BMC Public Health volume 25, Article number: 1206 (2025) Cite this article
Gatekeeper training (GKT) aims to enhance suicide gatekeepers’ (GKs) abilities in assessing suicide risk, identify those at-risk and refer them. However, existing randomized controlled trials (RCTs) on GKT have not produced definitive results. This study reviewed RCTs on GKT to provide evidence for developing effective suicide prevention strategies.
We conducted a systematic search of MEDLINE, PubMed, Web of Science, PsycINFO, CINAHL, Embase, Google Scholar, Medrxiv, and Ebsco for English-language RCTs published between January 1, 2000, and December 31, 2024. Two authors independently screened studies, extracted data, and assessed the risk of bias using the Cochrane risk-of-bias tool. Standardized mean differences (SMD) with 95% confidence intervals were calculated using a random-effect model. Heterogeneity was assessed by using I2 statistic, and publication bias was evaluated through funnel plots and Egger's regression. We stratified participants into subgroups by characteristics and categorized interventions by delivery mode (online vs. offline). Post-test and follow-up data were integrated into a unified model, with follow-ups classified as short-term (1–5 months) or long-term (> 5 months). All analyses were performed using R version 3.4.0, following the PRISMA guidelines (registration number: CRD42024507513).
Sixteen studies were included. Compared to the control group, gatekeepers showed increased suicide knowledge (SMD = 0.72, 95% CI: 0.32 – 1.13) and enhanced self-efficacy (SMD = 0.73, 95% CI: 0.33 – 1.13) for suicide prevention. For knowledge, the improvements were sustained in the short-term (SMD = 0.64, 95% CI: 0.22 – 1.06) but diminished in the long-term (SMD = 0.25, 95% CI: 0.05 – 0.45). Online interventions showed a significant improvement in self-efficacy (SMD = 1.02, 95% CI: 0.73 – 1.32), while offline interventions demonstrated a potential but non-significant improvement (SMD = 0.53, 95% CI: -0.08 – 1.17). Preparedness also showed a significant improvement (SMD = 0.69, 95% CI: 0.31 – 1.07).
This meta-analysis demonstrated GKT’s effectiveness in enhancing knowledge and self-efficacy (moderate effect) for suicide prevention. However, the positive effects tend to decrease over time. Online training offers significant benefits, making it a viable option for widespread implementation. Additionally, the selection of gatekeepers plays a crucial role in ensuring the effectiveness of online GKT programs.
Suicide is a global public health concern, with over 800,000 people dying by suicide each year [1]. An even greater number of people have considered or attempted suicide [2]. Suicide poses a significant threat to life safety, resulting in substantial economic losses and profound impacts on families and society [3]. However, by enhancing the ability to identify specific risk groups and potential suicide attempts, and by developing effective responses, the risk of suicide can be significantly reduced. Preventing suicide has become an important goal and a valuable indicator of global progress in promoting mental health and preventing mental disorders. Many countries are developing suicide prevention strategies that require up-to-date, high-quality evidence [4]. Research has shown that the Gatekeeper Training (GKT) program is an effective suicide prevention intervention [5] and one of the most common interventions in suicide prevention [6,7,8].
Before attempting suicide, individuals often send distress signals, such as verbal language, written messages, and behaviors to those around them and need time to decide to proceed with the act. It is during this critical period that gatekeepers can detect suicidal intent through these signs and provide timely help and referrals, thereby reducing the risk of suicide [9]. A qualified gatekeeper (GK) can identify whether someone may have suicidal tendencies. Anyone can contribute to suicide prevention by recognizing early warning signs in others and offering assistance, such as teachers, friends, family members, co-workers, and community members [10]. In particular, teachers, health care workers, and family members are more likely to be in contact with people at high risk of suicide. GKT programs aim to develop participants’ knowledge, attitudes, and skills for identifying individuals at risk, determining their level of risk, and then referring at-risk individuals for preventive treatment [1]. Training programs usually include the following components: raising awareness, increasing knowledge on suicidality, teaching intervention skills, and informing about local resources, and referral points [11].GKT programs typically consist of brief training sessions focused on suicide prevention [12].
As early as Isaac’s (2009) systematic review, it was found that GKT is associated with significant improvements in GKs’ knowledge and attitudes towards suicide prevention [13]. Subsequent studies have also demonstrated that GKT can enhance GKs’ understanding of suicide [14] and improve their suicide prevention skills [15, 16]. However, few studies have examined whether these effects can be sustained over time. Additionally, with the advancement of internet technology, online training methods are becoming increasingly prevalent. However, its effectiveness compared to traditional methods remains unclear.
