Development and temporal validation of a screening-based prediction model for suicidal ideation at first contact with university counseling services.

Development and temporal validation of a screening-based prediction model for suicidal ideation at first contact with university counseling services

Background

Suicidal ideation among university students is a major public health concern. Yet counseling services still lack practical tools that can identify students at elevated risk during their first clinical contact. This study aimed to develop and temporally validate a screening‑based prediction model for suicidal ideation at the time of initial presentation to university counseling services.

Methods

This retrospective observational study analyzed routinely collected data from a Japanese university counseling services. Predictor variables were exclusively derived from annual mental health screening data collected before participants used counseling services. The variables included psychological distress, self‑esteem, lifestyle‑related factors, and demographic characteristics. The outcome was suicidal ideation assessed during the first counseling intake interview. We developed a multivariable logistic regression model using a development cohort of students who accessed counseling services between 2017 and 2019 and temporally validated the model in an independent cohort from 2021 to 2023. We evaluated the model performance using discrimination, measured by the area under the receiver operating characteristic curve (AUC), calibration plots, and decision curve analysis (DCA). A reference value threshold was identified using a Youden Index–based cutoff.

Results

The development and validation cohorts included 1,136 and 1,286 students, respectively. The prevalence of suicidal ideation at intake was comparable across cohorts (14.0% and 14.7%). The model achieved moderate discrimination in the development cohort (AUC = 0.71) and good discrimination in the validation cohort (AUC = 0.74), with acceptable calibration in both cohorts. Higher levels of screening‑based psychological distress, particularly feelings of worthlessness, female sex, and lower self-esteem increased the odds of suicidal ideation at intake. Conversely, call-based or early outreach referral routes were associated with lower odds. DCA showed that the model provided greater net benefit than strategies that intervened in all or no students across clinically plausible threshold probabilities, including the Youden Index‑based cutoff (approximately 0.13).

Conclusions

Routinely collected mental health screening data can predict suicidal ideation at the time of initial counseling contact with reasonable accuracy and potential clinical utility. Screening‑based prediction models may strengthen existing student support systems by enabling earlier identification of elevated risk and guiding proportionate intervention strategies.

Background

Suicidal ideation: from passive desires for death to active thoughts of self-harm—is widely recognized as a critical precursor to suicide attempts and completed suicide [1–3]. Worldwide, suicide is a major public health crisis and claims more than 720,000 lives annually representing approximately 1 out of 100 deaths. For people aged 15–29 years, suicide ranks third as the leading cause of death [4]. In Japan, suicide has been the leading cause of death among youth aged 15–29 years since the late 1980s. While the overall suicide rate has declined after peaking in 2003, this demographic has demonstrated a concerning upward trend since 2016 [5]. Japan is the only G7 nation where suicide is the leading cause of death among young adults—underscoring a critical public health issue [6, 7].

Among Japanese university students, many report substantial psychological distress and reduced well-being. For example, 37.5% report life as difficult, with feelings of inadequacy and emptiness associated with suicidal ideation [8]. While Western models view emptiness as a symptom of depression or borderline personality disorder [9, 10], Japanese scholars emphasize an existential emptiness characterized by perceived insignificance and loss of ikigai—serving as predictors of suicidality independent of clinical diagnoses [11, 12]. The increase observed in 2016–2019 is mainly attributed to the negative impacts of social networking services, including cyberbullying and social comparison, alongside chronic school-related stressors and the absence of adequate support systems [5, 13, 14]. Recent large-scale stressors have underscored these vulnerabilities. During COVID-19, suicide rates increased among undergraduate students in Japanese national universities [15]. These examples demonstrate how external stressors exacerbate psychological distress and reveal structural weaknesses in university mental health systems.

These environmental and psychosocial pressures manifest as persistent suicidal ideation [5, 14]. As per global estimates, each suicide is accompanied by more than 20 attempts, with a high prevalence of suicidal ideation [16, 17]. International studies report prevalence rates ranging from 10 to 20% among university students [18–21]. Recent surveys in Japan indicate that approximately one out of four young adults has seriously considered suicide at certain points in life [22, 23]. This ideation functions as a critical precursor to suicidal behavior [1–3, 16], but many individuals remain isolated without seeking professional help [16, 22]. These findings highlight the need for timely identification of students at elevated risk. Early risk stratification remains a critical component of effective suicide prevention in university settings.

A substantial body of research has examined the risk and protective factors associated with suicidal ideation. Psychological distress consistently emerges as a strong predictor of suicidal ideation, whereas self-esteem functions as a robust protective factor across cultural contexts [24–29]. Several dimensions of psychological distress including worthlessness, hopelessness and depressive mood have been associated with suicidal ideation in previous studies [29–32]. However, their independent effects appear to vary after adjustment for overlapping psychological constructs and self-evaluative processes. For example, previous studies have suggested that hopelessness mediates the relationship between low self-esteem, heightened social anxiety, and suicidal ideation among university students [32]. In contrast, evidence indicates that suicidal ideation is more strongly linked to academic stress than to self-esteem in undergraduate populations [33]. Suicidal ideation has also been associated with lifestyle-related factors, including poor sleep duration (both short and prolonged sleep), physical inactivity, and unhealthy dietary patterns [34]. Longitudinal studies indicate that modifiable psychosocial and lifestyle factors observed at university admission often persist across subsequent academic years and increase the vulnerability to mental health problems [35]. These findings underscore the importance of such factors as contextual markers for early risk stratification rather than as isolated causal determinants.

