & potential solutions given within 24 hours after admissionTAUSelf-harmIndividuals aged 16+ years with self-reported self-harm and suicide attemptCEARCT (n = 1,308)Healthcare systemRichardson (2014), USSupport services: Post-discharge follow-up calls by eight call centers in collaboration with hospitals and emergency departmentsNo interventionSelf-harm & Suicidal ideationIn patients with previous self-harm or suicidal ideationROIModel (Decision tree assumed)PayerRyan (2015), the USSupport services: National Suicide Prevention Lifeline crisis center (telephone helpline): risk assessment, safety planning, referrals, emergency rescue activationNo interventionSelf-harm & SuicideCallers with self-harm and suicide ideation (38% male)CUA, CBARetrospective quasi-experimental study (N = 1,7202)Societal perspectiveStelmach (2022), Global (36 countries)Support services: Follow-up prevention in hospitals: patient education and follow-upNo interventionSelf-harm & SuicideAdolescents aged 10–19 years with previous self-harmCUA, ROIModel (Markov)Societal perspectiveAppleby (2000), UKCombined intervention: Suicide risk management training for professionals (2-hour sessions over 6 months) to provide assessment, clinical management, and problem-solving for patients at risk of suicideNo interventionSuicideSuicidal individuals in primary care, ED, mental health servicesCEAPre-Post StudyHealth and social care sector perspectiveDunlap (2019), USCombined intervention: Screening by physician plus 7 telephone sessions, self-personal safety plan, treatment information, suicide hotlineTAUSelf-harm & SuicideIndividuals with self-harm ideation or behavior upon entry to emergency careCEACluster RCT (n = 1,376)Program providerJackson-Morris ( 2024), UKCombined intervention: a 1-hour information session at a hospital, 9 brief follow-up contacts over 18 months delivered by mental health support workersNo interventionSelf-harmAdolescents with self-harm and treated at a hospitalROI, CUAModel (Markov)Partial societal perspective (assumed)Latimer (2014), CanadaCombined intervention: Rapid response team (RRT) (n = 158) included a psychiatrist, psychiatric nurse, and others offered assessment, treatment initiation (medication, CBT), and referralsUsual care (n = 128)Self-harmSuicidal adolescents aged 12 to 17 years at emergency departmentCEATrial (n = 286)Healthcare & societal perspectivesBotchway (2022), UKScreening: Oxford Mental Illness and Suicide (OxMIS): a suicide risk assessment for people with SMI (medication & clinical reviews, multi-disciplinary assessment, elective admission, ED visit)Usual careSuicidePeople with severe mental illness (SMI) admitted to the hospital due to nonfatal self-harmCUAModel (Decision tree)Healthcare system (NHS)Dunlap (2019), USScreening: Screening by ED nurses plus TAUTAUSelf-harm & SuicideIndividuals with self-harm ideation or behavior upon entry to emergency careCEACluster RCT (n = 1,376)Program providerMcDaid (2022), UKScreening: Psychosocial assessment administered upon arrival to hospital due to self-harmNo interventionSelf-harmIndividuals admitted to hospital due to self-harm (age: 15+, 41%)CUAModel (Markov)Healthcare and partial societal perspectivesde Beurs (2015), NetherlandsTraining for professionals: e-learning program -Train-the-Trainer: 1 day in-person training with e-learning module for professionals + IAU (n = 312)Usual implementation (IAU) (n = 254)Suicidal ideationPatients 18 and above with suicidal ideationCUA, CEARCT (n = 566)Societal perspective4. Multi-level interventionsAhern (2018), Global (10 EU countries)3-level School-based prevention:(1) Youth Aware of Mental Health (YAM): promote mental health knowledge(2) Question, Persuade, Refer (QPR): train gatekeepers to identify and intervene student’s risky behaviors.(3) Screening by Professionals (ProfScreen): screen, identify, and refer at-risk adolescentsControl group: only exposed to the same educational posters as in YAMSelf-harm & Suicidal ideationSchool-age children and adolescentsCEA, CUACluster-RCT (n = 11,110 students)Payer (education, health, & social care system)Crosland (2024),Australia(1) Emergency department-based suicide prevention (EDSP): Universal and secondary suicide risk screening, discharge resources, follow-up phone callsCAUSelf-harm and suicide(1) Young people (aged 25 and below) with psychological distressCUA, CEA, CBAModel (Dynamic simulation techniques)Healthcare and societal perspectivesCrosland (2024),Australia (Continued) (2) School-based suicide prevention program (SSP): universal mental health education and awareness training; indicated brief screening and provision of CBT; treatment seeking assistanceCAUSelf-harm and suicide(2) Children with psychological distressCUA, CEA, CBAModel (Dynamic simulation techniques)Healthcare and societal perspectivesDoran (2016), AustraliaMates in Construction: awareness training (1 hr), connector (4 hr), intervention skills (2 days), promotion materials, 24/7 helplineNo program assumedSelf-harm & SuicideConstruction industry workersROIModel (Decision tree)Societal perspective (assumed)Garraza (2018), USCommunity-based suicide prevention programs (GLS): gatekeeper training, prevention hotlines, education, screeningNo InterventionSelf-harmYouths aged 16–23ROI, CEAModel (Unclear)Healthcare systemKinchin (2017), AustraliaMates in Construction: awareness training (1 h), connector (4 h), intervention skills (2 days), promotion materials, 24/7 helplineNo interventionSelf-harm & SuicideAustralian workforce (mean age = 42)ROIModel (Markov assumed)Societal perspectiveLebenbaum (2020), CanadaAwareness campaigns, diagnostic training, educating primary-care physicians, & providing high-risk psychosocial interventions.No interventionSelf-harm & SuicideOntario residents aged 16+ years (M age = 46 years)CUAModel (Markov)Healthcare payer, Societal perspectivesStelmach (2022), Global (36 countries)Universal School-based prevention: gatekeeper training, mental health training for kids, screening, professionals’ referral (3 programs similar to Ahern)No interventionSelf-harm & SuicideAdolescents aged 10–19 yearsCUA, CBAModel (Markov)Societal perspectiveVasiliadis (2015), CanadaNuremberg Alliance Against Depression (NAD): community interventions for depression care and suicide prevention via universal awareness campaign, diagnostic and treatment training, and follow-up of individuals with previous self-harmNo interventionSelf-harmFirst responders and follow-up of individuals who attempted suicideCEAModel (Markov assumed)Healthcare system and Societal perspectiveLead author (year) countryTime horizonYear of pricing, discount ratesCost of interventionCost savingsHealth outcomes/benefitsResults ICERS (in 2022 USD value) thresholdQuality1. Universal interventionsBandara (2022), Australia5 years10 years2018 USD, 3%Initial installation and maintenance (cyclical and condition-based)Value of statistical lifeSuicide cases avertedROI = 1.7 (95% UI, 1.0–1.9) (5 yr) and 2.4 (95% UI, 1.5 to 2.7) (10 yr) for building barriers at bridges.ROI = 1.7 (95% UI, −0.9–3.3, ns) (5 yrs), and ROI = 2.0 (95% UI, −1.1–3.8, ns) (for 10 yr) for building at cliff sites.ICER = dominant for barriers over bridges BUT not for cliff sites.HDamerow (2020), Sri Lanka3 years2019 USD, 3%Direct personnel, travel, catering, admin, training materials, equipment, communicationFatal pesticide self-poisoning case prevented; Life-year savedCost-effective if preventing 0.23 fatal pesticide self-poisoning cases over 3 yearsThreshold: $4,860/life-year savedHLaw (2011), Hong Kong25 years2002 USD, 3%Costs of installing railway platform screen doorsPassenger time Fare revenue saved after installing Railway platform screen doorsDALYsNumber of suicidal and accidental fallsICER = $122,000 per DALY averted and decreased to $103,000 if accounting for the loss of fare revenue and passengers’ waiting timeThreshold: $117,000 per DALY avertedHLee (2021), Global (14 countries)100 years2017 I$, 3% for cost, 0% for health outcomeProgram costs (admin, training, media, regulation) and patient-level costsNoneHealthy life-years gained (HLYGs)Suicide Mortality rateICER = $110 per HLYG (95% UI 86–145) in LMICs and $280 per HLYG (95% UI 225–356) in UMHICsThreshold: $118/HLYGHWhitmer (2013), US20 years2009 USD (assumed), unknown discount rateCost of implementing suicide barrierValue of statistical lifeNumber of lives saved (Mortality reduction)DALYsCBR = 1; ICER = $6,528 per DALYThreshold: $193,000/DALY avertedLJackson-Morris (2024), UK10–80 years (lifetime)2021 GBP, 3%Salary, equipment (computer, phone call, internet access)Long-term employment and earningsDALYsICER = $4550 (10 years) & $970 (lifetime) /DALY avertedROIs = −0.4 (10 years) & 0.9 (lifetime)HKinchin (2020), Australia5 years2017–18 AUD, 5% for costsTravel and materials for local healthcare services and non-clinical groups.Ambulance, healthcare, police, investigation, funeral, productivity losses, grievingNoneHealth and justice perspective: ROI = 1.45 in Mackay; ROI = 0.19 in Queensland; ROI = 0.15 in AustraliaSocietal perspective: ROI = 31.2 in Mackay; ROI = 4.1 in Queensland; ROI = 3.3 in AustraliaHSari (2008), USLifetime (~40 years)2000 USD, 6.3%General suicide education: training & materialsDirect cost (ambulance & autopsy services); Indirect cost (earnings lost, productivity loss of family members)NoneROI: 2.03HFlego (2022), Australia5 years2018 AUD, 3%Staffing and non-staffing costs: review and update guidelines, training, support,Costs of suicide death prevented (value of a statistical life, total cost associated with a suicide death)Suicide deaths preventedQALYROI: 94 (95% CI 37–170) The intervention was dominant regarding cost per suicide prevented or QALYs gainedThreshold: ROI > 1H2. Selective interventionsComans (2013), Australia12 months 5 years2010–2011 AUD assumed, 5%Operating costs of providing intervention, community development, engagement, and training.Health costs, Productivity lost (HPQ): Absenteeism, Presenteeism, unpaid workQALY (EQ–5D)The StandBy service is dominanceWTP Threshold: $46,000/QALY gainedHPil (2013), Belgium10 years2012 EUR, 3% for cost and 1.5% for effectsCosts of salaries, transportation of personnel, stipends for volunteers, operation, and communicationAmbulance, medical examiner, ED, inpatient care, nursing home, productivity lossQALYs (EQ–5D)ICER = dominantNet societal savings of telephone and chat service: $3,600 (95%CI 2,960–4,330) and $3,460 (95%CI 2,800–4,200) in male users; $3,300 (95%CI 2,630–4,040) and $3,700 (95%CI 2,950–4,600) in female users.HRoss (2021), USLifetime2016 USD, 3%Intervention costs (No details)Healthcare, patient time, productivity lossSuicidal risk (Fatal and nonfatal suicide attempt) QALYsTo achieve ICERs below $50,000 per QALY gained, ACF needed a suicide prediction method with a sensitivity of ≥57% and specificity of ≥95%HSari (2008), USLifetime (~40 years)2000 USD, 6.3%Peer support group: housing & food for retreat, trainers’ salaries, stipend, computer, transportationDirect cost (ambulance services and autopsy services); Indirect cost (earnings lost & productivity loss of immediate family members)NoneROI: 3.71HWilson-Barthes (2021), Kenya10 years2018 USD, 3%Personnel salaries, overhead, training, education fees, capital costs (e.g. appliances, equipment)NoneReduction in suicidal ideation & intent (CDI-SF scale)ICER for institutional care: $526 [179, 959] per unit reduction in suicidal diagnosis and $5,680 per QALY gained among OSCAWTP threshold: $8,100/QALY gainedHGray (2011), US12 months2009 USD assumed, no discountTreatment costs & Cost of days spent in observation and assessment at detention centers and correctional facilitiesProductivity loss of parentsRecidivism, days in court placement, Suicidal ideation (Youth Outcome questionnaire-YOQ)Intervention is cost-saving in improving score on the youth outcome questionnaire for suicidal ideationLRoss (2021), USLifetime2016 USD, 3%Intervention costs (No details)Healthcare, patient time, and lost productivity costs associated with fatal and non-fatal suicide attemptsSuicidal risk (Fatal and nonfatal suicide attempt)QALYsTo achieve ICERs below $50,000 per QALY gained, CBT needed a suicide prediction method with a sensitivity of ≥75% and specificity o ≥ 97%HTebbett-Mock (2020), US8 months2017 USD assumed, no discountNone (Not provided)Cost of constant observation hoursIncidents of suicide attempts, self-injury; restraints; and days hospitalized, aggression patient-to-patient/staffIntervention is cost-saving: $141,000 for suicidal ideation, $90,000 for aggression, and $66,600 for self-injurious behaviors.AZaloshnja (2003), USLifetime1998 USD, 3%Costs of salaries and educationalCUA: medical and out-of-pocket costCBA: cost savings to individuals, local governments, insurance companies, employers, and the general publicQALYsBCR: 43ICER: $710 per QALY gained Threshold: $150,000/QALY gainedHDazelle (2026) the USLifetime2020 USD, 3%Costs of administering iron therapyMedical costs (blood transfusions, premature births), productivity lossesQALYAdverse outcomes averted: post-hemorrhage, hysterectomies, suicideICER: dominant, Incremental NMB = $1,141–$2,128 Ferric carboxymaltose confers the greatest overall benefitThreshold: $100,000 per QALY gainedHDuggan (2003), UK10 years2001 GBP assumed, 1.5% for outcomes & 6% for costsPatient support costsCosts of clozapine & other alternative drugsCosts of suicide: general hospital costs, inpatient costsLife-years savedClozapine in treatment-resistant schizophrenia is cost-effective.Cost-neutral scenario: $12,000 per life-year savedCost-saving scenario: $20,000,000 net annual savingAFreemantle (1994), UKLifetime1992 GBP, 5%Medication costsSuicide cases & deaths related to anti-depressants (administrative data)ICER = $54,600–$486,000 per life-year gainedSelective serotonin reuptake inhibitors are more costly and more effective than the older tricyclic.AGoren (2016), US12 months2015 USD, no discountLaboratory monitoring and follow-up visits, Medication costsDirect healthcare cost: prescriptions, inpatient & outpatient treatment, emergency services, and office-based medical visitsTreatment response, suicides, adverse drug reactions & associated mortality, hospitalization due to clozapineThe intervention resulted in cost-saving per case of suicide-related death averted.HGroessl (2018), US3 years2016 USD, 3%Cost of pharmacogenetic testDirect medical costs (medications, hospitalizations, physician visits, psycho-therapy)Indirect medical costs (lost productivity)QALYs (EQ–5D)Suicide rateIDGx is dominant for moderate to