To our knowledge, no meta-analysis has specifically examined the long-term follow-up effectiveness of GKT using data from RCTs. While previous systematic reviews have included RCTs, they have not quantitatively analyzed the differences between post-test and follow-up outcomes [7, 12, 17,18,19,20]. This meta-analysis aims to address the existing research gap by systematically examining both the immediate and sustained effects of GKT, thereby offering insights into the durability of its outcomes. Furthermore, recent studies, particularly those on online training, have been included to evaluate its effectiveness in comparison to traditional offline methods. A recent review of systematic reviews on GKT proposed that the lack of RCTs is a major limitation in the assessment of GKT and that more RCT studies are needed to confirm these findings [8]. In addition, the heterogeneity of GKT training approaches(e.g., training modules, duration, and delivery methods) and outcome measures, as well as the absence of standardized evidence levels in the literature, limits conclusions about the current effectiveness of GKT [4]. In recent years, there has been an increase in investigational studies related to GKT. Therefore, this review aims to include these RCTs in a comprehensive quantitative analysis. It will focus on attitudinal and behavioral changes in suicide GKs, incorporating the most recent evidence to support the development of effective suicide prevention strategies.
This project was registered in PROSPERO (CRD42024507513) and followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guidelines [21].
A search for relevant RCT studies was conducted from January 1, 2000, to December 31, 2024, in MEDLINE, PubMed, Web of Science, PsycINFO, CINAHL, Embase, Google Scholar, Medrxiv, and Ebsco. Boolean logic searches were performed using the keywords (suicide prevention, prevention, suicide, suicide awareness, awareness, suicide) and (gatekeeper) and (randomized controlled trials). During the search process, I search terms were applied search terms primarily within the full text and filters were applied to limit the timeframe to between January 1, 2000 and July 1, 2024 (S1 Table in the Supplement).
This review focused on integrated interventions for GKT, defined as training programs that combine multiple components (e.g., psychoeducation, skill-building, and role-playing) to enhance gatekeepers' ability to identify and support individuals at risk of suicide. Typical GKT sessions cover how to identify at-risk individuals, discuss suicide, ask about suicidal thoughts, and encourage seeking appropriate mental health care or crisis services [12].
Only RCTs related to GKT were included, with no restrictions on participants or study populations. The control groups typically received either no intervention, waitlist conditions, or alternative interventions (e.g., general mental health education). Two reviewers independently conducted article selection and quality assessment using a standardized screening protocol and the Cochrane risk-of-bias tool [17]. Disagreements were resolved through discussion or consultation with a third reviewer. The reference lists of identified articles and previous relevant systematic reviews were also checked to ensure comprehensiveness.
The main outcomes included self-efficacy (confidence in identifying and supporting at-risk individuals), knowledge (understanding of suicide risk factors), attitudes (beliefs and perceptions toward suicide), behaviors (frequency of suicide prevention actions), preparedness (readiness to intervene), stigma (negative attitudes toward suicide), and skill (ability to conduct risk assessments). The outcome measures were continuous, and although the measurement tools varied across studies, we ensured all tools were aligned in the same direction before calculating standardized mean differences (SMD) with 95% confidence intervals (CI) to pool effect sizes. Higher scores on all scales indicated better outcomes, and we reversed scoring if necessary to maintain uniformity. The design effect (DE = 1 + (m—1) * ICC) will be applied to adjust the sample size and variance estimates for cluster-randomized trials, ensuring comparability with traditional RCTs (where m represents the average cluster size and ICC (intra-cluster correlation coefficient) measures the similarity of outcomes within clusters). To ensure consistency in results, a random-effects model will be used to account for both within-study and between-study heterogeneity. For outcomes with fewer than 3 studies, meta-analysis will not be performed due to insufficient data.
We analyzed post-test data and stratified participants into subgroups based on their characteristics: specific populations (such as family members of individuals with mental illness, people dealing with youth issues, and nurses), general populations (representing the broader society, encompassing individuals without specific vulnerabilities or roles related to suicide prevention), students, and school workers. Additionally, we compared interventions based on delivery mode (online vs. offline). To assess the sustainability and time-varying efficacy of the GKT effect, we integrated post-test and follow-up data into a unified model to validate the effectiveness over time. Follow-up periods were categorized as short-term (1–5 months) and long-term (> 5 months).