Collectively, this evidence underscores the multifactorial nature of suicidal ideation, emerging from the interaction among psychological distress, reduced protective factors, lifestyle disruptions, and barriers to access to support services [36–40]. In addition, a consistent phenomenon observed across cultures is the gender paradox of suicide, characterized by higher rates of suicidal ideation and nonlethal attempts among women and significantly higher rates of completed suicide among men [41, 42]. Recent evidence reports that suicide rates among women have rapidly increased, and gender-specific risk factors have been increasingly recognized, underscoring the need for tailored, gender-sensitive prevention strategies [43, 44]. Despite these insights, Japanese studies examining the combined influence of these factors within an integrated predictive framework—particularly in university settings—remain scarce. This gap is especially consequential given constrained mental health resources in many institutions and the need to prioritize support for students at greatest risk.

To facilitate early identification of high-risk students, universities worldwide have implemented systematic mental health screening initiatives [45, 46]. In Western countries, programs such as the Interactive Screening Program in the United States, MHC-based screening initiatives in Canada, and StudiCare in Germany have shown efficacy in connecting distressed students to mental health services [47–49]. Similarly, Japanese universities have increasingly adopted digital health check-ups and screening instruments such as the Kessler Psychological Distress Scale (K10) to identify students with severe psychological distress [46, 50]. The K10 comprises 10 items for assessing the frequency of nonspecific psychological distress symptoms within the past month. Respondents rate the frequency of experiencing feelings such as fatigue without reason, nervousness, hopelessness, restlessness, depression, effortfulness, and worthlessness [51]. Items are rated using a five-point scale ranging from 0 (none of the time) to 4 (all of the time), with high scores indicating greater psychological distress. Compared with available tools, K10 demonstrated strong psychometric properties, brevity, and feasibility for large-scale screening, including clinical utility in identifying individuals with psychological distress [52, 53]. The K10 was selected over disorder-specific measures such as the Patient Health Questionnaire-9, because it captures nonspecific psychological distress relevant to early risk detection in nonclinical populations [54, 55]. Recent studies proposed that high scores on three subitems of the K10 (4: hopelessness; 7: depression; 10: worthlessness) predicted high levels of suicidal ideation [32, 56]. However, screening alone cannot address the practical challenge faced by university counseling services: how to prioritize limited mental health resources effectively. This challenge is particularly acute during initial counseling intake, when clinicians require brief, interpretable, and clinically actionable indicators to support rapid risk stratification.

In this context, predictive modeling offers a complementary framework for estimating the individual-level risk of suicidal ideation using information available before or at the time of first contact with counseling services. Several prediction models of suicide risk among university students have reported high discriminatory performance using machine learning and advanced statistical methods [57–60]. However, many of these models prioritize algorithmic complexity over clinical interpretability. They also rarely assess performance stability across different periods, which limits their practicality in real-world university counseling settings.

Accordingly, this study aims to develop and temporally validate a multivariable prediction model for suicidal ideation among Japanese university students accessing counseling services for the first time. The model uses routinely collected mental health screening data to predict suicidal ideation at initial counseling intake. Unlike many existing approaches, this study prioritizes interpretability, feasibility, and temporal robustness over causal inference or algorithmic sophistication alone. The study may support earlier risk stratification and more efficient allocation of mental health resources within higher education settings.

Methods

Study design and setting

This retrospective observational study analyzed data collected from a university counseling service in Japan. The primary aim of this study was to develop and temporally validate a multivariable prediction model for suicidal ideation at first contact with the counseling service. The study followed the TRIPOD (Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis) guidelines.

Study procedures and data sources

At the beginning of each academic year, the university administered a mental health screening survey to all enrolled undergraduate students as part of routine student support activities. The survey operated independently of counseling service use and collected data on psychological distress, self‑esteem, lifestyle factors, and demographic characteristics.

The present study focused exclusively on students who subsequently accessed the university counseling service for the first time. During routine intake procedures, students completed an intake questionnaire that assessed current suicidal ideation.

Predictor variables were derived from the annual campus‑wide screening survey. The outcome variable, suicidal ideation at initial counseling intake assessed exclusively during counseling intake and was not obtained from the screening survey.

Participants

Participants were limited to students who accessed the university counseling service for the first time during the study periods and completed the intake questionnaire, which included an item assessing suicidal ideation, at initial contact. Two temporally distinct cohorts were defined: a development cohort (2017–2019) and a temporal validation cohort (2021–2023). Owing to disruptions related to the COVID-19 pandemic, data from 2020 were excluded.