severe MDD patientsThreshold: $60,000 per QALY gainedHHughes (2023), US20 years2021 USD, 3%Personnel, designation fee, maintenance, and individual prescribing psychologistsMedical costs (ED visits and hospitalization due to suicide)Societal costs (productivity, quality of life) following the suicideSuicide attempt, suicide death, QALYsICER = $48,300 per QALY gained; $28,700 per nonfatal suicide attempt averted; 28,161 per suicide death avertedIntervention is dominant at WTP threshold of $53,000HPark (2016), UK10 years2009 GBP, 3.5%Costs of consultation with health professionals, formal and informal admissionProductivity losses, health service use, reduced quality of lifeSuicide case avertedICER = dominant (Societal cost savings of $1,815 per person per year)H3. Indicated interventionsAcolin (2022), US12 months2022 USD, no discountCost of therapist’s timeCosts of hospitalizationLife yearsQALYICER = $26,362 per QALY gainedICERs for life-years gained or attempted suicides prevented were negligible.Threshold = $50,000 per QALYHAlvi (2022), Pakistan6 months2019 USD, no discountPersonnel travel and time, transportation and reimbursement, materials, supplies, and communication costsHealth services: Inpatient & outpatient, GP & other health professional visitsSuicidal ideation (Beck Scale for Suicidal ideation)QALY (EQ–5D–3L)ICER for C-MAP = $18,410 (95% CI $8,060–$112,203) per QALY gainedThreshold: no formal WTP threshold for Pakistan, thus used WTP threshold $22,600–$34,000HBernecker (2020), USLifetime2017 USD, Federal discount rate & 3%Psychotherapy sessions, psychiatric inpatient days, & patients’ intake evaluationsMortality costs, Suicide attempt costs, retired disability pay costsSuicide attempts avertedSuicide deaths avertedCost saving = $17,700–$19,640/patient Except for the worst-case scenario,ICER of BCBT was dominant for $/suicide attempt & suicide death averted.HBjureberg (2025) Sweden1 & 3 months2022 USD, no discountTherapist’s timeHealthcare visits, medication, social support, productivity lossesSelf-harm frequency (DSHI-Y),QALY (CHU–9D)– ICERs - societal perspective: $18,677 (1 month) & $27,563 (3 month) per one adverted case; & $792,244 (1 month) & $471,800 (3 month) per QALY gained– ICERs – healthcare perspective: $15,502 (1 month) & $29,972 (3 month) per one adverted case; & $660,793 (1 month) & $361,041 (3 month) per QALY gainedWTP threshold: $84,000 per QALY gainedHBojke (2024), UK1.5 and 5 years2022 GBP (assumed), 3.5%Personnel costs, telephone contact, supervisionCosts of hospital admissionsNon-hospital cost: primary & community care, medicationQALYsICERs = $53,000/QALY gained (1.5 years) – dominated (5 years)Threshold: $43,000/QALY gainedHByford (1999), UK6 months1997–1998 GPB, no discountCosts of personnel, travel, supervision, hospital services, GP, social worker, and other healthcare and social servicesSuicidal ideation (Suicidal Ideation Scale)Family functioning (Family Assessment Device) General HealthThe intervention is as cost-effective as routine care aloneFor children with no depression, intervention is more effective in reducing suicidal ideation with no difference in costs.HByford (2003), UK12 months1999/2000 GBP, no discountProductivity losses, living expenses, hospital, community health, social services, criminal justiceRepeated episodes of deliberate self-harm (Parasuicide History Interview) HRQoL (EQ–5D)At 12 months, MACT is dominant in reducing repeat self-harm, and ICER = $156,750/QALY gained.WTP Threshold: 0-$156,750/QALY gainedHCottrell (2018), UK18 months 5 years2014 GBP, 3.5%Personnel costs, telephone contact, supervisionHealthcare costs (health, social services, in- & out-patient care) MedicationsHospitalization due to self-harm Suicidal ideation (BSS) QALY (PQ-LES-Q & GHQ–12)At 18 months (base case), FT is dominated in preventing self-harm events & ICER = $64,185/QALYAt 5 years, ICER = $34,000/QALY gainedThreshold: $35,000–$52,000 per QALY gainedHDenchev (2018), US54 weeks2014 USD assumed, no discountDirect cost of intervention deliveryCosts of hospitalization & healthcare (ED visits, inpatient & outpatient care)Self-harm and suicide incidents averted Life-years savedICER = $22,100 per life-year saved. WTP threshold: $61,000 per life-year savedAGreen (2011), UK12 months2005–6 GBP, no discountSalary of therapist, overhead, indirect therapist time (preparation), travel costTotal costs of health, social care, education, and criminal justice sector services, plus family-related costs and productivity lossesFrequency of repeated self-harm (interview), severity of subsequent self-harm (triangulated reports), suicidal ideation (SIQ)Group psychotherapy is not a cost-effective addition to the routine care deliveredHHaddock (2019), UK6 months2015–16 GBP, no discountIntervention costCommunity services, primary care, accident and emergency care, in- & out-patient careSuicide behavior & ideation (SBQ, BSSI, BHS, clinical records); QoL (QoL-Brief); QALY (EQ–5D)ICER = $637,000/QALY gained Intervention is less costly and less effective than TAU. 95CI suggests no difference in costs and QALYs for both groupsHHuntjens (2025), Netherlands12 months2023 EUR, no discountCosts of administering DBT & supervision costs for therapists providing DBTDirect healthcare costs: health consultations, hospitalizations, medication Non-healthcare costs: carer costs, transportation, productivity losses.QALYAverted suicidal ideation on SIDAS scaleICER: dominant (healthcare perspective); & $1,527 per QALY gained & $1,642 per point improved on SIDAS scale (societal perspective)WTP threshold: $61,000 per QALY gainedHHaga (2018), Norway12 months2012 EUR, no discountWages for staff, equipment, IT, house rent, telephone consultations, medication, therapist team consultations.Health services due to self-harm (inpatient treatment, emergency care, and GP consultationsSelf-harm episodes (LPC)Global functioning (CGAS)ICER: less costly & more effective with 99.5% cost-effective at thresholds: $2,000/self-harm episode avoidedHKrysinska (2024), Australia10 years2018 AUD, 3%Costs of referral pathway, delivering intervention, weekly health professionals’ supervision costCosts of hospital admission related to self-harm, suicide, and death due to other causesQALYROI = 2.3–6.0ICERs = dominant Threshold: ROI > 1, $41,100 per QALY gainedHLelie (2025), Belgium20 years2023 EUR, 3% for costs, 1.5% for benefitsASSIP start-up costsHealthcare costs: psychiatric consultation, hospitalization Productivity loss: recuperation, and economic replacement costsSuicide re-attempt or suicide death QALYICER: dominant (societal perspective), $360 per QALY gained (healthcare perspective)WTP threshold: $52,000 per QALYHMartínez-Alés (2021), Spain12 months2019 EUR, no discountCosts of TAU plus 8 psychotherapy sessions.Extra-hospital emergency care, in- & out-patient care. Indirect costs of death & disability leaveSuicide reattempt within 1 year after hospital discharge Time to reattemptICER = $10,600 per averted attempt WTP Threshold: $47,300–67,500/suicide avertedHMavranezouli (2024), UK5 years2020 GBP, 3.5%(1) Costs of individual sessions(2) Costs of individual psychotherapy, group skill training, team and individual consultationsHealthcare costs associated with self-harm: primary care, emergency department, inpatient care, medication, social careQALYs(1) ICER = $13,800/QALY gained(2) ICER = $407,000/QALY gainedThresholds: $30,000–45,500 per QALY gainedHMcCutchan (2022), US12 months2018 USD, no discountImplementation costs (training, personnel, travel, space, material, treatment costs (e.g. facilities, admin, insurance)Healthcare expenditures, medical services, crisis services, & medicationsSuicidal ideation (The Scale for Suicidal ideation-Current - SSI-C)Mean CBRs at 12-month: −13.26 (SD = 55.10) for ETAU and 1.68 (SD = 40.53) for CAMS. Mean within-subject net benefit at 12-months: -$4,300 (SD = $17,000) for ETAU and -$480 (SD = $14,900) for CAMS.HPark (2018), Switzerland24 months2015 CHF, 3.5% for costs & 1.5% for health outcomesTraining costs for hospital staffInpatient care, outpatient careSuicide attempt avoidedASSIP was dominant, with significantly fewer reattempts at lower overall cost.WTP threshold: $0 and $25,600HPriebe (2012), UK12 months2010 GBP assumed, no discountService costs (no further details provided)Costs of primary, secondary & social care: hospital, outpatient visits (GP), medication, lost workdaysSelf-harm rate Borderline personality disorderICER = $67 per one percentage point reduction in the incidence of self-harmHTubeuf (2019), UK12 months2017 GBP assumed, No discountCosts of therapies sessionsHealth service use (accident and ED visits and inpatient stays), psychotropic medicationsQALYs (EuroQoL 5 Dimensions 3 Levels (EQ–5D–3L) for the adolescent and the Health Utility Index (HUI2) for one parent)ICER = $67,500 per QALY gained (include adolescent’s health gain only)ICER = $45,400 - $68,200 per QALY gained (including adolescent’s health gain, health spill over to one parent)Threshold: $33,400–50,000/QALY gainedHTyrer (2004), UK12 months2002 GBP assumed, No discountNo detailsCost incurred by all service-providing sectors, productivity lossRepetition in self-harm Time until repetitionNo significant difference between self-harm cases in the MACT (39%) compared to TAU (46%) (P = 0.20), MAC was cost-effective (10% cheaper than TAU)HSpijker (2012), Netherlands1.5 month (6 weeks)2009 EUR, no discountCosts of intervention, visits to psychologists, website maintenance, and hostingDirect costs: medical (healthcare, prescription); non-medical (participants’ travel, loss of leisure time, informal care); Productivity lossesChange in suicidal ideation: Beck Scale for Suicidal ideation (BSS)/Suicidal ideation (BSS) Treatment responseICER was cost saving = −$55,000 per additional treatment response.WTP Threshold: 0/treatment responseHvanBentum (2025), Netherlands12 months2024 EUR, no discount assumedPsychologists’ time providing add-on moduleHealthcare costs: psychiatric healthcare, medication, out-of-pocket costs Productivity losses, travel expensesQALYICER = $160,256 (societal perspective) & $116,605 (healthcare perspective) per QALY gained WTP threshold: $93,000 per QALY gainedHVasiliadis (2017), Canada40 years2012 CAD assumed, no discount specifiedCosts of psychological therapy services for individuals received inadequate careMedical, nonmedical, and productivity costsQALYsICER = dominant (Societal cost savings: $2,700 per person (95% CI $1,340–$6,700))ROI = 2.00 (1.78–3.15)HBorschmann (2013), UK6 months2009–2010 GBP, no discountCosts of salaries, overheads, preparation & trainingCosts of health and social care services: Hospital, community health and social care, medication, criminal justice sector servicesSelf-report self-harm post 6 months, Depression and anxiety (HADS), Well-being (WEMWBS), Perceived coercion (TES), QALY (EQ–5D)JCP + TAU is dominantHDenchev (2018), US54 weeks2014 USD assumed, no discountDirect cost of intervention deliveryCosts of hospitalization & healthcare (ED visits, inpatient & outpatient care)Self-harm and suicide incidents averted Life-years savedPostcards is dominant ICER = $5,280 (Telephone) per life-year saved.WTP threshold: $61,000 per life-year savedAGallien (2025), France12 months2022 EUR, unspecified discountCosts of administering BCI (unclear cost components)Direct healthcare costs (assumed)Suicide re-attempt or suicide deathROI: 2.06NMB per patient: $331 in healthcare cost savingHLe (2023), Australia10 years2018 AUD, 3%Costs of salaries, staff training, professional development, and staff travel.Healthcare cost-savings: (a) Base case: direct costs (healthcare, police & ambulance, & counseling), indirect costs (productivity loss); (b) SA: the value of a statistical lifeQALYsROI = 1.58 (95% CI: −0.17 – 5.33)ICER = dominant (95% CI: dominant - $156,000/QALY)WTP Threshold: $41,000/QALY gainedHLindkvist (2024), Sweden1 & 5 years2020 EUR, 3%Costs of brief admission: bed occupancyCosts of psychiatric clinic visits, in-patient care, emergency department visitsQALYsAfter 1 year: ICERs = cost-saving (capacity adjusted) - $79,000 (Base case) per QALY gained.All analyses indicated health gains and substantially reduced costs in 5 years.HMartínez-Alés (2021), Spain12 months2019 EUR, no discountTAU costs plus service use costs (phone call, specialized visits).Extra-hospital emergency care, in- & out-patient care. Indirect costs of death & disability leaveSuicide reattempt within 1 year after hospital discharge Time to reattemptICER = $4,000 per averted attemptWTP Threshold: $47,300–67,500/suicide avertedHO’Connor (2017), UK6 months2014 GBP, no discountCosts of health service use & routine implementation, excluding development costs (e.g. salaries)Costs of hospital care, ambulance, treatment within EDs, hospital care following injuryself-harm event avertedVHS is dominant compared to TAU (lower costs and higher effective, although these differences are not statistically significant)WTP threshold: $0HRichardson (2014), US1 month2011 USD, no discountCosts of follow-up calls, reports, admin work, and additional treatmentMental health and substance abuse hospital readmissionsNonePatients discharged from (1) hospital: ROI = 1.76 (commercial insurance), 2.43 (Medicaid); (2) ED: ROI = 1.70 (commercial insurance), 2.05 (Medicaid)HRyan (2015), the US5 years2013 USD, unspecified discountCosts of staff, daily operations, facilities, technology, trainingEmergency department costs, productivity lossesQALYBCR = 0.27ICER = $6,540WTP threshold: $50,000 per QALY gainedHStelmach (2022), Global (36 countries)80 years (Lifetime)2019 USD, 3%Salaries, medications, laboratory tests, computer equipment, material, transportationWage returns to education, productivity gains & lossesUtility weights converted into monetary unit DALY avertedROI = 62ICER = approximately <$110/DALYHAppleby (2000), UKLifetime1998 GBP assumed, no discountMaterial development, training, materials and equipment, and attending trainingSuicide prevented:Life-year gainedICER: $243,000 per suicide prevented and $8,300 per life-year gained.LDunlap (2019), US12 months2015 USD, no discountWage & Salary, Rental costs for hospital spaceSuicide attempts avertedSuicide deaths avertedICER = $6,100 per averted suicide attempt or deathWTP threshold of $7,300HJackson-Morris (2024), UK10–80 years (lifetime)2021 GBP, 3%Salary, equipment (computer, phone call, internet access)Long-term employment and earningsDALYsICER = $10,900 (10 years) & $2,150 (lifetime) /DALY avertedROIs = 0 (10 years) & 2.9 (lifetime)HLatimer (2014), Canada6 