Heterogeneity between studies was assessed using I2statistic, classified as low (I2 ≤ 25%), moderate (25% < I2 ≤ 50%), or high (I2 > 50%) [18]. High heterogeneity necessitated sensitivity analyses, which involved excluding studies with a high risk of bias. The random-effects model, which accounts for errors at both the sampling and study levels, was deemed more suitable for this scenario [19]. Publication bias was evaluated using funnel plots or Egger’s regression, with the Trim and Fill method used to adjust for potential bias [20, 22]. All data were analyzed using R version 3.4.0.
A total of 7,766 articles were retrieved from the database. After excluding duplicates (n = 22), 7,744 publications were screened for inclusion. By reviewing the titles and abstracts, 7,710 studies were excluded. These studies were excluded based on the following criteria: studies that did not focus on suicide prevention; studies that were not randomized controlled trials (RCTs); studies that did not involve GKT interventions. Out of the 34 full-text articles assessed for eligibility, 16 studies provided sufficient data for meta-analysis. The search process and selection phases followed the PRISMA protocol flowchart, as shown in Fig. 1.
Flow chart of literature search strategy
The studies included in this review were published between 2000 and 2024 (Table 1). Six studies were from the USA [6, 22,23,24,25,26], two from Canada [27, 28],two from Korea [29, 30],two from Germany [31, 32], Australia [33],the Netherlands [34], Japan [35] and Taiwan [36] each contributed one study. Participants included school workers [6, 22], all students enrolled in schools, colleges and universities as well as graduate students [23, 24, 26, 32, 35], specific population (nurses in a general hospital [36],pastors, and Christian counselors [33], and veterans contacts [25], family members of patients with mental disorders [29], people dealing with youth issues [27, 34]), general population( recruited from the community) [28, 30, 31]. The duration of follow-up was mainly no follow-up [24, 36], short-term follow-up [6, 23, 26, 29,30,31, 33,34,35] and long-term follow-up [22, 25, 27, 28, 32].
Most studies demonstrated a low risk of bias in random sequence generation and allocation concealment. However, there were some concerns regarding the blinding of participants and personnel, as well as incomplete outcome data. Specifically, the inability to blind participants and personnel in GKT studies was a common limitation due to the nature of the intervention. (S3 Table and S4 Figure in the Supplement).
The pooled SMD for knowledge about suicide prevention in post-test data was 0.72 (95% CI: 0.32 – 1.13). However, significant heterogeneity was observed (I2 = 90%, p < 0.01). Sensitivity analysis revealed that excluding Hill’ study [24] increased the SMD to 0.73 (95% CI: 0.59 – 0.87). For follow-up data, the improvement in knowledge was sustained in the short-term (SMD = 0.64, 95% CI: 0.22 – 1.06), but diminished in the long-term (SMD = 0.25, 95% CI: 0.05 – 0.45). Offline interventions showed a moderate effect size (SMD = 0.68, 95% CI: 0.17 – 1.19). The specific population showed a large improvement (SMD = 1.10, 95% CI: 0.58 – 1.63). (Fig. 2, Tables 2 and 3, S5 Figure. a in the Supplement).
GKT Forest plot of the effect on knowledge
Egger’s regression test indicated no significant publication bias for post-test data (Bias estimate = 5.34, SE = 2.76, p = 0.09). Visual inspection of the funnel plot suggested potential asymmetry. The Trim and Fill method yielded an adjusted SMD of 0.13 (95% CI: -0.40 – 0.66; p = 0.63). Although the adjusted effect size was not statistically significant (p = 0.63), the confidence interval included the original pooled effect size (SMD = 0.72). (S7 Table, S8 Figure. cd and S9 Table in the Supplement).
The pooled SMD for self-efficacy was 0.73 (95% CI: 0.33 – 1.13). However, significant heterogeneity was observed (I2 = 89%; p < 0.01). Sensitivity analysis revealed that excluding individual studies did not substantially alter the overall effect size. (Fig. 3, S5 Figure. b in the Supplement). For follow-up data, the improvement in self-efficacy was sustained in the short-term (SMD = 0.86, 95% CI: 0.51–1.20), but diminished in the long-term (SMD = 0.69, 95% CI: 0.04–1.34). Online interventions showed a large effect size (SMD = 1.02, 95% CI: 0.73 –1.32). A moderate but significant improvement in self-efficacy among students (SMD = 0.66, 95% CI: 0.11–1.20). (Tables 2 and Tables 3).