Demographic characteristics

The participants in the development cohort (2017–2019; n = 1,136) were predominantly aged 18–21 years (38.8%, 22.9%, 15.4%, and 11.5%, respectively), with women and men representing 53.6% and 46.4%, respectively. More than half of the participants (54.0%) were first-year students, while 17.8%, 13.6%, 12.8%, and 1.9% were second-, third-, fourth-, and fifth and sixth-year students, respectively. Similarly, the participants in the temporal validation cohort (2021–2023; n = 1,286) were predominantly aged 18–21 years (40.2%, 21.9%, 16.3%, and 12.7%, respectively), with women and men accounting for 56.2% and 43.8%, respectively. More than half (55.1%) were first-year students, followed by 17.7%, 13.8%, 11.8%, and 1.6% in the second-, third-, fourth-, and fifth and sixth-year students, respectively.

Outcome variable

Suicidal ideation at the time of initial counseling intake was defined as the dependent variable. Suicidal ideation was assessed using a single dichotomous item included in the intake questionnaire and coded as present or absent [61]. This measure reflected current suicidal ideation at intake rather than the individual’s lifetime history.

Predictor variables

Predictors were prespecified based on prior literature and clinical relevance and were defined and identically coded in the development and validation cohorts. All predictors were treated as risk markers rather than causal or intervention variables.

Psychological distress was assessed using three subitems of the Kessler Psychological Distress Scale (K10): worthlessness, hopelessness, and depressive mood. Each item was originally measured on a five‑point Likert scale (0–4). For the purpose of prediction modeling, responses were dichotomized a priori, with scores of 0–2 coded as low (0) and 3–4 coded as high (1).

Self‑esteem was assessed using the total score of the Rosenberg Self‑Esteem Scale (SE_total) and treated as a continuous variable [62].

Lifestyle‑related variables included sleep duration (short or long sleep duration [≤ 5 h or ≥ 8 h] versus typical sleep duration [6–7 h]), lack of regular exercise, skipping breakfast, and loss of appetite, all of which were coded as dichotomous variables. Sex was coded as female or male. The access route to the counseling service was coded as call‑based or early outreach contact versus voluntary self‑initiated access. Call‑based or early outreach contact referred to cases in which students were proactively contacted by the counseling service when routine screening identified elevated psychological distress, defined as a total score of 15 or higher on the Kessler Psychological Distress Scale (K10; range = 0–40) [63]. The analysis treats this variable as an indicator of route of access to services rather than as an intervention variable.

Additionally, multicollinearity among predictors was assessed using variance inflation factor (VIF) in the development cohort. All VIF values were below 2 and close to 1, indicating no evidence of problematic multicollinearity.

Model development

A multivariable logistic regression model was developed using data from the 2017–2019 development cohort. All prespecified predictors were entered simultaneously into the model without variable selection or penalization (forced‑entry method). Regression coefficients were estimated using maximum likelihood estimation.

Model discrimination in the development cohort was assessed using the area under the receiver operating characteristic curve (AUC). Apparent calibration in the development cohort was evaluated using decile‑based calibration plots comparing predicted probabilities with observed outcome frequencies.

Determination of optimal risk threshold

To identify a representative probability threshold for classifying elevated risk of suicidal ideation, the Youden Index (J = sensitivity + specificity − 1) was calculated using the development cohort. The optimal cutoff value was defined as the predicted probability that maximized the Youden Index.

Additionally, sensitivity and specificity across the range of predicted probabilities were graphically examined, and the intersection of sensitivity and specificity curves was used to visually confirm the Youden‑based cutoff.

The identified cutoff was treated as a reference value for evaluating model performance and clinical utility and was not intended to define a fixed clinical decision rule.

Temporal external validation

Temporal external validation was performed using an independent cohort of students who accessed the counseling service between 2021 and 2023. Regression coefficients and intercepts estimated in the development cohort were applied unchanged to the validation cohort to generate predicted probabilities of suicidal ideation.

Model discrimination in the validation cohort was evaluated using the AUC. Calibration in the validation cohort was assessed using decile‑based calibration plots comparing the predicted probabilities with observed outcome frequencies. No model refitting or recalibration was performed during validation.

Decision curve analysis

To evaluate the potential clinical utility of the prediction model in the validation cohort, decision curve analysis (DCA) was conducted. Net benefit was calculated across a range of threshold probabilities and compared with default strategies of intervening either in all students or in none.

The Youden‑based cutoff probability was superimposed on the decision curve as a reference point to examine whether the use of the prediction model conferred greater net benefit than alternative strategies at clinically relevant risk thresholds.

Statistical analysis

Descriptive statistics were used to summarize participant characteristics. Continuous variables were summarized using means and standard deviations, and categorical variables were summarized using frequencies and proportions. Logistic regression analyses and post‑estimation procedures were conducted using Stata (ver.19.0). Missing data were minimal (< 2% across variables) and were handled using multiple imputation. Statistical significance testing was considered secondary because the primary objectives were prediction, temporal validation, and evaluation of potential clinical utility rather than hypothesis testing.