months2011 CAD, no discountCosts of RRT intervention and psychiatrist from RRTHealthcare: ED visits, inpatient & outpatient care, medicationSocietal perspective: visits to other hospitals, out-of-pocket healthcare expendituresPatient’s level of functioning (CGAS)Severity of suicidal behavior (SSBS)Hospital perspective: ICER = $14,700 (dominant to $150,000) per point improved on the SSBSSocietal perspective: ICER = $7,800 (dominant to $184,000) per point improved on the SSBSHBotchway (2022), UK12 months2021 GBP, no discountRisk assessment administration, High risk management (e.g. additional 4 clinical reviews)Direct costs of immediate emergency service and health-related.QALYsOxMIS is dominantHDunlap (2019), US12 months2015 USD, no discountWage & Salary, Rental costs for hospital spaceSuicide attempts avertedSuicide deaths avertedICER = $3,400 per averted suicide attempt or death WTP threshold of $7,300HMcDaid (2022), UK2 years2020 GBP, 3.5%Administering psychosocial assessment (No details)Hospital costs of treating self-harm, ambulance costs, productivity lossQALYsICERs = $16,600 (healthcare perspective); $15,100 (societal perspective) per QALY gainedThresholds: $30,000–45,500 per QALY gainedHde Beurs (2015), Netherlands3 months2013 EUR assumed, no discountTraining costs & personnel costsDirect medical costs (visits to a psychiatrist or a psychologist)Indirect medical costs (costs due to sick leave & productivity losses)Suicidal ideation (BSS) QALYs (EQ–5D)Intervention is not cost-effective compared to IAU (ICER = $3,900/point gained on BSS; ICER = $254,000/QALY gained) But cost-effective for suicidal patients if societal WTP > $14,000/point gained on BSSH4. Multi-level interventionsAhern (2018), Global (10 EU countries)12 months2010 EUR, no discountTraining, material adaptation, clinical interview, implementation, photocopySuicide attempt & Severe suicidal ideation (Paykel Hierarchical Suicide Ladder)QALY (SDQ)ICERs per % reduction in suicide attempt = $134 (QPR), $52 (YAM), $78 (ProfScreen); ICERs per QALY gained = $179,187 (QPR), $69,876 (YAM), $95,191 (ProfScreen). WTP Threshold: $69,850/QALY gainedICERs per % point reduction in suicidal ideation = dominated (QPR), $68 (YAM), $152 (ProfScreen); ICERs per QALY gained = dominated (QPR), $71,660 (YAM), $161,684 (ProfScreen). WTP Threshold: $71,300/QALY gained YAM is the most cost-effectiveHCrosland (2024), Australia10 & 30 years2019–20 AUD, 5%Intervention cost: program cost, administration costHealthcare, productivity, and carer costsQALYs, Suicide deaths, and self-harm(1) Healthcare perspective: ICER = $83,000 per QALY gained, $2,800 per self-harm prevented, $76,800 per suicide death prevented; INMB = −1;Societal perspective: ICER = $80,600 per QALY gained, $74,000 per suicide death prevented, $2,700 per self-harm prevented; INMB = −1;Thresholds: $63,700/QALY gained, INMB >0, no threshold for cost/suicide death or self-harmHCrosland (2024), Australia (Continued) 10 & 30 years2019–20 AUD, 5%Intervention cost: program cost, administration costHealthcare, productivity, and carer costsQALYs, Suicide deaths, and self-harm(2) Healthcare perspective: ICERs = dominated, $16.1 million per suicide death prevented, $0.22 million per self-harm prevented; INMB = −92;Societal perspective: ICER = dominated, $19.76 million per suicide death prevented, $0.27 million per self-harm prevented; INMB = −109;Thresholds: $63,700/QALY gained, INMB >0, no threshold for cost/suicide death or self-harmHDoran (2016), Australia5 years2010 AUD, 4.1%No details on implementation costProduction disturbance costs, human capital costs, medical costs, administrative costs, transfer costs, othersPotential suicide and self-harm cases averted (Administrative data)ROI = 4.6AGarraza (2018), USLifetime2010 USD, 3%Staff salaries & benefits, travel, rent, phone, and other overhead expenses.Avoided healthcare costs: inpatient stays, readmission, follow-up, transport, rehabilitationSelf-harm case avoidedROI = 4.5 (95% CI: 1.59, 7.40) per dollar invested ICER = dominantHKinchin (2017), AustraliaLifetime (~40 years)2014 AUD, 4.58%Average cost MIC per worker per yearAvoided production disturbance, human capital, medical, administrative, and other costs (carers, postvention services)Suicide and self-harm cases avertedROI = 1.50 (95% CI 1.11–3.07)HLebenbaum (2020), CanadaLifetime (50 years)2016 CAD, 1.5%Intervention cost per person (no further details)Healthcare costs (hospitalization, ED, and family physicians), non-medical costs (autopsy, police & funeral services), productivity lossesQALYs (EQ–5D)ICER = $19,300 per QALY (Societal perspective); $21,600/QALY (Healthcare payer perspective)Threshold: $51,000 per QALY gainedHStelmach (2022), Global (36 countries)80 years (Lifetime)2019 USD, 3%Costs of salaries, salaries, medications, laboratory tests, computer equipment, materials, transportationWage returns to education, productivity gains & lossesUtility weights converted into monetary unit DALY avertedROI = 3.5 ICER = approximately <$110/DALYHVasiliadis (2015), CanadaLifetime (~60 years)2010 CAD, 3%Cost of prevention programHospitalization & ED visits, outpatient medications, non-medical costs, police & funeral services, productivity losses & short-term disabilityLife-years saved Suicide attempt caseNAD was cost-saving of $4,400 per life-year saved. ICER = $61,000 per case avertedHComplex table summarizing characteristics, interventions, comparators, outcomes, costs, and cost-effectiveness results for studies on suicide and self-harm prevention. See long description. Note: H, high quality; A, average quality; L, low quality. Table 2.Characteristics of included studies Table 2. long description.Lead author (year) countryDescription of interventionsDescription of comparatorsTypes of suicide targetedTarget population (age, %male)Evaluation methodStudy designPerspective1. Universal interventionsBandara (2022), AustraliaMeans restriction: Barriers installed at bridge and cliff sitesNo InterventionSuicidePeople who attempt suicide at bridges and cliffsROI, CEAModel (Decision tree)Partial societalDamerow (2020), Sri LankaMeans restriction: Shop-based gatekeeper training program (n = 535 shops): delivered at local pesticide shops in 2-hour sessions by field staff membersNo programSuicidePeople attempt suicide via pesticide poisoningCEAModel (Decision tree)Government perspectiveLaw (2011), Hong KongMeans restriction: Railway platform screen doors (PSDs) installation at the underground railway tracksNo interventionSelf-harmGeneral populationCUAQuasi-experimentalSocietal perspectiveLee (2021), Global (14 countries)Means restriction: Banning highly hazardous pesticidesNo interventionSuicideGeneral populationCUAModel (Markov)Healthcare systemWhitmer (2013), USMeans restriction: Suicide barrier on the Golden Gate BridgeNo barrierSuicidePeople who attempt suicide at Golden Gate BridgeCBA, CUARetrospective studyPartial societal perspective (assumed)Jackson-Morris ( 2024), UKAwareness training: Universal school-based training (5-h in-person sessions over 4 weeks by healthcare workers)No interventionSelf-harmAll adolescents enrolled in a school (aged 10–19)ROI, CUAModel (Markov)Partial societal perspective (assumed)Kinchin (2020), AustraliaAwareness training: Universal suicide prevention training (safeTALK) 3-hour workshop using presentations, videos, discussion, and questionsNo interventionSelf-harmStudents aged 15–16 yearsROIModel (Markov)Health and justice system & Societal perspectivesSari (2008), USAwareness training: General suicide education: 5 lessons to provide college students with suicide facts, warning signs, and seeking helpNo interventionSuicideYoung adults aged 18–24 yearsROIModel (Markov assumed)Societal perspectiveFlego (2022), AustraliaGuidelines/Policies: Mindframe – Media guidelines for journalists as a universal suicide prevention strategyNo interventionSuicideGeneral populationROI, CEA, CUAModel (Decision tree)Societal perspective2. Selective interventionsComans (2013), AustraliaSupport services: Standby Response Service: 24-hour multiple-center crisis response telephone number & in-person outreach (n = 90)Usual care: (n = 670)SuicidePeople bereaved by suicideCUAModel (Markov)Societal perspectivePil (2013), BelgiumSupport services: Chat helpline for suicide preventionTelephone helpline for suicide preventionNo interventionSelf-harm & SuicideUnique users of the suicide helpline, who called/chatted about their personal problemsCUAModel (Markov)Societal perspectiveRoss (2021), USSupport services: An active contact and follow-up (ACF) interventionNo interventionSelf-harm & SuicidePrimary care patientsCEA, CUAModel (Decision analytic)Healthcare system and Societal perspectiveSari (2008), USSupport services: Peer support group program: exploration of personal issues, build leadership & empathy, promote social skillsNo interventionSuicideYoung adults aged 18–24 years who may be at risk of suicideROIModel (Markov assumed)Societal perspectiveWilson-Barthes ( 2021), KenyaSupport services: Charitable Children’s Institutions: institutional care for orphanagesFamily-based careSuicidal ideationOrphaned and separated children and adolescents (OSCA)CEA, CUAProspective cohort study (n = 1,680)Societal perspectiveGray ( 2011), USPsychotherapy: Tailored and intensive family-centered in-home services (80–100 hours in 3 months); psychiatric treatment in mental health clinic (Pilot n = 22; At-scale n = 192)TAU (Pilot n = 22, At-scale n = 611)Suicidal ideationMentally ill teenagers (13–16 years old) at high risk for suicide within the juvenile court systemCEARCT for Pilot n = 44; Quasi-experimentation (At-scale n = 803)Societal perspective (assumed)Ross (2021), USPsychotherapy: A cognitive behavioral therapy (CBT) interventionNo interventionSelf-harm & SuicidePrimary care patientsCEA, CUAModel (Decision analytic)Healthcare system and Societal perspectiveTebbett-Mock ( 2020), USPsychotherapy: Dialectical behavior therapy (DBT): 9 weekly group training sessions for skill acquisition, daily intensive psychotherapy sessions in therapeutic groups (n = 425)TAUSelf-harmAdolescents (12–17 years) admitted to a coeducational, inpatient acute-careCEARetrospective cohort study (n = 801)Healthcare assumedZaloshnja (2003), USPsychotherapy: Suicide prevention & intervention program: mental health technician support, 2 weekly psychologist sessions, monthly consultations with psychiatristNo interventionSuicideNative American adolescents aged 15–19 years in communities with high risk of suicideCUA, CBAModel (Unclear)Societal perspectiveDazelle (2026) the USMedication: Oral iron therapy; Intravenous iron therapy (intravenous iron sucrose & ferric carboxymaltose)Do nothingSuicidePregnant women (M age = 28.5) with anemiaCEA, CUAModel (Markov)Healthcare & societal perspectivesDuggan (2003), UKMedication:ClozapineConventional neuroleptic therapySuicidePatients with severe mental illness & treatment-resistant schizophreniaCEA, ROIModel (Markov)Healthcare system (NHS assumed)Freemantle (1994), UKMedication:Selective serotonin reuptake inhibitors or newer tricyclic and related antidepressantsOlder tricyclicsSuicidePeople with depression & take antidepressantsCEAModel (Unclear)Healthcare perspective assumedGoren (2016), USMedication:Clozapine: daily dose of 600 mg, monitoring (weekly for the first 6 month and biweekly for the next 6 months) (n = 18,000)No clozapineSuicideVeterans with treatment-resistant schizophreniaCEAModel (Decision tree)Healthcare provider (Veteran Health Admin)Groessl (2018), USMedication:Pharmacogenetic test (IDGx): buccal cell swabs with results used to guide medication selectionStandard of care (SOC)SuicidePatients with treatment-naïve or inadequately controlled major depressive disorderCUA, CEA, CBAModel (Markov)Societal perspectiveHughes (2023), USHealth Policies: Psychologist prescriptive Authority: state policy granting prescriptive authority to qualified psychologistsNo interventionSelf-harm & SuicidePeople with mental health issuesCBA, CEA, CUAModel (Markov)Societal perspectivePark (2016), UKCombined intervention: Early intervention: mental health services from 10 professionals over a year: medication, psychotherapy, counseling, vocational support.Standard careSuicidePeople with first-episode psychosisCEAModel (Decision tree)Broader economic (societal perspective assumed)3. Indicated interventionsAcolin (2022), USPsychotherapy: Dialectical behavioral therapy (DBT): weekly individual & group sessions & clinical supervision for therapists for 1 year (n = 120)Cognitive behavioral therapy (CBT): a brief group-based treatment (5–12 sessions) (n = 180)Self-harm & SuicideAdults aged 18–64 with previous suicide attemptsCEA, CUAModel (Markov)Healthcare systemAlvi (2022), PakistanPsychotherapy: Culturally adapted manual-assisted problem-solving training (C-MAP) plus TAU (n = 108): 6–8 sessions by a psychologist at home or healthcare settingTAU (n = 113)Suicidal ideationPatients admitted after an episode of self-harmCUARCT (n = 221)Healthcare providerBernecker (2020), USPsychotherapy: Brief cognitive behavioral therapy (BCBT): 12 individual sessions + TAUTAUSelf-harm & SuicideSoldiers with recent suicidal crises in the US ArmyCEAModel (Decision tree)Program providerBjureberg (2025) SwedenPsychotherapy: Internet-delivered emotion regulation individual therapy for adolescents (IERITA) + TAU: 11 digital modules for adolescents & 6 modules for parents over 12 weeks with support from a dedicated therapistTAU: regular mental health services (therapy sessions biweekly)Self-harmAdolescents (M age = 15, 4% male) with self-harm in the past monthCEA, CUARCT (n = 166)Healthcare & societal perspectivesBojke (2024), UKPsychotherapy: Family therapy: monthly 1.25-hour therapy session over 6 monthsTAUSelf-harmYouths (11–17 years) with self-harm at least twiceCUAModel (Quasi-Markov)HealthcareByford (1999), UKPsychotherapy: Home-based social work intervention: assessment & 4 intensive, family-centered, home-based intervention sessions by psychiatric social workers + Routine care (n = 85)Routine care (n = 77)Suicidal ideationChildren aged <16 with deliberate self-poisoningCEARCT (n = 162)Societal perspectiveByford (2003), UKPsychotherapy: Manual-assisted cognitive behavior therapy (MACT): treatment manual and up to 7 treatment sessions with a trained therapist (n = 239)TAU (n = 241)Self-harmPatients aged 16–65 with a history of recurrent deliberate self-harmCUA, CEARCT (n = 480)Societal perspective (assumed)Cottrell (2018), UKPsychotherapy: Family therapy: monthly group supervision of 2 hours, monthly