GKT Forest plot of the effect on self-efficacy
Egger’s regression test indicated no significant publication bias for post-test data (Bias estimate = 3.43, SE = 3.48, p = 0.37). Visual inspection of the funnel plot suggested potential asymmetry. The Trim and Fill method yielded an adjusted SMD of 0.44 (95% CI: -0.05 – 0.92; p = 0.08). Although the adjusted effect size was not statistically significant (p = 0.08), the confidence interval included the original pooled effect size (SMD = 0.73). (S7 Table, S8 Figure. cd and S9 Table in the Supplement).
The pooled SMD for skills using the random-effects model was 0.62 (95% CI: -0.47–1.70), there was no statistically significant improvement in skills among gatekeepers following GKT. Significant heterogeneity was observed (I2 = 93%, p < 0.01). In the short-term follow-up, the effect size was small and not statistically significant. (Fig. 4, Tables 2 and 3).
GKT Forest plot of the effect on skill
The pooled SMD for preparedness using the random-effects model was 0.69(95% CI: 0.31 –1.07) with significant heterogeneity (I2 = 77%, p < 0.01). In the short-term follow-up, the effect size was large and statistically significant (SMD = 0.97, 95% CI: 0.42 – 1.53). For students, the SMD ranged from 0.20 to 0.80, while offline interventions demonstrated a small improvement with an SMD of 0.38 (95% CI: 0.18 – 0.57) and very low heterogeneity (I2 = 0%, p = 0.50). (Fig. 5, Tables 2 and 3).
GKT Forest plot of the effect on preparedness
Egger’s regression test indicated no significant publication bias for preparedness (Bias estimate = 3.26, SE = 2.19, p = 0.23). Visual inspection of the funnel plot suggested potential asymmetry. The Trim and Fill method yielded an adjusted SMD of 0.41 (95% CI: -0.05 – 0.88; p = 0.09). Although the adjusted effect size was not statistically significant (p = 0.09), the confidence interval included the original pooled effect size (SMD = 0.69). (S7 Table, S8 Figure, and S9 Table in the Supplement).
To our knowledge, this is the first and most comprehensive pooled analysis of RCTs related to GKT, encompassing a diverse and broad population without restriction to specific population. Preliminary data suggested that GKT can enhance participants’ knowledge of suicide prevention, improve self-efficacy and preparedness, and reduce stigma toward suicide. However, definitive evidence on its effects on attitudinal and behavioral outcomes remains lacking [7, 13, 47, 48]. Because acquiring knowledge is relatively straightforward, changing attitudes is more challenging. Attitudes reflect deeper values and may require extended durations to achieve significant change. Short training sessions (e.g., 1 -2 h) may be insufficient to induce meaningful attitudinal shifts [29]. Insufficient attitudinal changes can undermine the sustainability of GKT outcomes, as improved knowledge and skills in suicide prevention do not always translate into behavioral change. The short follow-up periods of most studies may not be long enough to capture significant behavioral changes among participants [49]. To ensure a lasting impact on suicide prevention behaviors, future GKT programs should focus more on altering attitudes, as this is a crucial factor in achieving sustained behavioral change [47].
Our study confirms that GKT is effective in enhancing both knowledge and self-efficacy, consistent with previous research [7]. However, we observed a significant decline in suicide knowledge scores from post-test to follow-up, indicating that knowledge gained may diminish over time. This novel finding suggests that acquired knowledge tends to decline over time if not regularly applied [50]. Although the effectiveness of GKT decreases over time, its importance is not diminished. Similar to other forms of skills training, GKT should be periodically refreshed to maintain its impact [50, 51]. Strategies such as spaced learning, which involves distributing training over time rather than concentrating it in a single phase, and practice-based interventions, have been identified as effective methods to enhance training outcomes [52].