Additionally, to characterize temporal patterns in the broader screening population and in service utilization, supplementary chi-square analyses were conducted. In particular, we examined differences across academic years in (1) the proportion of students classified as high risk based on screening data and (2) access routes to counseling services (call-based/early outreach vs. voluntary). The effect sizes were quantified using Cramer’s V, and adjusted residuals were examined to identify the specific cells that contributed to the overall associations. These analyses were conducted to provide contextual interpretation of temporal changes in help-seeking behavior, supplementing the primary prediction modeling analysis.

Ethical approval

This study analyzed anonymized data collected as part of routine university counseling service procedures. Kitasato University School of Medicine and Hospital Ethics Committee approved the study (Approval No. B23-166). All procedures adhered to the Declaration of Helsinki (2013 revision). Informed consent was obtained for the use of annual screening data, and intake questionnaire data were handled under an opt-out procedure in accordance with national ethical guidelines.

AI-assisted language editing

A large language model (Microsoft Copilot) was used solely for English-language editing and to improve the clarity and readability of the manuscript. All AI-assisted suggestions were reviewed, verified, and revised by the authors, and they retain full responsibility for the final content. No AI tools were used for data analysis, statistical procedures, or interpretation.

Results

Study population

The development cohort included students who first accessed the university counseling service between 2017 and 2019, while the temporal external validation cohort included those who accessed the service between 2021 and 2023. Missing data were handled using multiple imputation; therefore, all participants were retained in the analyses.

Although the baseline characteristics of the two cohorts were broadly comparable, the proportion of students accessing counseling via call‑based or early outreach routes was higher in the validation cohort. The prevalence of suicidal ideation at intake was similar between the cohorts (development cohort, 14.0%; validation cohort, 14.7%). Detailed cohort characteristics are presented in Table 1.Table 1Baseline characteristics of students accessing university counseling services for the first time

Variables Development cohort (2017–2019), n = 1,136 Validation cohort (2021–2023), n = 1,286

Suicidal ideation, [Outcome variable], % 14.0 14.7

Self-esteem score (SE_total), mean (SD) 22.46 (4.44) 22.80 (4.75)

Call-based access (vs. voluntary), % 64.7 75.7

Female, % 53.6 56.2

K10 worthlessness (high*), % 22.6 23.3

K10 hopelessness (high*), % 14.3 15.5

K10 depression (high*), % 31.0 34.3

Short or long sleep duration, % 37.9 32.1

Lack of regular exercise, % 42.9 37.1

Skipping breakfast, % 20.1 23.1

Loss of appetite, % 29.6 32.8

Baseline characteristics of the development cohort (2017–2019) and the temporal validation cohort (2021–2023) are shown. Predictor variables were obtained from a routinely conducted campus‑wide mental health screening administered prior to counseling service use, including psychological distress indicators, self‑esteem, lifestyle‑related factors, and demographic characteristics. Suicidal ideation was assessed at the time of initial counseling intake. Values are percentages unless otherwise indicated. Self‑esteem is presented as the mean (standard deviation). Kessler Psychological Distress Scale (K10) subitems (worthlessness, hopelessness, depressive mood) were dichotomized as low (scores 0–2) versus high (scores 3–4)

Additional information concerning the screening population and the flow of students to counseling service utilization is presented in Table 2. The proportion of students identified as high risk based on screening criteria remained relatively stable across academic years, ranging from approximately 10% to 13%. In contrast, patterns of service utilization varied across years. The proportion of students who attended counseling through call-based or early outreach was higher in the validation cohort (2021–2023), whereas voluntary visits were relatively more common in the development cohort (2017–2019). Supplementary chi-square analyses indicated a significant association between academic year and access route (call-based vs. voluntary) (χ2(5) = 55.18, p < 0.001), with a moderate effect size (Cramer’s V = 0.151). Adjusted residuals showed that voluntary visits were significantly more frequent in 2017–2019, whereas call-based access was significantly more frequent in 2021. Notably, although the association between academic year and screening-based high-risk classification was statistically significant, the effect size was negligible (χ2(5) = 32.997, p < 0.001, Cramer’s V = 0.035), indicating minimal variation in the underlying distribution of psychological risk across years.Table 2Annual screening population and flow from screening to counseling service utilization

Academic year and subtotal of each cohort Target populations Number of valid responses Number of outreach targets (highrisk;%) Number of high-risk interviews Number of unresponsive students Number of voluntary visitors

2017 5,238 4,677 577 (12.3) 250 327 142

2018 5,248 4,794 527(11.0) 252 275 135

2019 5,274 4,821 480(10.0) 233 247 124

Development Cohort 15,760 14,292 1,584(11.1) 735 849 401

2021 5,275 3,972 526(13.2) 357 169 74

2022 5,317 4,594 576(12.5) 345 231 115

2023 5,405 4,525 553(12.2) 272 281 123

Validation Cohort 15,997 13,091 1,655(12.6) 974 681 312

This table summarizes the flow of students from campus-wide mental health screening to counseling service utilization for each academic year. The total number of students targeted for screening, number of valid responses, number of students identified as high risk based on screening criteria, and subsequent engagement with counseling services (call-based outreach interviews, voluntary visits, and nonresponse after outreach) are presented. The table provides contextual information concerning the relationship between the screened population and the analytic sample used in the prediction model. In particular, it illustrates the proportion of high-risk students who did not attend counseling services despite outreach efforts, highlighting potential selection bias and temporal changes in help-seeking behavior across academic years

Model development and performance in the development cohort

A multivariable logistic regression model predicting suicidal ideation at initial counseling intake was developed using the 2017–2019 cohort, with all prespecified predictors entered simultaneously. In the development cohort, the model demonstrated moderate discriminatory performance, with an area under the receiver operating characteristic curve (AUC) of 0.71 (Fig. 1).