therapy session of ~1.25 hours over 6 months (n = 415)TAU (n = 417)Self-harmYouths (11–17 years) with self-harm at least twiceCUAPragmatic, multi-center RCT (n = 832) + Model (Markov)Healthcare system (NHS)Denchev (2018), USPsychotherapy: Suicide-focused (CBT) (9 sessions (bi)weekly psychotherapy).TAUSelf-harm & SuicideAdults with self-harm presenting to emergency careCEAModel (Markov)Healthcare perspective assumedGreen (2011), UKPsychotherapy: Group psychotherapy: 6 weekly psychotherapy sessions (acute phase) followed by a booster phase of weekly groups as needed + routine care (n = 183)Routine care (n = 183)Self-harm & Suicidal ideationAdolescent (12–17 years) with at least 2 past episodes of self-harm in the previous yearCEARCT (n = 366)Societal perspectiveHaddock (2019), UKPsychotherapy: Cognitive behavioral suicide prevention therapy: up to 20 sessions of 1 hour over 6 months in the community following discharge. + TAU (n = 24)TAU (n = 27)Suicidal ideationIn-patients (aged 18–65) in acute psychiatric wards with suicidal thoughts/behaviorsCUARCT (n = 51)Healthcare (NHS), social care, and patient perspectivesHuntjens (2025), NetherlandsPsychotherapy: Dialectical behavioral therapy (DBT) (n = 63): weekly 45 min individual CBT, weekly 2-hour skill training group, telephone coaching, weekly 1-hour therapist consultationCAU (n = 60): weekly 45-minute sessions with a psychotherapist or social workerSuicide ideationAutistic adults with suicidal behaviors (M age = 37.4, 52% male)CEA, CUARCT (n = 123)Healthcare & societal perspectivesHaga (2018), NorwayPsychotherapy: Dialectical behavioral therapy (DBT): 19 weekly 60-minute sessions of individual therapy, 120 min sessions of skills training, family therapy, telephone coaching, outpatient treatment (n = 39)Enhanced usual care (EUC) (n = 38)Self-harmAdolescents aged 12–18 years with repeated self-harm, borderline personality disorderCEARCT (n = 77)Healthcare perspectiveKrysinska (2024), AustraliaPsychotherapy: Cognitive behavioral/problem-solving therapyNo interventionSelf-harmAdults presented to hospitals with self-harmROI, CUAModel (Markov)Partial societal perspectiveLelie (2025), BelgiumAttempted Suicide Short Intervention Program (ASSIP): 3 sessions & one optional session after 3 week post discharge from hospitalCAU: short psychiatric evaluation, 3 outpatient sessions, 2 nights in psychiatric wardSelf-harm & SuicidePatients admitted to hospital with previous suicide attempt (M age = unknown, 39% male)CUAModel (Markov)Healthcare & societal perspectivesMartínez-Alés ( 2021), SpainPsychotherapy: Problem-solving Psychotherapy program: an outpatient appointment, weekly individual psychotherapy over 2 monthsTAUSelf-harmAdults aged 18 years or above with previous suicide attemptsCEAModel (Decision tree)Societal perspectiveMavranezouli (2024), UKPsychotherapy:(1) Cognitive behavioral therapy (CBT) + TAU: individual face-to-face psycho-education, problem solving(2) Dialectical behavioral therapy (DBT): weekly individual and group therapy sessions and support (consultation)TAU: treatment provided by community mental health.Self-harm(1) Adults who have self-harm and present at hospitals(2) Children and young people who have self-harm and present at hospitalsCUAModel (Decision analytic)Healthcare perspectivesMcCutchan (2022), USPsychotherapy: The Collaborative Assessment and Management of Suicidality (CAMS): builds a therapeutic relationship and addresses suicidal thoughts through 4–11 structured weekly sessionsEnhanced treatment as usual (ETAU)Suicidal ideationSoldiers with current suicidal ideationROIRCT (n = 148)Healthcare systemPark (2018), SwitzerlandPsychotherapy: Attempted Suicide Short Intervention Program (ASSIP): 3 60-minute face-to-face sessions, then semistructured personalized letters sent 6 times over 2 yearsTAUSelf-harmIndividuals with suicide attempt and treated at hospitalCEARCT (n = 120)Healthcare systemPriebe (2012), UKPsychotherapy: Dialectical behavior therapy (DBT): weekly hour-long individual therapy sessions, 2-hour skills training group, and telephone coaching (n = 40)Treatment as Usual (TAU) (n = 40)Self-harmParticipants with personality disorder, ≥5 days of self-harm in the previous yearCEARCT (n = 80)Societal perspective assumedTubeuf (2019), UKPsychotherapy: Family therapy (FT): involved monthly sessions lasting 1.25 hour over 6 months, initially more frequent, totaling six to eight sessions. (n = 383)Treatment as usual (TAU) (n = 371)Self-harmAdolescents aged 11–17 years with previous self-harm episodeCUARCT (n = 754)HealthcareTyrer (2004), UKPsychotherapy: Manual-assisted cognitive behavioral therapy (MACT): a booklet, 7 treatment sessions with a trained therapist (n = 239)TAU (n = 241)Self-harmIndividuals with a previous episode of self-harm and personality disorderCEARCT (n = 480)Broad economic (societal perspective)Spijker (2012), NetherlandsPsychotherapy: Unguided Web-Based Self-help: six weekly modules on suicidal thoughts, intense emotions, negative thinking, and relapse prevention (n = 116)Waitlist, a website that provided information on suicidality (n = 120)Suicidal ideationAdults (over 18 years old) with mild to moderate suicidal thoughtsCEARCT (n = 236)Societal perspectivevanBentum (2025), NetherlandsPsychotherapy: Add-on module to reduce intrusive suicidal mental images + TAU (n = 46): 2–6 60-minute weekly sessions at mental healthcare center by trained therapistsCAU (n = 45): psychotherapy and/or pharmacotherapySuicide ideationAdult psychiatric outpatients with depressive symptoms and distressing suicidal intrusions (M age = unknown, 32% male)CUARCT (n = 91)Healthcare & societal perspectivesVasiliadis (2017), CanadaPsychotherapy: Increasing publicly funded access to psychotherapy: two visits with a GP, 8 sessions with a mental health professionalStatus quoSelf-harm & SuicideAdults aged 20–85 years with incident depression or self-harmCUA, ROIDiscrete event simulationHealthcare system and Societal perspectiveBorschmann (2013), UKSupport services: Joint crisis plans (JCP) developed collaboratively and finalized with care coordinators, included in treatment plans alongside TAU to enhance crisis preparedness and support (N = 46)Treatment as usual (N = 42)Self-harmAdults with borderline personality disorder & a self-harm record (Age M = 35.8, 19.3% male)CUARCT (N = 88)Health and social care sector perspectiveDenchev (2018), USSupport services:(1) postcards/ letters (bi)monthly, (2) telephone outreach (1–3 months)TAUSelf-harm & SuicideAdults with self-harm presenting to EDCEAModel (Markov)Healthcare perspective assumedGallien (2025), FranceSupport services: Brief contact intervention (BCI): post-hospital follow-up with structured crisis card, telephone calls, postcardsCAUSelf-harm & SuicideIndividuals admitted to hospital with suicide attempt (M age = 39, 37% male)ROI, CBARetrospective cohort study (N = 2,3146)Healthcare perspective (assumed)Le (2023), AustraliaSupport services: Way Back Support Service: 4 brief aftercare interventions: (a) follow-up telephone contacts, (b) care coordination, (c) safety planning, (d) brief therapyNo interventionSelf-harmIndividuals aged 15+ years admitted to hospital/ emergency after a self-harm episodeROI, CUAModel (Markov)Partial societal perspectiveLindkvist (2024), SwedenSupport services: Brief Admission by self-referral (BA) + TAU: up to 3 nights in the hospital, bring their own medication, supportive meeting with hospital staffsTAU: access to outpatient and acute psychiatric clinicsSelf-harmIndividuals with self-harm at risk of suicide (M age = 32, 12% Males)CUARCT (n = 117)Payer perspectiveMartínez-Alés ( 2021), SpainSupport services: Enhanced Contact intervention: one-time priority outpatient appointment, in-person, and telephone follow-upTAUSelf-harmAdults aged 18 years or above with previous suicide attemptsCEAModel (Decision tree)Societal perspectiveO’Connor (2017), UKSupport services: Volitional help sheet (VHS): a brief encouragement statement, a list of self-harm situations & potential solutions given within 24 hours after admissionTAUSelf-harmIndividuals aged 16+ years with self-reported self-harm and suicide attemptCEARCT (n = 1,308)Healthcare systemRichardson (2014), USSupport services: Post-discharge follow-up calls by eight call centers in collaboration with hospitals and emergency departmentsNo interventionSelf-harm & Suicidal ideationIn patients with previous self-harm or suicidal ideationROIModel (Decision tree assumed)PayerRyan (2015), the USSupport services: National Suicide Prevention Lifeline crisis center (telephone helpline): risk assessment, safety planning, referrals, emergency rescue activationNo interventionSelf-harm & SuicideCallers with self-harm and suicide ideation (38% male)CUA, CBARetrospective quasi-experimental study (N = 1,7202)Societal perspectiveStelmach (2022), Global (36 countries)Support services: Follow-up prevention in hospitals: patient education and follow-upNo interventionSelf-harm & SuicideAdolescents aged 10–19 years with previous self-harmCUA, ROIModel (Markov)Societal perspectiveAppleby (2000), UKCombined intervention: Suicide risk management training for professionals (2-hour sessions over 6 months) to provide assessment, clinical management, and problem-solving for patients at risk of suicideNo interventionSuicideSuicidal individuals in primary care, ED, mental health servicesCEAPre-Post StudyHealth and social care sector perspectiveDunlap (2019), USCombined intervention: Screening by physician plus 7 telephone sessions, self-personal safety plan, treatment information, suicide hotlineTAUSelf-harm & SuicideIndividuals with self-harm ideation or behavior upon entry to emergency careCEACluster RCT (n = 1,376)Program providerJackson-Morris ( 2024), UKCombined intervention: a 1-hour information session at a hospital, 9 brief follow-up contacts over 18 months delivered by mental health support workersNo interventionSelf-harmAdolescents with self-harm and treated at a hospitalROI, CUAModel (Markov)Partial societal perspective (assumed)Latimer (2014), CanadaCombined intervention: Rapid response team (RRT) (n = 158) included a psychiatrist, psychiatric nurse, and others offered assessment, treatment initiation (medication, CBT), and referralsUsual care (n = 128)Self-harmSuicidal adolescents aged 12 to 17 years at emergency departmentCEATrial (n = 286)Healthcare & societal perspectivesBotchway (2022), UKScreening: Oxford Mental Illness and Suicide (OxMIS): a suicide risk assessment for people with SMI (medication & clinical reviews, multi-disciplinary assessment, elective admission, ED visit)Usual careSuicidePeople with severe mental illness (SMI) admitted to the hospital due to nonfatal self-harmCUAModel (Decision tree)Healthcare system (NHS)Dunlap (2019), USScreening: Screening by ED nurses plus TAUTAUSelf-harm & SuicideIndividuals with self-harm ideation or behavior upon entry to emergency careCEACluster RCT (n = 1,376)Program providerMcDaid (2022), UKScreening: Psychosocial assessment administered upon arrival to hospital due to self-harmNo interventionSelf-harmIndividuals admitted to hospital due to self-harm (age: 15+, 41%)CUAModel (Markov)Healthcare and partial societal perspectivesde Beurs (2015), NetherlandsTraining for professionals: e-learning program -Train-the-Trainer: 1 day in-person training with e-learning module for professionals + IAU (n = 312)Usual implementation (IAU) (n = 254)Suicidal ideationPatients 18 and above with suicidal ideationCUA, CEARCT (n = 566)Societal perspective4. Multi-level interventionsAhern (2018), Global (10 EU countries)3-level School-based prevention:(1) Youth Aware of Mental Health (YAM): promote mental health knowledge(2) Question, Persuade, Refer (QPR): train gatekeepers to identify and intervene student’s risky behaviors.(3) Screening by Professionals (ProfScreen): screen, identify, and refer at-risk adolescentsControl group: only exposed to the same educational posters as in YAMSelf-harm & Suicidal ideationSchool-age children and adolescentsCEA, CUACluster-RCT (n = 11,110 students)Payer (education, health, & social care system)Crosland (2024),Australia(1) Emergency department-based suicide prevention (EDSP): Universal and secondary suicide risk screening, discharge resources, follow-up phone callsCAUSelf-harm and suicide(1) Young people (aged 25 and below) with psychological distressCUA, CEA, CBAModel (Dynamic simulation techniques)Healthcare and societal perspectivesCrosland (2024),Australia (Continued) (2) School-based suicide prevention program (SSP): universal mental health education and awareness training; indicated brief screening and provision of CBT; treatment seeking assistanceCAUSelf-harm and suicide(2) Children with psychological distressCUA, CEA, CBAModel (Dynamic simulation techniques)Healthcare and societal perspectivesDoran (2016), AustraliaMates in Construction: awareness training (1 hr), connector (4 hr), intervention skills (2 days), promotion materials, 24/7 helplineNo program assumedSelf-harm & SuicideConstruction industry workersROIModel (Decision tree)Societal perspective (assumed)Garraza (2018), USCommunity-based suicide prevention programs (GLS): gatekeeper training, prevention hotlines, education, screeningNo InterventionSelf-harmYouths aged 16–23ROI, CEAModel (Unclear)Healthcare systemKinchin (2017), AustraliaMates in Construction: awareness training (1 h), connector (4 h), intervention skills (2 days), promotion materials, 24/7 helplineNo interventionSelf-harm & SuicideAustralian workforce (mean age = 42)ROIModel (Markov assumed)Societal perspectiveLebenbaum (2020), CanadaAwareness campaigns, diagnostic training, educating primary-care physicians, & providing high-risk psychosocial interventions.No interventionSelf-harm & SuicideOntario residents aged 16+ years (M age = 46 years)CUAModel (Markov)Healthcare payer, Societal perspectivesStelmach (2022), Global (36 countries)Universal School-based prevention: gatekeeper training, mental health training for kids, screening, professionals’ referral (3 programs similar to Ahern)No interventionSelf-harm & SuicideAdolescents aged 10–19 yearsCUA, CBAModel (Markov)Societal perspectiveVasiliadis (2015), CanadaNuremberg Alliance Against Depression (NAD): community interventions for depression care and suicide prevention via universal awareness campaign, diagnostic and treatment training, and follow-up of individuals with previous self-harmNo interventionSelf-harmFirst responders and follow-up of individuals who attempted suicideCEAModel (Markov assumed)Healthcare system and Societal