The results indicated that both online and offline delivery methods have an impact on GKT. Specifically, online methods appear to be more effective in improving self-efficacy among trainees. Innovative training modalities are essential as advancements in science and technology increasingly influence GKT. Recent years have seen a growing reliance on online technologies for GKT [26, 53,54,55]. Online training offers several benefits: it overcomes time and space constraints, providing accessibility and flexibility that benefit gatekeepers with limited resources [34]. Online training utilizes flexible, user-friendly and relatively inexpensive internet technologies [24, 34, 56]. The online training can accommodate for its flexibility and scalability. This format has proven effective in reducing training costs, as demonstrated during the COVID-19 pandemic when remote meetings enabled interaction with national suicide prevention experts without incurring high travel costs [25]; Additionally, online training can enhance privacy and engagement by addressing concerns associated with sharing sensitive content in traditional settings. Some participants may be reluctant to participate in this program due to privacy concerns. Online GKT courses may help overcome these obstacles to promote more people's participation [53]; Studies have shown that online training can be as effective as face-to-face training, sometimes even reducing trainee reluctance to intervene more effectively [57]. It also increases learner control over the training process and extends accessibility to individuals who previously had limited opportunities for suicide prevention training [7, 58]. However, online training does have limitations: participants may lack foundational knowledge, be less adequately prepared, and show lower efficiency or exhibit reluctance to assume the role of gatekeepers. There is also a shortage of comprehensive videos offering detailed implementation guidance or targeted content for specific demographics [59]. Therefore, combining online and offline training modalities can significantly enhance the overall effectiveness of GKT.
The results indicated that the effectiveness of GKT varies depending on the participants, highlighting the need for careful consideration in selecting individuals for the program. Preliminary analysis suggested that targeted training for specific professional groups, such as healthcare workers, teachers, and first responders, may yield higher benefits as gatekeepers. This is likely due to their frequent interactions with high-risk individuals and their ability to apply GKT skills in their daily roles. In contrast, the general population may derive less immediate benefit. Therefore, given the challenging nature of the gatekeeper role, individuals must be selected thoughtfully, choosing more suitable trainees for the program. GKs should possess the capacity to effectively support suicide prevention efforts for at-risk populations. They need to have the skills and sense of responsibility required to intervene and take action to prevent suicide [60]. Previous studies have often overlooked the importance of participant selection. Future research should address aspects such as the suitability of trainers, the selection criteria for trainees, and the content of the training. Additionally, it is crucial to evaluate training tailored to specific backgrounds or populations [4]. Equally important is the consideration of health support for GKs. Establishing a robust support system, such as a buddy system for GKs to train together and regular meetings to share experiences and technical resources, is essential [6].
Some limitations of this study should be noted. Significant heterogeneity was observed among the included studies, indicating substantial variability in results. The limited number of studies for certain outcomes, particularly in subgroup analyses, may restrict the reliability and generalizability of the findings. Future research should include more high-quality studies to validate these findings and enhance the robustness of the analysis. Methodological challenges were also identified, particularly the difficulty in achieving blinding of participants and personnel. Given that GKT involves active participation and skill-building, blinding is often impractical, potentially introducing performance bias. Despite this limitation, the consistency of results across studies with varying levels of blinding suggests that the impact of performance bias may be mitigated. Researchers should explore innovative approaches, such as using objective outcome measures or independent assessors, to address these challenges. Addressing these limitations, particularly those related to blinding and incomplete data, will be crucial for advancing our understanding of GKT's effectiveness and its long-term impact on suicide prevention behaviors, ultimately informing the development of more effective strategies.
Based on the meta-analysis of RCTs, this review provides robust evidence reaffirming the effectiveness of GKT programs in enhancing knowledge, self-efficacy, and preparedness. However, the positive effects on knowledge and self-efficacy tend to diminish over time. Online training has gained popularity as a delivery method for GKT, and our analysis compared its effectiveness with traditional in-person approaches. The selection of training participants is critical, given that the effectiveness of GKT varies across different populations. To maximize outcomes, we recommend that future GKT programs incorporate regular sessions, extended durations, innovative modalities, and more targeted participant selection, thereby fostering lasting attitude changes and promoting sustained intervention behaviors.
All data generated or analyzed in this study are included in this manuscript. This means that a list of all studies included in this review is provided as supplementary material.
Not applicable.
This project was supported by Senior Talent Startup Fund of Nanchang University [grant number28170120/9167] and Postgraduate Innovation Special Fund of Jiangxi Province [grant number YC2023–S181].
Not applicable.
Not applicable.
The authors declare no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Liu, H., Zheng, C., Cao, Y. et al. Gatekeeper training for suicide prevention: a systematic review and meta-analysis of randomized controlled trials. BMC Public Health 25, 1206 (2025). https://doi.org/10.1186/s12889-025-21736-1