Higher levels of screening‑based psychological distress—particularly high levels of worthlessness—as well as female sex were associated with increased odds of suicidal ideation. In contrast, higher self‑esteem scores and access to the counseling service via call‑based or early outreach contact were associated with lower odds of suicidal ideation. Lifestyle‑related variables, including lack of regular exercise and skipping breakfast, were not independently associated with suicidal ideation but were retained in the model as covariates owing to their theoretical relevance and potential clinical utility.

Odds ratios, corresponding 95% confidence intervals, and p-values for all predictors included in the development model are shown in Table 3.Table 3Screening‑based multivariable logistic regression model predicting suicidal ideation at initial intake (Development cohort, 2017–2019)

Predictors Odds ratio (95% CI) p-value

Self-esteem (SE_total) (per 1-point increase) 0.94 (0.90–0.99) 0.014

K10 worthlessness (1) 2.04 (1.27–3.27) 0.003

K10 hopelessness (1) 1.39 (0.83–2.31) 0.207

K10 depression (1) 1.46 (0.93–2.32) 0.103

Female (1) 1.60 (1.10–2.31) 0.014

Call-based/early outreach access (1) 0.35 (0.23–0.52)  < 0.001

Short or long sleep duration (1) 1.17 (0.81–1.67) 0.404

Lack of regular exercise (1) 1.12 (0.72–1.65) 0.541

Skipping breakfast (1) 1.24 (0.80–1.93) 0.328

Loss of appetite (1) 1.32 (0.76–2.29) 0.329

Multivariable logistic regression results from the development cohort. The outcome variable was suicidal ideation assessed at the time of initial counseling intake (yes/no). All predictors were obtained from pre‑intake mental health screening. K10 subitems were dichotomized as low (scores 0–2) versus high (scores 3–4). Odds ratios (ORs) greater than 1 indicate higher odds of suicidal ideation, whereas ORs less than 1 indicate lower odds

Determination of optimal risk threshold

Using the development cohort, a representative probability threshold for identifying individuals at elevated risk of suicidal ideation was explored using Youden’s J statistic. The Youden Index was maximized at a predicted probability of approximately 0.13, corresponding to a sensitivity of approximately 0.70 and a specificity of approximately 0.64.

This cutoff was treated as a reference value to illustrate the balance between sensitivity and specificity and was not intended to define a fixed clinical decision rule.

Temporal external validation

Temporal external validation was conducted using the independent cohort of students who first accessed the counseling service between 2021 and 2023. Regression coefficients and intercepts estimated in the development cohort were applied unchanged to the validation cohort to generate predicted probabilities of suicidal ideation.

In the validation cohort, the model demonstrated good discriminatory performance, yielding an AUC of 0.74, indicating that predictive performance was maintained across temporally distinct cohorts (Fig. 2).

Model calibration was assessed using decile‑based calibration plots comparing predicted probabilities with observed proportions of suicidal ideation. Overall, the predicted risks agreed well with the observed outcomes in the validation cohort. Although some variability was observed in the higher predicted risk range, no systematic over‑ or underestimation of risk was evident (Fig. 3).

Decision curve analysis

DCA was performed in the validation cohort to evaluate the potential clinical utility of the prediction model. Across a range of clinically plausible threshold probabilities, the prediction model demonstrated greater net benefit than default strategies of either intervening in all students or in none (Fig. 4).

At a threshold probability corresponding to the Youden-based cutoff (approximately 0.13), use of the prediction model was associated with a positive net benefit, supporting its potential value for risk stratification at the time of first contact with university counseling services.

Summary of model performance

Overall, the prediction model demonstrated stable discrimination, acceptable calibration, and favorable decision‑analytic performance in the development and temporal validation cohorts. These findings support the robustness of the model and suggest its potential utility for identifying students at increased risk of suicidal ideation at initial counseling intake.

Discussion

Principal findings

In this study, we developed and temporally validated a multivariable prediction model for suicidal ideation assessed at the time of first contact with a university counseling service, using information obtained from an annual mental health screening conducted prior to counseling service use. The model demonstrated stable discrimination, acceptable calibration, and favorable decision‑analytic performance across temporally distinct cohorts. These findings indicate that routinely collected screening data can be used to identify students at an elevated risk of suicidal ideation at intake.