perspectiveLead author (year) countryTime horizonYear of pricing, discount ratesCost of interventionCost savingsHealth outcomes/benefitsResults ICERS (in 2022 USD value) thresholdQuality1. Universal interventionsBandara (2022), Australia5 years10 years2018 USD, 3%Initial installation and maintenance (cyclical and condition-based)Value of statistical lifeSuicide cases avertedROI = 1.7 (95% UI, 1.0–1.9) (5 yr) and 2.4 (95% UI, 1.5 to 2.7) (10 yr) for building barriers at bridges.ROI = 1.7 (95% UI, −0.9–3.3, ns) (5 yrs), and ROI = 2.0 (95% UI, −1.1–3.8, ns) (for 10 yr) for building at cliff sites.ICER = dominant for barriers over bridges BUT not for cliff sites.HDamerow (2020), Sri Lanka3 years2019 USD, 3%Direct personnel, travel, catering, admin, training materials, equipment, communicationFatal pesticide self-poisoning case prevented; Life-year savedCost-effective if preventing 0.23 fatal pesticide self-poisoning cases over 3 yearsThreshold: $4,860/life-year savedHLaw (2011), Hong Kong25 years2002 USD, 3%Costs of installing railway platform screen doorsPassenger time Fare revenue saved after installing Railway platform screen doorsDALYsNumber of suicidal and accidental fallsICER = $122,000 per DALY averted and decreased to $103,000 if accounting for the loss of fare revenue and passengers’ waiting timeThreshold: $117,000 per DALY avertedHLee (2021), Global (14 countries)100 years2017 I$, 3% for cost, 0% for health outcomeProgram costs (admin, training, media, regulation) and patient-level costsNoneHealthy life-years gained (HLYGs)Suicide Mortality rateICER = $110 per HLYG (95% UI 86–145) in LMICs and $280 per HLYG (95% UI 225–356) in UMHICsThreshold: $118/HLYGHWhitmer (2013), US20 years2009 USD (assumed), unknown discount rateCost of implementing suicide barrierValue of statistical lifeNumber of lives saved (Mortality reduction)DALYsCBR = 1; ICER = $6,528 per DALYThreshold: $193,000/DALY avertedLJackson-Morris (2024), UK10–80 years (lifetime)2021 GBP, 3%Salary, equipment (computer, phone call, internet access)Long-term employment and earningsDALYsICER = $4550 (10 years) & $970 (lifetime) /DALY avertedROIs = −0.4 (10 years) & 0.9 (lifetime)HKinchin (2020), Australia5 years2017–18 AUD, 5% for costsTravel and materials for local healthcare services and non-clinical groups.Ambulance, healthcare, police, investigation, funeral, productivity losses, grievingNoneHealth and justice perspective: ROI = 1.45 in Mackay; ROI = 0.19 in Queensland; ROI = 0.15 in AustraliaSocietal perspective: ROI = 31.2 in Mackay; ROI = 4.1 in Queensland; ROI = 3.3 in AustraliaHSari (2008), USLifetime (~40 years)2000 USD, 6.3%General suicide education: training & materialsDirect cost (ambulance & autopsy services); Indirect cost (earnings lost, productivity loss of family members)NoneROI: 2.03HFlego (2022), Australia5 years2018 AUD, 3%Staffing and non-staffing costs: review and update guidelines, training, support,Costs of suicide death prevented (value of a statistical life, total cost associated with a suicide death)Suicide deaths preventedQALYROI: 94 (95% CI 37–170) The intervention was dominant regarding cost per suicide prevented or QALYs gainedThreshold: ROI > 1H2. Selective interventionsComans (2013), Australia12 months 5 years2010–2011 AUD assumed, 5%Operating costs of providing intervention, community development, engagement, and training.Health costs, Productivity lost (HPQ): Absenteeism, Presenteeism, unpaid workQALY (EQ–5D)The StandBy service is dominanceWTP Threshold: $46,000/QALY gainedHPil (2013), Belgium10 years2012 EUR, 3% for cost and 1.5% for effectsCosts of salaries, transportation of personnel, stipends for volunteers, operation, and communicationAmbulance, medical examiner, ED, inpatient care, nursing home, productivity lossQALYs (EQ–5D)ICER = dominantNet societal savings of telephone and chat service: $3,600 (95%CI 2,960–4,330) and $3,460 (95%CI 2,800–4,200) in male users; $3,300 (95%CI 2,630–4,040) and $3,700 (95%CI 2,950–4,600) in female users.HRoss (2021), USLifetime2016 USD, 3%Intervention costs (No details)Healthcare, patient time, productivity lossSuicidal risk (Fatal and nonfatal suicide attempt) QALYsTo achieve ICERs below $50,000 per QALY gained, ACF needed a suicide prediction method with a sensitivity of ≥57% and specificity of ≥95%HSari (2008), USLifetime (~40 years)2000 USD, 6.3%Peer support group: housing & food for retreat, trainers’ salaries, stipend, computer, transportationDirect cost (ambulance services and autopsy services); Indirect cost (earnings lost & productivity loss of immediate family members)NoneROI: 3.71HWilson-Barthes (2021), Kenya10 years2018 USD, 3%Personnel salaries, overhead, training, education fees, capital costs (e.g. appliances, equipment)NoneReduction in suicidal ideation & intent (CDI-SF scale)ICER for institutional care: $526 [179, 959] per unit reduction in suicidal diagnosis and $5,680 per QALY gained among OSCAWTP threshold: $8,100/QALY gainedHGray (2011), US12 months2009 USD assumed, no discountTreatment costs & Cost of days spent in observation and assessment at detention centers and correctional facilitiesProductivity loss of parentsRecidivism, days in court placement, Suicidal ideation (Youth Outcome questionnaire-YOQ)Intervention is cost-saving in improving score on the youth outcome questionnaire for suicidal ideationLRoss (2021), USLifetime2016 USD, 3%Intervention costs (No details)Healthcare, patient time, and lost productivity costs associated with fatal and non-fatal suicide attemptsSuicidal risk (Fatal and nonfatal suicide attempt)QALYsTo achieve ICERs below $50,000 per QALY gained, CBT needed a suicide prediction method with a sensitivity of ≥75% and specificity o ≥ 97%HTebbett-Mock (2020), US8 months2017 USD assumed, no discountNone (Not provided)Cost of constant observation hoursIncidents of suicide attempts, self-injury; restraints; and days hospitalized, aggression patient-to-patient/staffIntervention is cost-saving: $141,000 for suicidal ideation, $90,000 for aggression, and $66,600 for self-injurious behaviors.AZaloshnja (2003), USLifetime1998 USD, 3%Costs of salaries and educationalCUA: medical and out-of-pocket costCBA: cost savings to individuals, local governments, insurance companies, employers, and the general publicQALYsBCR: 43ICER: $710 per QALY gained Threshold: $150,000/QALY gainedHDazelle (2026) the USLifetime2020 USD, 3%Costs of administering iron therapyMedical costs (blood transfusions, premature births), productivity lossesQALYAdverse outcomes averted: post-hemorrhage, hysterectomies, suicideICER: dominant, Incremental NMB = $1,141–$2,128 Ferric carboxymaltose confers the greatest overall benefitThreshold: $100,000 per QALY gainedHDuggan (2003), UK10 years2001 GBP assumed, 1.5% for outcomes & 6% for costsPatient support costsCosts of clozapine & other alternative drugsCosts of suicide: general hospital costs, inpatient costsLife-years savedClozapine in treatment-resistant schizophrenia is cost-effective.Cost-neutral scenario: $12,000 per life-year savedCost-saving scenario: $20,000,000 net annual savingAFreemantle (1994), UKLifetime1992 GBP, 5%Medication costsSuicide cases & deaths related to anti-depressants (administrative data)ICER = $54,600–$486,000 per life-year gainedSelective serotonin reuptake inhibitors are more costly and more effective than the older tricyclic.AGoren (2016), US12 months2015 USD, no discountLaboratory monitoring and follow-up visits, Medication costsDirect healthcare cost: prescriptions, inpatient & outpatient treatment, emergency services, and office-based medical visitsTreatment response, suicides, adverse drug reactions & associated mortality, hospitalization due to clozapineThe intervention resulted in cost-saving per case of suicide-related death averted.HGroessl (2018), US3 years2016 USD, 3%Cost of pharmacogenetic testDirect medical costs (medications, hospitalizations, physician visits, psycho-therapy)Indirect medical costs (lost productivity)QALYs (EQ–5D)Suicide rateIDGx is dominant for moderate to severe MDD patientsThreshold: $60,000 per QALY gainedHHughes (2023), US20 years2021 USD, 3%Personnel, designation fee, maintenance, and individual prescribing psychologistsMedical costs (ED visits and hospitalization due to suicide)Societal costs (productivity, quality of life) following the suicideSuicide attempt, suicide death, QALYsICER = $48,300 per QALY gained; $28,700 per nonfatal suicide attempt averted; 28,161 per suicide death avertedIntervention is dominant at WTP threshold of $53,000HPark (2016), UK10 years2009 GBP, 3.5%Costs of consultation with health professionals, formal and informal admissionProductivity losses, health service use, reduced quality of lifeSuicide case avertedICER = dominant (Societal cost savings of $1,815 per person per year)H3. Indicated interventionsAcolin (2022), US12 months2022 USD, no discountCost of therapist’s timeCosts of hospitalizationLife yearsQALYICER = $26,362 per QALY gainedICERs for life-years gained or attempted suicides prevented were negligible.Threshold = $50,000 per QALYHAlvi (2022), Pakistan6 months2019 USD, no discountPersonnel travel and time, transportation and reimbursement, materials, supplies, and communication costsHealth services: Inpatient & outpatient, GP & other health professional visitsSuicidal ideation (Beck Scale for Suicidal ideation)QALY (EQ–5D–3L)ICER for C-MAP = $18,410 (95% CI $8,060–$112,203) per QALY gainedThreshold: no formal WTP threshold for Pakistan, thus used WTP threshold $22,600–$34,000HBernecker (2020), USLifetime2017 USD, Federal discount rate & 3%Psychotherapy sessions, psychiatric inpatient days, & patients’ intake evaluationsMortality costs, Suicide attempt costs, retired disability pay costsSuicide attempts avertedSuicide deaths avertedCost saving = $17,700–$19,640/patient Except for the worst-case scenario,ICER of BCBT was dominant for $/suicide attempt & suicide death averted.HBjureberg (2025) Sweden1 & 3 months2022 USD, no discountTherapist’s timeHealthcare visits, medication, social support, productivity lossesSelf-harm frequency (DSHI-Y),QALY (CHU–9D)– ICERs - societal perspective: $18,677 (1 month) & $27,563 (3 month) per one adverted case; & $792,244 (1 month) & $471,800 (3 month) per QALY gained– ICERs – healthcare perspective: $15,502 (1 month) & $29,972 (3 month) per one adverted case; & $660,793 (1 month) & $361,041 (3 month) per QALY gainedWTP threshold: $84,000 per QALY gainedHBojke (2024), UK1.5 and 5 years2022 GBP (assumed), 3.5%Personnel costs, telephone contact, supervisionCosts of hospital admissionsNon-hospital cost: primary & community care, medicationQALYsICERs = $53,000/QALY gained (1.5 years) – dominated (5 years)Threshold: $43,000/QALY gainedHByford (1999), UK6 months1997–1998 GPB, no discountCosts of personnel, travel, supervision, hospital services, GP, social worker, and other healthcare and social servicesSuicidal ideation (Suicidal Ideation Scale)Family functioning (Family Assessment Device) General HealthThe intervention is as cost-effective as routine care aloneFor children with no depression, intervention is more effective in reducing suicidal ideation with no difference in costs.HByford (2003), UK12 months1999/2000 GBP, no discountProductivity losses, living expenses, hospital, community health, social services, criminal justiceRepeated episodes of deliberate self-harm (Parasuicide History Interview) HRQoL (EQ–5D)At 12 months, MACT is dominant in reducing repeat self-harm, and ICER = $156,750/QALY gained.WTP Threshold: 0-$156,750/QALY gainedHCottrell (2018), UK18 months 5 years2014 GBP, 3.5%Personnel costs, telephone contact, supervisionHealthcare costs (health, social services, in- & out-patient care) MedicationsHospitalization due to self-harm Suicidal ideation (BSS) QALY (PQ-LES-Q & GHQ–12)At 18 months (base case), FT is dominated in preventing self-harm events & ICER = $64,185/QALYAt 5 years, ICER = $34,000/QALY gainedThreshold: $35,000–$52,000 per QALY gainedHDenchev (2018), US54 weeks2014 USD assumed, no discountDirect cost of intervention deliveryCosts of hospitalization & healthcare (ED visits, inpatient & outpatient care)Self-harm and suicide incidents averted Life-years savedICER = $22,100 per life-year saved. WTP threshold: $61,000 per life-year savedAGreen (2011), UK12 months2005–6 GBP, no discountSalary of therapist, overhead, indirect therapist time (preparation), travel costTotal costs of health, social care, education, and criminal justice sector services, plus family-related costs and productivity lossesFrequency of repeated self-harm (interview), severity of subsequent self-harm (triangulated reports), suicidal ideation (SIQ)Group psychotherapy is not a cost-effective addition to the routine care deliveredHHaddock (2019), UK6 months2015–16 GBP, no discountIntervention costCommunity services, primary care, accident and emergency care, in- & out-patient careSuicide behavior & ideation (SBQ, BSSI, BHS, clinical records); QoL (QoL-Brief); QALY (EQ–5D)ICER = $637,000/QALY gained Intervention is less costly and less effective than TAU. 95CI suggests no difference in costs and QALYs for both groupsHHuntjens (2025), Netherlands12 months2023 EUR, no discountCosts of administering DBT & supervision costs for therapists providing DBTDirect healthcare costs: health consultations, hospitalizations, medication Non-healthcare costs: carer costs, transportation, productivity losses.QALYAverted suicidal ideation on SIDAS scaleICER: dominant (healthcare perspective); & $1,527 per QALY gained & $1,642 per point improved on SIDAS scale (societal perspective)WTP threshold: $61,000 per QALY gainedHHaga (2018), Norway12 months2012 EUR, no discountWages for staff, equipment, IT, house rent, telephone consultations, medication, therapist team consultations.Health services due to self-harm (inpatient treatment, emergency care, and GP consultationsSelf-harm episodes (LPC)Global functioning (CGAS)ICER: less costly & more effective with 99.5% cost-effective at thresholds: $2,000/self-harm episode avoidedHKrysinska (2024), Australia10 years2018 AUD, 3%Costs of referral pathway, delivering intervention, weekly health professionals’ supervision costCosts of hospital admission related to self-harm, suicide, and death due to other causesQALYROI = 2.3–6.0ICERs = dominant Threshold: ROI > 1, $41,100 per QALY gainedHLelie (2025), Belgium20 years2023 EUR, 3% for costs, 1.5% for benefitsASSIP start-up costsHealthcare costs: psychiatric consultation, hospitalization Productivity loss: recuperation, and economic replacement costsSuicide re-attempt or suicide death