A major strength of the model is the clear temporal separation between predictors and outcome. Psychological distress indicators, self‑esteem, lifestyle‑related factors, and demographic characteristics were collected through routine campus‑wide screening independently of counseling service utilization, whereas suicidal ideation was assessed later at intake. This temporal ordering minimizes information leakage and supports interpretation of the model as a screening‑based prediction tool rather than a cross‑sectional diagnostic aid.

Among screening‑based predictors, higher levels of psychological distress—particularly feelings of worthlessness—were consistently associated with increased odds of suicidal ideation at intake. Female sex and lower self‑esteem were also associated with higher odds of suicidal ideation, whereas access to the counseling service via call‑based or early outreach routes was associated with lower odds.

Lifestyle‑related factors, such as lack of regular exercise and skipping breakfast, were not independently associated with suicidal ideation in the multivariable model, but were retained as theoretically relevant covariates to support overall risk adjustment and model stability.

The finding regarding call-based or early outreach access requires further consideration. This is supported by the supplementary analyses of the screening population and service utilization. As shown in Table 2, the overall proportion of students identified as high risk based on screening criteria remained relatively stable across different academic years, whereas patterns of help-seeking behavior varied substantially. In particular, voluntary visits were more common in the pre-pandemic period (2017–2019), whereas call-based outreach accounted for a larger proportion of counseling access in 2021. These findings indicate that the observed protective association of call-based access may not simply reflect lower-risk individuals; rather, it may represent the role of outreach as a mechanism to engage students who might otherwise not seek help. An alternative explanation is that outreach-based contact may facilitate earlier engagement before symptoms escalate, considering that students who seek help voluntarily may do so after experiencing more severe distress, which could result in higher levels of suicidal ideation at intake. These interpretations are consistent with the shift in help-seeking behavior observed during the COVID-19 pandemic, when barriers to voluntary service use may have increased, thereby elevating the relative importance of proactive outreach strategies. This perspective on engagement could also help explain why certain lifestyle-related variables were retained in the model although they lacked statistical significance, given their potential role as accessible entry points for engagement in real-world counseling settings. In practice, such factors may serve as less intrusive and more approachable entry points during outreach-based or early contact interventions, compared with direct interventions addressing sensitive psychological symptoms. Therefore, even in the absence of statistical significance, such variables may facilitate engagement with students and contribute to the practical usefulness of the model without compromising its overall performance.

Taken together, the observations at the individual and system levels suggest that the COVID-19 pandemic had a greater impact on help-seeking behavior and access pathways than on the overall distribution of screening-based psychological risk. The proportion of high-risk individuals remained relatively stable; however, the pathway through which students accessed counseling services shifted, with call-based outreach playing a more prominent role when voluntary visits declined. These findings highlight the importance of considering both individual- and system-level factors when interpreting and applying such models.

Comparison with previous research

Previous studies examining suicidal ideation among university students have often relied on cross‑sectional data in which psychological symptoms and suicidal thoughts were assessed simultaneously, limiting interpretation of temporal relationships [27–30, 33]. In contrast, the present study leveraged routinely collected screening data obtained before the use of counseling services to predict suicidal ideation assessed later at intake. This design strengthens the evidence that elevated psychological distress and low self‑esteem identified during screening are meaningful indicators of subsequent risk at the point of help‑seeking.

The importance of worthlessness and related cognitive‑affective symptoms observed in this study is consistent with prior research linking negative self‑evaluation and hopeless cognitions to suicidal ideation [30, 32]. However, by embedding these variables within a prediction modeling framework and validating the model temporally, the present study extends existing literature from explanatory associations toward practical risk stratification.

Clinical and practical implications

From a practical perspective, the findings suggest that annual mental health screening data may be used to support the early identification of students who present to counseling services with elevated risk of suicidal ideation. The Youden Index was used to identify a reference probability threshold that balanced sensitivity and specificity in the development cohort. This threshold was not intended to function as a rigid clinical cutoff but rather to facilitate interpretation of model performance and downstream evaluation of clinical utility.

DCA further demonstrated that, across a range of clinically plausible threshold probabilities—including the Youden‑based reference point—the prediction model yielded greater net benefit than default strategies of intervening in all students or intervening in none. This indicates that using model‑estimated risk could, in principle, improve the efficiency of targeted risk assessment or follow‑up without increasing unnecessary interventions. The model is intended to support, rather than replace, clinical judgment within existing student support systems.

As all predictors and their corresponding coefficients are explicitly specified, the model can be reproduced in other institutions using comparable screening data. However, its applicability may depend on the similarity of screening procedures, variable definitions, and student populations. The model may be applied to identify students at elevated risk of suicidal ideation based on screening data and to support prioritization for outreach or early intervention.

Strengths

This study has several notable strengths. First, the clear temporal ordering of predictors and outcome enhances interpretability and reduces the likelihood of information leakage, which is a common limitation in mental health prediction studies. Second, the use of routinely collected screening data improves feasibility and potential scalability in real‑world university settings. Third, we conducted temporal external validation and DCA, providing evidence of not only predictive performance but also potential clinical utility. In addition, predictors were selected a priori and entered using a forced‑entry approach, reducing the risk of overfitting associated with data‑driven variable selection. Together, these features support the robustness and practical relevance of the proposed model.

Limitations

Several limitations should be acknowledged. First, this study was conducted within a single university counseling service, which may limit its generalizability to other institutions or cultural contexts. Second, suicidal ideation was assessed using a single intake item and may not capture the full complexity or persistence of suicidal thoughts over time. Third, to enhance robustness and interpretability, selected psychological distress measures were dichotomized; although exploratory analyses supported this approach, some information loss is inevitable. Finally, although the model demonstrated stable performance across temporal cohorts, additional external validation in independent populations and prospective evaluation are required before routine implementation.

In addition, Table 2 shows that a substantial proportion of students identified as high risk in the screening did not attend counseling services despite outreach efforts. This finding may be attributable to potential selection bias, as the present analyses were restricted to students who completed the screening survey and subsequently accessed counseling services. Therefore, the study findings should be interpreted within this specific context and may not fully represent the entire screened student population.

Conclusions

In conclusion, this study demonstrates that routinely collected mental health screening data can be used to predict suicidal ideation with reasonable accuracy and potential clinical utility at the time of initial counseling contact. Screening‑based prediction models of this kind may complement existing student support systems by facilitating earlier recognition of elevated risk and informing timely, proportionate intervention strategies.

References

1.Rizvi A, Harmer B, Saadabadi A. Suicidal ideation. In: StatPearls. Treasure Island (FL): StatPearls Publishing. 2025 Jan. https://www.ncbi.nlm.nih.gov/books/NBK565877/. Accessed 28 Dec 2025.33351435

2.Wastler HM Khazem LR Ammendola E Baker JC Bauder CR Tabares J An empirical investigation of the distinction between passive and active ideation: understanding the latent structure of suicidal thought content Suicide Life Threat Behav 2023 53 219 226 36504400 10.1111/sltb.12935 Wastler HM, Khazem LR, Ammendola E, Baker JC, Bauder CR, Tabares J, et al. An empirical investigation of the distinction between passive and active ideation: understanding the latent structure of suicidal thought content. Suicide Life Threat Behav. 2023;53:219–26.10.1111/sltb.12935 36504400

3.Van Orden KA Witte TK Cukrowicz KC Braithwaite SR Selby EA Joiner TE Jr.The interpersonal theory of suicide Psychol Rev 2010 117 575 600 20438238 PMC3130348 10.1037/a0018697 Van Orden KA, Witte TK, Cukrowicz KC, Braithwaite SR, Selby EA, Joiner TE Jr. The interpersonal theory of suicide. Psychol Rev. 2010;117:575–600.10.1037/a0018697 PMC3130348 20438238

4.World Health OrganizationSuicide 2024 GenevaWHOWorld Health Organization. Suicide. Geneva: WHO; 2024.

5.Ministry of Health, Labour and WelfareWhite paper on suicide prevention in Japan 2024 2024 TokyoMHLWMinistry of Health, Labour and Welfare. White paper on suicide prevention in Japan 2024. Tokyo: MHLW; 2024.

6.Ministry of Health, Labour and WelfareSummary of vital statistics 2017 (approximate numbers) 2018 TokyoMHLWMinistry of Health, Labour and Welfare. Summary of vital statistics 2017 (approximate numbers). Tokyo: MHLW; 2018.

7.Ministry of Health, Labour and WelfareWhite paper on suicide prevention in Japan 2025 2025 TokyoMHLWMinistry of Health, Labour and Welfare. White paper on suicide prevention in Japan 2025. Tokyo: MHLW; 2025.

8.Otsuka H Anamizu S Japanese university students’ difficulty in living and its association with suicidal ideation Asian J Psychiatr 2019 43 50 52 Accessed 28 Dec 202531079008 10.1016/j.ajp.2019.05.004 Otsuka H, Anamizu S. Japanese university students’ difficulty in living and its association with suicidal ideation. Asian J Psychiatr. 2019;43:50–2 . Accessed 28 Dec 2025.10.1016/j.ajp.2019.05.004 31079008

9.Miller CE Townsend ML Grenyer BFS Understanding chronic feelings of emptiness in borderline personality disorder: a qualitative study Borderline Personal Disord Emot Dysregul 2021 8 24 34365966 PMC8351135 10.1186/s40479-021-00164-8 Miller CE, Townsend ML, Grenyer BFS. Understanding chronic feelings of emptiness in borderline personality disorder: a qualitative study. Borderline Personal Disord Emot Dysregul. 2021;8:24.10.1186/s40479-021-00164-8 PMC8351135 34365966

10.Miller CE Townsend ML Day NJS Grenyer BFS Measuring the shadows: a systematic review of chronic emptiness in borderline personality disorder PLoS ONE 2020 15 e0233970 32609777 PMC7329066 10.1371/journal.pone.0233970 Miller CE, Townsend ML, Day NJS, Grenyer BFS. Measuring the shadows: a systematic review of chronic emptiness in borderline personality disorder. PLoS ONE. 2020;15:e0233970.10.1371/journal.pone.0233970 PMC7329066 32609777