QALYICER: dominant (societal perspective), $360 per QALY gained (healthcare perspective)WTP threshold: $52,000 per QALYHMartínez-Alés (2021), Spain12 months2019 EUR, no discountCosts of TAU plus 8 psychotherapy sessions.Extra-hospital emergency care, in- & out-patient care. Indirect costs of death & disability leaveSuicide reattempt within 1 year after hospital discharge Time to reattemptICER = $10,600 per averted attempt WTP Threshold: $47,300–67,500/suicide avertedHMavranezouli (2024), UK5 years2020 GBP, 3.5%(1) Costs of individual sessions(2) Costs of individual psychotherapy, group skill training, team and individual consultationsHealthcare costs associated with self-harm: primary care, emergency department, inpatient care, medication, social careQALYs(1) ICER = $13,800/QALY gained(2) ICER = $407,000/QALY gainedThresholds: $30,000–45,500 per QALY gainedHMcCutchan (2022), US12 months2018 USD, no discountImplementation costs (training, personnel, travel, space, material, treatment costs (e.g. facilities, admin, insurance)Healthcare expenditures, medical services, crisis services, & medicationsSuicidal ideation (The Scale for Suicidal ideation-Current - SSI-C)Mean CBRs at 12-month: −13.26 (SD = 55.10) for ETAU and 1.68 (SD = 40.53) for CAMS. Mean within-subject net benefit at 12-months: -$4,300 (SD = $17,000) for ETAU and -$480 (SD = $14,900) for CAMS.HPark (2018), Switzerland24 months2015 CHF, 3.5% for costs & 1.5% for health outcomesTraining costs for hospital staffInpatient care, outpatient careSuicide attempt avoidedASSIP was dominant, with significantly fewer reattempts at lower overall cost.WTP threshold: $0 and $25,600HPriebe (2012), UK12 months2010 GBP assumed, no discountService costs (no further details provided)Costs of primary, secondary & social care: hospital, outpatient visits (GP), medication, lost workdaysSelf-harm rate Borderline personality disorderICER = $67 per one percentage point reduction in the incidence of self-harmHTubeuf (2019), UK12 months2017 GBP assumed, No discountCosts of therapies sessionsHealth service use (accident and ED visits and inpatient stays), psychotropic medicationsQALYs (EuroQoL 5 Dimensions 3 Levels (EQ–5D–3L) for the adolescent and the Health Utility Index (HUI2) for one parent)ICER = $67,500 per QALY gained (include adolescent’s health gain only)ICER = $45,400 - $68,200 per QALY gained (including adolescent’s health gain, health spill over to one parent)Threshold: $33,400–50,000/QALY gainedHTyrer (2004), UK12 months2002 GBP assumed, No discountNo detailsCost incurred by all service-providing sectors, productivity lossRepetition in self-harm Time until repetitionNo significant difference between self-harm cases in the MACT (39%) compared to TAU (46%) (P = 0.20), MAC was cost-effective (10% cheaper than TAU)HSpijker (2012), Netherlands1.5 month (6 weeks)2009 EUR, no discountCosts of intervention, visits to psychologists, website maintenance, and hostingDirect costs: medical (healthcare, prescription); non-medical (participants’ travel, loss of leisure time, informal care); Productivity lossesChange in suicidal ideation: Beck Scale for Suicidal ideation (BSS)/Suicidal ideation (BSS) Treatment responseICER was cost saving = −$55,000 per additional treatment response.WTP Threshold: 0/treatment responseHvanBentum (2025), Netherlands12 months2024 EUR, no discount assumedPsychologists’ time providing add-on moduleHealthcare costs: psychiatric healthcare, medication, out-of-pocket costs Productivity losses, travel expensesQALYICER = $160,256 (societal perspective) & $116,605 (healthcare perspective) per QALY gained WTP threshold: $93,000 per QALY gainedHVasiliadis (2017), Canada40 years2012 CAD assumed, no discount specifiedCosts of psychological therapy services for individuals received inadequate careMedical, nonmedical, and productivity costsQALYsICER = dominant (Societal cost savings: $2,700 per person (95% CI $1,340–$6,700))ROI = 2.00 (1.78–3.15)HBorschmann (2013), UK6 months2009–2010 GBP, no discountCosts of salaries, overheads, preparation & trainingCosts of health and social care services: Hospital, community health and social care, medication, criminal justice sector servicesSelf-report self-harm post 6 months, Depression and anxiety (HADS), Well-being (WEMWBS), Perceived coercion (TES), QALY (EQ–5D)JCP + TAU is dominantHDenchev (2018), US54 weeks2014 USD assumed, no discountDirect cost of intervention deliveryCosts of hospitalization & healthcare (ED visits, inpatient & outpatient care)Self-harm and suicide incidents averted Life-years savedPostcards is dominant ICER = $5,280 (Telephone) per life-year saved.WTP threshold: $61,000 per life-year savedAGallien (2025), France12 months2022 EUR, unspecified discountCosts of administering BCI (unclear cost components)Direct healthcare costs (assumed)Suicide re-attempt or suicide deathROI: 2.06NMB per patient: $331 in healthcare cost savingHLe (2023), Australia10 years2018 AUD, 3%Costs of salaries, staff training, professional development, and staff travel.Healthcare cost-savings: (a) Base case: direct costs (healthcare, police & ambulance, & counseling), indirect costs (productivity loss); (b) SA: the value of a statistical lifeQALYsROI = 1.58 (95% CI: −0.17 – 5.33)ICER = dominant (95% CI: dominant - $156,000/QALY)WTP Threshold: $41,000/QALY gainedHLindkvist (2024), Sweden1 & 5 years2020 EUR, 3%Costs of brief admission: bed occupancyCosts of psychiatric clinic visits, in-patient care, emergency department visitsQALYsAfter 1 year: ICERs = cost-saving (capacity adjusted) - $79,000 (Base case) per QALY gained.All analyses indicated health gains and substantially reduced costs in 5 years.HMartínez-Alés (2021), Spain12 months2019 EUR, no discountTAU costs plus service use costs (phone call, specialized visits).Extra-hospital emergency care, in- & out-patient care. Indirect costs of death & disability leaveSuicide reattempt within 1 year after hospital discharge Time to reattemptICER = $4,000 per averted attemptWTP Threshold: $47,300–67,500/suicide avertedHO’Connor (2017), UK6 months2014 GBP, no discountCosts of health service use & routine implementation, excluding development costs (e.g. salaries)Costs of hospital care, ambulance, treatment within EDs, hospital care following injuryself-harm event avertedVHS is dominant compared to TAU (lower costs and higher effective, although these differences are not statistically significant)WTP threshold: $0HRichardson (2014), US1 month2011 USD, no discountCosts of follow-up calls, reports, admin work, and additional treatmentMental health and substance abuse hospital readmissionsNonePatients discharged from (1) hospital: ROI = 1.76 (commercial insurance), 2.43 (Medicaid); (2) ED: ROI = 1.70 (commercial insurance), 2.05 (Medicaid)HRyan (2015), the US5 years2013 USD, unspecified discountCosts of staff, daily operations, facilities, technology, trainingEmergency department costs, productivity lossesQALYBCR = 0.27ICER = $6,540WTP threshold: $50,000 per QALY gainedHStelmach (2022), Global (36 countries)80 years (Lifetime)2019 USD, 3%Salaries, medications, laboratory tests, computer equipment, material, transportationWage returns to education, productivity gains & lossesUtility weights converted into monetary unit DALY avertedROI = 62ICER = approximately <$110/DALYHAppleby (2000), UKLifetime1998 GBP assumed, no discountMaterial development, training, materials and equipment, and attending trainingSuicide prevented:Life-year gainedICER: $243,000 per suicide prevented and $8,300 per life-year gained.LDunlap (2019), US12 months2015 USD, no discountWage & Salary, Rental costs for hospital spaceSuicide attempts avertedSuicide deaths avertedICER = $6,100 per averted suicide attempt or deathWTP threshold of $7,300HJackson-Morris (2024), UK10–80 years (lifetime)2021 GBP, 3%Salary, equipment (computer, phone call, internet access)Long-term employment and earningsDALYsICER = $10,900 (10 years) & $2,150 (lifetime) /DALY avertedROIs = 0 (10 years) & 2.9 (lifetime)HLatimer (2014), Canada6 months2011 CAD, no discountCosts of RRT intervention and psychiatrist from RRTHealthcare: ED visits, inpatient & outpatient care, medicationSocietal perspective: visits to other hospitals, out-of-pocket healthcare expendituresPatient’s level of functioning (CGAS)Severity of suicidal behavior (SSBS)Hospital perspective: ICER = $14,700 (dominant to $150,000) per point improved on the SSBSSocietal perspective: ICER = $7,800 (dominant to $184,000) per point improved on the SSBSHBotchway (2022), UK12 months2021 GBP, no discountRisk assessment administration, High risk management (e.g. additional 4 clinical reviews)Direct costs of immediate emergency service and health-related.QALYsOxMIS is dominantHDunlap (2019), US12 months2015 USD, no discountWage & Salary, Rental costs for hospital spaceSuicide attempts avertedSuicide deaths avertedICER = $3,400 per averted suicide attempt or death WTP threshold of $7,300HMcDaid (2022), UK2 years2020 GBP, 3.5%Administering psychosocial assessment (No details)Hospital costs of treating self-harm, ambulance costs, productivity lossQALYsICERs = $16,600 (healthcare perspective); $15,100 (societal perspective) per QALY gainedThresholds: $30,000–45,500 per QALY gainedHde Beurs (2015), Netherlands3 months2013 EUR assumed, no discountTraining costs & personnel costsDirect medical costs (visits to a psychiatrist or a psychologist)Indirect medical costs (costs due to sick leave & productivity losses)Suicidal ideation (BSS) QALYs (EQ–5D)Intervention is not cost-effective compared to IAU (ICER = $3,900/point gained on BSS; ICER = $254,000/QALY gained) But cost-effective for suicidal patients if societal WTP > $14,000/point gained on BSSH4. Multi-level interventionsAhern (2018), Global (10 EU countries)12 months2010 EUR, no discountTraining, material adaptation, clinical interview, implementation, photocopySuicide attempt & Severe suicidal ideation (Paykel Hierarchical Suicide Ladder)QALY (SDQ)ICERs per % reduction in suicide attempt = $134 (QPR), $52 (YAM), $78 (ProfScreen); ICERs per QALY gained = $179,187 (QPR), $69,876 (YAM), $95,191 (ProfScreen). WTP Threshold: $69,850/QALY gainedICERs per % point reduction in suicidal ideation = dominated (QPR), $68 (YAM), $152 (ProfScreen); ICERs per QALY gained = dominated (QPR), $71,660 (YAM), $161,684 (ProfScreen). WTP Threshold: $71,300/QALY gained YAM is the most cost-effectiveHCrosland (2024), Australia10 & 30 years2019–20 AUD, 5%Intervention cost: program cost, administration costHealthcare, productivity, and carer costsQALYs, Suicide deaths, and self-harm(1) Healthcare perspective: ICER = $83,000 per QALY gained, $2,800 per self-harm prevented, $76,800 per suicide death prevented; INMB = −1;Societal perspective: ICER = $80,600 per QALY gained, $74,000 per suicide death prevented, $2,700 per self-harm prevented; INMB = −1;Thresholds: $63,700/QALY gained, INMB >0, no threshold for cost/suicide death or self-harmHCrosland (2024), Australia (Continued) 10 & 30 years2019–20 AUD, 5%Intervention cost: program cost, administration costHealthcare, productivity, and carer costsQALYs, Suicide deaths, and self-harm(2) Healthcare perspective: ICERs = dominated, $16.1 million per suicide death prevented, $0.22 million per self-harm prevented; INMB = −92;Societal perspective: ICER = dominated, $19.76 million per suicide death prevented, $0.27 million per self-harm prevented; INMB = −109;Thresholds: $63,700/QALY gained, INMB >0, no threshold for cost/suicide death or self-harmHDoran (2016), Australia5 years2010 AUD, 4.1%No details on implementation costProduction disturbance costs, human capital costs, medical costs, administrative costs, transfer costs, othersPotential suicide and self-harm cases averted (Administrative data)ROI = 4.6AGarraza (2018), USLifetime2010 USD, 3%Staff salaries & benefits, travel, rent, phone, and other overhead expenses.Avoided healthcare costs: inpatient stays, readmission, follow-up, transport, rehabilitationSelf-harm case avoidedROI = 4.5 (95% CI: 1.59, 7.40) per dollar invested ICER = dominantHKinchin (2017), AustraliaLifetime (~40 years)2014 AUD, 4.58%Average cost MIC per worker per yearAvoided production disturbance, human capital, medical, administrative, and other costs (carers, postvention services)Suicide and self-harm cases avertedROI = 1.50 (95% CI 1.11–3.07)HLebenbaum (2020), CanadaLifetime (50 years)2016 CAD, 1.5%Intervention cost per person (no further details)Healthcare costs (hospitalization, ED, and family physicians), non-medical costs (autopsy, police & funeral services), productivity lossesQALYs (EQ–5D)ICER = $19,300 per QALY (Societal perspective); $21,600/QALY (Healthcare payer perspective)Threshold: $51,000 per QALY gainedHStelmach (2022), Global (36 countries)80 years (Lifetime)2019 USD, 3%Costs of salaries, salaries, medications, laboratory tests, computer equipment, materials, transportationWage returns to education, productivity gains & lossesUtility weights converted into monetary unit DALY avertedROI = 3.5 ICER = approximately <$110/DALYHVasiliadis (2015), CanadaLifetime (~60 years)2010 CAD, 3%Cost of prevention programHospitalization & ED visits, outpatient medications, non-medical costs, police & funeral services, productivity losses & short-term disabilityLife-years saved Suicide attempt caseNAD was cost-saving of $4,400 per life-year saved. ICER = $61,000 per case avertedHComplex table summarizing characteristics, interventions, comparators, outcomes, costs, and cost-effectiveness results for studies on suicide and self-harm prevention. See long description. Note: H, high quality; A, average quality; L, low quality. Characteristics of included studies Table 2. long description. Note: H, high quality; A, average quality; L, low quality. Summary findings Universal interventions Universal means restriction: The cost-effectiveness of universal means restriction was inconsistent across four high-quality (Bandara et al., 2022; Damerow et al., 2020; Law & Yip, 2011; Lee et al., 2021) and one low-quality (Whitmer & Woods, 2013) studies. Barriers at bridges were cost-effective over 5–20 years from a partial societal perspective in Australia (Bandara et al., 2022) and the US (Whitmer & Woods, 2013) but not at cliff sites in Australia (Bandara et al., 2022) and railway platforms in Hong Kong (Law & Yip, 2011). Hazardous pesticide access restriction was cost-effective from a healthcare perspective across 14 countries over a lifetime (Lee et al., 2021). In Sri Lanka, a threshold analysis indicated that preventing as few as 0.23 suicides per 100,000 population over 3 years would render pesticide restriction cost-effective (Damerow et al., 2020). Universal awareness training: The cost-effectiveness of universal awareness training varied across three high-quality studies (Jackson-Morris et al., 2024; Kinchin et al., 2020; Sari et al., 2008). The intervention was cost-effective for self-harm prevention in Australia and in the US from a societal perspective (Kinchin et al., 2020; Sari et al., 2008). However, in the UK, Jackson-Morris et al. (2024) reported uncertain cost-effectiveness for self-harm prevention from a partial societal perspective. Universal media guidelines: In Australia, a high-quality study found media guidelines for reporting suicide were cost-effective in preventing self-harm and suicide in the general population over 5 years, compared to no intervention, from a societal perspective (Flego et al., 2022). Selective interventions Selective support services: Support services were cost-effective in preventing self-harm, suicide, and suicidal ideation among individuals displaying risk factors (e.g. individuals facing life challenges) according to five high-quality studies (Comans, Visser, & Scuffham, 2013; Pil et al., 2013; Ross et al., 2021; Sari et al., 2008; Wilson-Barthes et al., 2021). In the US, active follow-up and peer support groups were cost-effective over a lifetime from healthcare and societal perspectives, respectively (Ross et al., 2021; Sari et al., 2008). Telephone and chat helplines also demonstrated cost savings from a societal perspective in Australia (Comans, Visser, & Scuffham, 2013) and Belgium (Pil et al., 2013). In Kenya, institutional care for orphaned children was more cost-effective than family-based care (Wilson-Barthes et al., 2021). Selective psychotherapy: The cost-effectiveness of selective psychotherapy for preventing self-harm, suicide, and suicidal ideation in populations displaying risk factors (e.g. mental illness) was demonstrated in four studies of mixed quality in the US (Gray et al., 2011; Ross et al., 2021; Tebbett-Mock et al., 2020; Zaloshnja et al., 2003). The intervention was cost-saving for adolescents at risk of mental illness over a 1-year period and a lifetime from healthcare and societal perspectives (Gray et al., 2011; Tebbett-Mock et al., 2020; Zaloshnja et al., 2003). However, among primary care patients, the intervention’s cost-effectiveness depended on whether suicide prediction scales had a minimum sensitivity of 75% and specificity of 97% (Ross et al., 2021). Selective medication: The cost-effectiveness of medication for populations displaying risk factors (e.g. mental illness, pregnant women with anemia) showed inconsistent results across five average-to-high quality studies (Dazelle et al., 2026; Duggan et al., 2003; Freemantle et al., 1994; Goren et al., 2016; Groessl et al., 2018). In the US, clozapine and antidepressants were cost-saving over 1–3 years from healthcare (Goren et al. 2016) and societal (Groessl et al., 2018) perspectives, respectively. Similarly, iron supplementation therapy was also cost-effective among pregnant women with anemia over a lifetime from healthcare and societal perspectives (Dazelle et al., 2026). However, in the UK, antidepressants (e.g. selective serotonin reuptake inhibitors) had uncertain cost-effectiveness, being more costly and more effective than older tricyclics, with ICERs of $54,600–$486,000 per life-year gained over a lifetime from a healthcare perspective (Freemantle et al., 1994). A threshold analysis over 10 years also showed a cost per life-year saved of $12,000 if clozapine were cost-neutral, and a cost saving if it reduced annual support costs by 10% from a healthcare perspective (Duggan et al., 2003). Selective health policies: A high-quality study in the US demonstrated that a health policy granting prescriptive authority for psychotropic medications to qualified psychologists was cost-effective in preventing self-harm and suicide among people with mental illness from a societal perspective over 20 years (Hughes et al., 2023). Selective combined intervention: Based on one high-quality study in the UK, the selective combined intervention (including medication, psychotherapy, counselling, and vocational therapies) was cost-saving compared to TAU for individuals with mental disorders over 10 years from a societal perspective (Park, McCrone, & Knapp, 2016). Indicated interventions Indicated psychotherapy: Psychotherapy for individuals with previous self-harm episodes was largely cost-effective in preventing self-harm, suicide, and suicidal ideation across 25 studies, mostly high quality (n = 24) and conducted in HICs (n = 24). Employing a 1-year time horizon, psychotherapy was found cost-effective for preventing self-harm, suicide, and suicidal ideation in adolescents and adults with previous self-harm across six studies, mostly of high-quality, undertaken in the US (Acolin, 2022; Denchev et al., 2018; McCutchan et al., 2022), Pakistan (Alvi et al., 2022), Norway (Haga et al., 2018), and Spain (Martínez-Alés et al., 2021). Specifically, this intervention was either cost-saving (McCutchan et al., 2022) or had acceptable ICERs (Alvi et al., 2022; Denchev et al., 2018; Haga et al., 2018; Martínez-Alés et al., 2021) compared to TAU from healthcare (Alvi et al., 2022; Denchev et al., 2018; Haga et al., 2018; McCutchan et al., 2022) and societal (Martínez-Alés et al., 2021) perspectives in these locations. In the US, dialectical behavior therapy was more cost-effective than cognitive behavior therapy from a healthcare perspective (Acolin, 2022). However, mixed or unfavorable evidence was found in other locations and populations. Across seven high-quality studies in the UK, psychotherapy showed inconsistent cost-effectiveness for adults with previous self-harm (Byford et al., 2003; Haddock et al., 2019; Priebe et al., 2012; Tyrer et al., 2004) and was not cost-effective for adolescents with previous self-harm over 1 year or less (Byford et al., 1999; Green et al., 2011; Tubeuf, Saloniki, & Cottrell, 2019) from healthcare (Haddock et al., 2019; Tubeuf, Saloniki, & Cottrell, 2019) and societal perspectives (Byford et al., 1999, 2003; Green et al., 2011; Priebe et al., 2012; Tyrer et al., 2004). Similarly, three high-quality studies in the Netherlands reported inconsistent cost-effectiveness over a 1-year horizon or less from both healthcare and societal perspectives (Huntjens et al., 2025; Spijker et al., 2012; van Bentum et al., 2026). While psychotherapy demonstrated good value-for-money among adults with previous self-harm (Huntjens et al., 2025) or suicidal ideation (Spijker et al., 2012), it was not cost-effective among psychiatric outpatient adults experiencing suicidal intrusions (van Bentum et al., 2026). Over a shorter time horizon (1–3 months), a high-quality Swedish study found psychotherapy for adolescents with previous self-harm was not cost-effective (Bjureberg et al., 2025). Over 2 years to a lifetime, the cost-effectiveness of psychotherapy for preventing self-harm and suicide among individuals with previous self-harm remained inconsistent across eight high-quality studies. In the US (Bernecker et al., 2020), Canada (Vasiliadis et al., 2017), Australia (Krysinska et al., 2024), and Switzerland (Park et al., 2018), psychotherapy was cost-saving compared to TAU or no intervention from healthcare (Park et al., 2018), payer (Bernecker et al., 2020), partial societal (Krysinska et al., 2024), and societal (Vasiliadis et al., 2017) perspectives. However, in the UK, the cost-effectiveness of psychotherapy compared with TAU was inconsistent across three studies (Bojke et al., 2024; Cottrell et al., 2018; Mavranezouli et al., 2024). Specifically, over a 5-year horizon, one study found psychotherapy cost-effective among adolescents (Cottrell et al., 2018), another demonstrated the intervention was dominated (Bojke et al., 2024), and a third found it cost-effective among adults but not among children (Mavranezouli et al., 2024), all from a healthcare perspective. Indicated support services: Support services were mostly cost-effective compared to TAU or no intervention in preventing self-harm, suicide, and suicidal ideation among individuals with previous self-harm, based on 10 studies of average (Denchev et al., 2018) to high (Borschmann et al., 2013; Gallien et al., 2025; Le et al., 2023; Lindkvist et al., 2024; Martínez-Alés et al., 2021; O’Connor et al., 2017; Richardson, Mark, & McKeon, 2014; Ryan, 2015; Stelmach et al., 2022) quality. Under 1-year time horizon, support services (e.g. postcard/telephone outreach) were cost-effective from payer, healthcare, and societal perspectives in the US (Denchev et al., 2018; Richardson, Mark, & McKeon, 2014), the UK (Borschmann et al., 2013; O’Connor et al., 2017), France (Gallien et al., 2025), and Spain (Martínez-Alés et al., 2021). However, findings from Sweden and the US showed unclear cost-effectiveness of support services, including brief hospital admission by self-referral (Lindkvist et al., 2024) and crisis helpline (Ryan, 2015) over 1 and 5 year(s), respectively. Over 10–80 years, support services (e.g. aftercare, follow-up contact) were dominant in Australia (Le et al., 2023) and across 36 HICs and LMICs (Stelmach et al., 2022) from societal and partial societal perspectives. Indicated suicide risk screening: Three high-quality studies from the US and the UK found suicide risk screening cost-effective for self-harm and suicide prevention among adults with previous self-harm or suicidal ideation in emergency departments from healthcare and societal perspectives over 1–2 year time horizon (Botchway et al., 2022; Dunlap et al., 2019; McDaid et al., 2022). Indicated training for professionals: A high-quality study found training for mental health professionals not cost-effective in preventing suicidal ideation among adults with prior suicidal ideation, but potentially cost-effective for those with depression and suicidal ideation compared to TAU over 3 months from a societal perspective in the Netherlands (de Beurs et al., 2015). Indicated combined interventions: The cost-effectiveness of indicated combined interventions for self-harm and suicide prevention among individuals with previous self-harm compared to TAU or no intervention remained inconsistent across one low-quality (Appleby et al., 2000) and three high-quality studies (Dunlap et al., 2019; Jackson-Morris et al., 2024; Latimer, Gariépy, & Greenfield, 2014) in HICs. In Canada, the combined intervention (assessment, referrals, and treatments) was less costly and less effective over 6 months from healthcare and societal perspectives (Latimer, Gariépy, & Greenfield, 2014). Conversely, in the US, the intervention (screening, safety/treatment plan, hotlines) was cost-effective for adults over 12 months from a program provider perspective (Dunlap et al., 2019). Nevertheless, in the UK, the cost-effectiveness of combined interventions (training professionals, follow-up contact, problem-solving for patients) remained uncertain from healthcare and partial societal perspectives over the lifetime (Appleby et al., 2000; Jackson-Morris et al., 2024). Multi-level interventions Multi-level interventions were mostly cost-effective in preventing self-harm and suicide over long time horizons (10 years to a lifetime) compared to no intervention, based on one average quality (Doran et al., 2016) and seven high-quality studies (Ahern et al., 2018; Crosland et al., 2024; Garraza et al., 2018; Kinchin & Doran, 2017; Lebenbaum et al., 2020; Stelmach et al., 2022; Vasiliadis et al., 2015). In the US, a multi-level intervention (gatekeeper training, hotlines, education, screening) was cost-saving for youths (ages 16–23) with a ROI of 4.5 from a healthcare perspective (Garraza et al., 2018). However, similar interventions in Australia were not cost-effective for children and young people (aged below 25) with psychological distress from healthcare and societal perspectives (Crosland et al., 2024). Nevertheless, a multi-level intervention (awareness trainings, suicide intervention skills) for construction workers in Australia demonstrated positive ROIs of 1.5–4.6 across 5-year and lifetime horizons from a societal perspective (Doran et al., 2016; Kinchin & Doran, 2017). In Canada, two studies also found multi-level interventions (awareness campaigns, training, psychosocial interventions) to be cost-effective (Lebenbaum et al., 2020) or cost-saving (Vasiliadis et al., 2015) compared to no intervention over a lifetime from both healthcare and societal perspectives. Globally, a study across 36 HICs and LMICs found a multi-level intervention (awareness campaigns, training, and screening) cost-effective among adolescents from a societal perspective over a lifetime (Stelmach et al., 2022). However, when components were examined separately in 10 European countries, their cost-effectiveness was uncertain from the education, health, and social care system perspective over 12 months (Ahern et al., 2018). Synthesis of findings As shown in Table 2, the included studies showed high heterogeneity in the intervention characteristics and economic evaluation methods, leading to the use of a dominance ranking framework with three classifications – ‘favored’, ‘unfavored’, and ‘unclear’ (Figure 2). The majority (67%) of studies reported ‘favored’ outcomes. While a few interventions (e.g. universal means restriction, indicated psychotherapy) demonstrated either ‘unclear’ or ‘unfavored’ results, several selective (e.g. psychotherapy, support services) and indicated (e.g. suicide risk screening) interventions consistently generated ‘favored’ cost-effectiveness. Notably, cost-effectiveness varied across countries and health system contexts, with several interventions appearing more consistently cost-effective in the US than in countries with universal healthcare systems (e.g. Australia, the UK), likely reflecting higher baseline healthcare costs and greater potential for cost offsets.Figure 2.Cost-effectiveness results and implications for decision makers. Figure 2. long description.A color-coded table summarizing the number and outcome of studies on suicide prevention interventions by strategy and type. See long description. Cost-effectiveness results and implications for decision makers. Figure 2. long description. Although psychotherapy was the most frequently evaluated indicated intervention (n = 26), its cost-effectiveness remained inconsistent across the studies. Most studies (n = 15) reported ‘favored’ cost-effectiveness for psychotherapy in many countries (e.g. Australia, the US, Canada), though the results remained ‘unclear’ for adults and ‘unfavored’ for adolescents in the UK. In cases of ‘unclear’ cost-effectiveness, the primary reason was the absence of defined cost-effectiveness thresholds for outcome-specific measures (i.e. cost per suicide case averted). Observations from interventions evaluated by multiple studies across three 10-year periods (1995–2025) revealed notable cost-effectiveness trends. Selective support services (Comans, Visser, & Scuffham, 2013; Pil et al., 2013; Ross et al., 2021; Sari et al., 2008) and selective psychotherapy (Gray et al., 2011; Ross et al., 2021; Zaloshnja et al., 2003) consistently demonstrated ‘favored’ cost-effectiveness across all periods. The cost-effectiveness of selective medication improved over time, shifting from ‘unclear’ in earlier periods (Duggan et al., 2003; Freemantle et al., 1994) to ‘favored’ in the most recent period (Goren et al., 2016; Groessl et al., 2018). Nevertheless, indicated psychotherapy exhibited inconsistent cost-effectiveness outcomes across multiple studies at different periods (Haddock et al., 2019; Spijker et al., 2012; Tubeuf, Saloniki, & Cottrell, 2019; Tyrer et al., 2004). Quality assessment As shown in Table 3, most studies were of high quality (n = 61, 88%), with the majority of these (n = 47, 77%) published between 2015 and 2025, demonstrating an increase in publication quality and quantity over time. However, a few limitations persisted across several studies. Uncertainty and sensitivity analyses were often neglected or poorly conducted in many studies (e.g. Gray et al., 2011; Tebbett-Mock et al., 2020). Some studies inadequately addressed the generalizability of their findings and barriers to implementation (e.g. Tubeuf, Saloniki, & Cottrell, 2019; Whitmer & Woods, 2013). Additionally, the range of costs analyzed did not fully align with the stated perspective in certain studies (e.g. Appleby et al., 2000; Latimer, Gariépy, & Greenfield, 2014).Table 3.Results of quality assessment using Drummond’s 10-item checklist Table 3. long description.#Lead author (year)1. Was a well-defined question posed in answerable form?2. Was a comprehensive description of the competing alternatives given?3. Was the effectiveness of the program or services established?4. Were all the important and relevant costs and consequences for each alternative identified?5. Were costs and consequences measured accurately in appropriate physical units?6. Were the costs and consequences valued credibly?7. Were costs and consequences adjusted for differential timing?8. Was an incremental analysis of costs and consequences of alternatives performed?9. Was allowance made for uncertainty in the estimates of costs and consequences?10. Did the presentation and discussion of study results include all issues of concern to users?TotalQuality1Acolin (2022)11110.51110.519High quality2Ahern (2018)111111111110High quality3Alvi (2022)111111111110High quality4Appleby (2000)10.50.50.50001014.5Low quality5Bandara (2022)111111111110High quality6Bernecker (2020)111111111110High quality7Bjureberg (2025)111111111110High quality8Bojke (2024)111110.511119.5High quality9Borschmann (2013)111111110.50.59High quality10Botchway (2022)111111111110High quality11Byford (1999)111111111110High quality12Byford (2003)1111111110.59.5High quality13Comans (2013)110.51110.51119High quality14Cottrell (2018)1111110.51119.5High quality15Crosland (2024)111111111110High quality16Damerow (2020)111111111110High quality17Dazelle (2026)110.510.510.51118.5High quality18de Beurs (2015)111110.511119.5High quality19Denchev (2018)0110.510.51110.57.5Average quality20Doran (2016)0.50.50.50.511110.50.57Average quality21Duggan (2003)0.50.510.500110.50.55.5Average quality22Dunlap (2019)111111110.519.5High quality23Flego (2022)1111110.51119.5High quality24Freemantle (1994)00.510110.510.516.5Average quality25Gallien (2025)0.5110.50.5111118.5High quality26Garraza (2018)111111111110High quality27Goren (2016)1111111110.59.5High quality28Gray (2011)0110.5001100.55Low quality29Green (2011)111111111110High quality30Groessl (2018)111111110.519.5High quality31Haddock (2019)110.51111110.59High quality32Haga (2018)111111111110High quality33Hughes (2023)111111111110High quality34Huntjens (2025)111111111110High quality35Jackson-Morris ( 2024)0.5110.51111119High quality36Kinchin (2017)110.50.511110.50.58High quality37Kinchin (2020)110.50.511110.518.5High quality38Krysinska (2024)111111111110High quality39Latimer (2014)1110.5111110.59High quality40Law (2011)1110.5111100.58High quality41Le (2023)1111110.51119.5High quality42Lebenbaum (2020)111111111110High quality43Lee (2021)111111111110High quality44Lelie (2025)1111110.51119.5High quality45Lindkvist (2024)111111111110High quality46Martínez-Alés ( 2021)111111111110High quality47Mavranezouli (2024)111111111110High quality48McCutchan (2022)111111110.519.5High quality49McDaid (2022)111111111110High quality50O’Connor (2017)111111111110High quality51Park (2016)111111110.519.5High quality52Park (2018)111111111110High quality53Pil (2013)110.51111110.59High quality54Priebe (2012)111110.5110.519High quality55Richardson (2014)1110.5111110.59High quality56Ross (2021)111111111110High quality57Ryan (2015)110.511101118.5High quality58Sari (2008)111110.5110.519High quality59Stelmach (2022)111111111110High quality60Tebbett-Mock ( 2020)011010.511016.5Average quality61Tubeuf (2019)11111111109High quality62Tyrer (2004)11111110.50.508High quality63vanBentum (2025)1111110.51119.5High quality64Spijker (2012)111111111110High quality65Vasiliadis (2015)1111110.510.519High quality66Vasiliadis (2017)110.5110.501118High quality67Whitmer (2013)010.50.50001003Low quality68Wilson-Barthes ( 2021)111111111110High quality69Zaloshnja (2003)110.51111100.58High qualityA data table summarizing quality assessment scores for 69 studies using Drummond’s 10-item checklist, with total scores and quality ratings ranging from low to high quality. See long description. Table 3.Results of quality assessment using Drummond’s 10-item checklist Table 3. long description.#Lead author (year)1. Was a well-defined question posed in answerable form?2. Was a comprehensive description of the competing alternatives given?3. Was the effectiveness of the program or services established?4. Were all the important and relevant costs and consequences for each alternative identified?5. Were costs and consequences measured accurately in appropriate physical units?6. Were the costs and consequences valued credibly?7. Were costs and consequences adjusted for differential timing?8. Was an incremental analysis of costs and consequences of alternatives performed?9. Was allowance made for uncertainty in the estimates of costs and consequences?10. Did the presentation and discussion of study results include all issues of concern to users?TotalQuality1Acolin (2022)11110.51110.519High quality2Ahern (2018)111111111110High quality3Alvi (2022)111111111110High quality4Appleby (2000)10.50.50.50001014.5Low quality5Bandara (2022)111111111110High quality6Bernecker (2020)111111111110High quality7Bjureberg (2025)111111111110High quality8Bojke (2024)111110.511119.5High quality9Borschmann (2013)111111110.50.59High quality10Botchway (2022)111111111110High quality11Byford (1999)111111111110High quality12Byford (2003)1111111110.59.5High quality13Comans (2013)110.51110.51119High quality14Cottrell (2018)1111110.51119.5High quality15Crosland (2024)111111111110High quality16Damerow (2020)111111111110High quality17Dazelle (2026)110.510.510.51118.5High quality18de Beurs (2015)111110.511119.5High quality19Denchev (2018)0110.510.51110.57.5Average quality20Doran (2016)0.50.50.50.511110.50.57Average quality21Duggan (2003)0.50.510.500110.50.55.5Average quality22Dunlap (2019)111111110.519.5High quality23Flego (2022)1111110.51119.5High quality24Freemantle (1994)00.510110.510.516.5Average quality25Gallien (2025)0.5110.50.5111118.5High quality26Garraza (2018)111111111110High quality27Goren (2016)1111111110.59.5High quality28Gray (2011)0110.5001100.55Low quality29Green (2011)111111111110High quality30Groessl (2018)111111110.519.5High quality31Haddock (2019)110.51111110.59High quality32Haga (2018)111111111110High quality33Hughes (2023)111111111110High quality34Huntjens (2025)111111111110High quality35Jackson-Morris ( 2024)0.5110.51111119High quality36Kinchin (2017)110.50.511110.50.58High quality37Kinchin (2020)110.50.511110.518.5High quality38Krysinska (2024)111111111110High quality39Latimer (2014)1110.5111110.59High quality40Law (2011)1110.5111100.58High quality41Le (2023)1111110.51119.5High quality42Lebenbaum (2020)111111111110High quality43Lee (2021)111111111110High quality44Lelie (2025)1111110.51119.5High quality45Lindkvist (2024)111111111110High quality46Martínez-Alés ( 2021)111111111110High quality47Mavranezouli (2024)111111111110High quality48McCutchan (2022)111111110.519.5High quality49McDaid (2022)111111111110High quality50O’Connor (2017)111111111110High quality51Park (2016)111111110.519.5High quality52Park (2018)111111111110High quality53Pil (2013)110.51111110.59High quality54Priebe (2012)111110.5110.519High quality55Richardson (2014)1110.5111110.59High quality56Ross (2021)111111111110High quality57Ryan (2015)110.511101118.5High quality58Sari (2008)111110.5110.519High quality59Stelmach (2022)111111111110High quality60Tebbett-Mock ( 2020)011010.511016.5Average quality61Tubeuf (2019)11111111109High quality62Tyrer (2004)11111110.50.508High quality63vanBentum (2025)1111110.51119.5High quality64Spijker (2012)111111111110High quality65Vasiliadis (2015)1111110.510.519High quality66Vasiliadis (2017)110.5110.501118High quality67Whitmer (2013)010.50.50001003Low quality68Wilson-Barthes ( 2021)111111111110High quality69Zaloshnja (2003)110.51111100.58High qualityA data table summarizing quality assessment scores for 69 studies using Drummond’s 10-item checklist, with total scores and quality ratings ranging from low to high quality. See long description. Results of quality assessment using Drummond’s 10-item checklist Table 3. long description. Discussion This review provides the most comprehensive analysis of existing economic evidence on self-harm and suicide prevention, covering a wide range of interventions and identifying significant growth in quality and quantity of publications, particularly in the past decade. Despite high heterogeneity, around 67% of the studies found suicide prevention to be cost-effective or cost-saving. Most cost-effectiveness evidence supported universal (e.g. awareness training), selective (e.g. support services), indicated (e.g. suicide risk screening, support services, psychotherapy for adults in Australia, the US, Canada), and multi-level interventions. However, our review also shows limited cost-effectiveness evidence of self-harm and suicide prevention in children, adolescents, and especially older adults, despite the high prevalence of such phenomena in these age groups (WHO, 2023). Similarly, there is significantly limited evidence from LMICs, which accounted for more than two-thirds of total suicides worldwide in 2021 (WHO, 2023). Policy implications and directions for future research More economic evaluations are urgently needed, particularly in the underrepresented populations and regions mentioned above, to accelerate the implementation of cost-effective interventions. Critically, scaling up cost-effective interventions requires careful consideration of important local factors (Lee et al., 2021) because what works (or is cost-effective) in one context might not work (or be cost-effective) in another. For instance, system inefficiencies and higher baseline healthcare costs in the US may increase the potential for cost offsets from prevention, which may partly explain why preventive interventions often appear more cost-effective in US-based studies than in countries with universal healthcare. Furthermore, large-scale implementation requires a multidisciplinary approach impacting multiple sectors (WHO, 2021), which can lead to fragmentation of responsibility and funding across stakeholders. Methodological issues of current evidence While most studies were high-quality, heterogeneity in outcome measures and comparators still exists, making direct comparisons and firm conclusions difficult across studies. Although suicide impacts broader society, many studies did not adopt a societal perspective, underestimating costs related to lost productivity, social services, and burden on families and communities (Werdin & Wyss, 2024). Additionally, inadequate discussion of broader implementation issues (e.g. generalizability and implementation barriers) and unclear cost-effectiveness thresholds, especially for clinical outcomes, might create hesitation and sub-optimal outcomes while implementing the interventions in the real-world. Strengths and limitations To our knowledge, this is the first comprehensive and up-to-date review of the cost-effectiveness of self-harm and suicide interventions that employed broad inclusion criteria, including ROI studies. These ROI studies, accounting for 10% of the studies included, showed strong returns for self-harm and suicide interventions. However, the present review may have missed unpublished studies, studies published in grey literature (due to our limited search of grey literature), and studies written in non-English languages, which may underestimate the cost-effectiveness evidence from LMICs. In addition, the available evidence may be influenced by publication and commissioning practices. Nevertheless, the inclusion of studies reporting unfavorable or inconclusive cost-effectiveness findings suggests that negative economic results are not systematically excluded from the peer-reviewed literature. Conclusions This review found strong and promising evidence for the cost-effectiveness of self-harm and suicide interventions, including several universal (e.g. awareness training), selective (e.g. psychotherapy, support services), indicated (e.g. suicide risk screening, support services, and psychotherapy for adults in countries including Australia, the US, Canada), and multi-level interventions. However, research on children, adolescents, and especially older adults remains limited, and evidence from LMICs, where most suicides occur, is also scarce. More economic evaluations for underrepresented populations and regions, as well as translational research on real-world implementation of suicide prevention, are urgently needed. Supporting information Le et al. supplementary material Le et al. supplementary material