11.Iwasawa A Tanji F Miyamoto S Nomura K The association between loneliness, suicidal ideation, and psychological distress considering family compositions: a cross-sectional study in a Japanese rural area Sci Rep 2025 15 11676 10.1038/s41598-025-96205-1 40188227 PMC11972342 Iwasawa A, Tanji F, Miyamoto S, Nomura K. The association between loneliness, suicidal ideation, and psychological distress considering family compositions: a cross-sectional study in a Japanese rural area. Sci Rep. 2025;15:11676.10.1038/s41598-025-96205-1 PMC11972342 40188227

12.Ozawa-de SC Too lonely to die alone: internet suicide pacts and existential suffering in Japan Cult Med Psychiatry 2008 32 516 551 10.1007/s11013-008-9108-0 18800195 Ozawa-de SC. Too lonely to die alone: internet suicide pacts and existential suffering in Japan. Cult Med Psychiatry. 2008;32:516–51.10.1007/s11013-008-9108-0 18800195

13.Cabinet Office, Government of JapanWhite paper on children and young people 2020 2020 TokyoCabinet OfficeCabinet Office, Government of Japan. White paper on children and young people 2020. Tokyo: Cabinet Office; 2020.

14.Imataka G Shiraishi H Youth suicide in Japan: exploring the role of subcultures, internet addiction, and societal pressures Diseases 2025 13 2 10.3390/diseases13010002 PMC11764159 39851466 Imataka G, Shiraishi H. Youth suicide in Japan: exploring the role of subcultures, internet addiction, and societal pressures. Diseases. 2025;13:2.10.3390/diseases13010002 PMC11764159 39851466

15.Fuse-Nagase Y Marutani T Tachikawa H Iwami T Yamamoto Y Moriyama T Increase in suicide rates among undergraduate students in Japanese national universities during the COVID-19 pandemic Psychiatry Clin Neurosci 2021 75 351 352 34355847 PMC8446983 10.1111/pcn.13293 Fuse-Nagase Y, Marutani T, Tachikawa H, Iwami T, Yamamoto Y, Moriyama T, et al. Increase in suicide rates among undergraduate students in Japanese national universities during the COVID-19 pandemic. Psychiatry Clin Neurosci. 2021;75:351–2.10.1111/pcn.13293 PMC8446983 34355847

16.World Health OrganizationPreventing suicide: a global imperative 2014 GenevaWHOWorld Health Organization. Preventing suicide: a global imperative. Geneva: WHO; 2014.

17.World Health OrganizationSuicide worldwide in 2019: global health estimates 2021 GenevaWHOWorld Health Organization. Suicide worldwide in 2019: global health estimates. Geneva: WHO; 2021.

18.Mortier P Auerbach RP Alonso J Axinn WG Cuijpers P Ebert DD Suicidal thoughts and behaviors among college students and same-aged peers: results from the WHO World Mental Health Surveys Soc Psychiatry Psychiatr Epidemiol 2018 53 279 288 10.1007/s00127-018-1481-6 29340781 PMC5896296 Mortier P, Auerbach RP, Alonso J, Axinn WG, Cuijpers P, Ebert DD, et al. Suicidal thoughts and behaviors among college students and same-aged peers: results from the WHO World Mental Health Surveys. Soc Psychiatry Psychiatr Epidemiol. 2018;53:279–88.10.1007/s00127-018-1481-6 PMC5896296 29340781

19.Teixeira KOM Lisboa JL Dias BMF Ferreira RCF Zarzar PMPA Sampaio AA Suicidal ideation among university students during the COVID 19 pandemic: a rapid systematic review with meta-analysis Res Soc Dev 2022 11 1 19 10.33448/rsd-v11i4.26426 Teixeira KOM, Lisboa JL, Dias BMF, Ferreira RCF, Zarzar PMPA, Sampaio AA. Suicidal ideation among university students during the COVID 19 pandemic: a rapid systematic review with meta-analysis. Res Soc Dev. 2022;11:1–19. 10.33448/rsd-v11i4.26426 .

20.Li ZZ Li YM Lei XY Zhang D Liu L Tang SY Prevalence of suicidal ideation in Chinese college students: a meta-analysis PLoS ONE 2014 9 e104368 25285890 PMC4186746 10.1371/journal.pone.0104368 Li ZZ, Li YM, Lei XY, Zhang D, Liu L, Tang SY, et al. Prevalence of suicidal ideation in Chinese college students: a meta-analysis. PLoS ONE. 2014;9:e104368.10.1371/journal.pone.0104368 PMC4186746 25285890

21.Wang M Kou C Bai W Song Y Liu X Yu W Prevalence and correlates of suicidal ideation among college students: a mental health survey in Jilin Province, China J Affect Disord 2019 246 166 173 30583141 10.1016/j.jad.2018.12.055 Wang M, Kou C, Bai W, Song Y, Liu X, Yu W, et al. Prevalence and correlates of suicidal ideation among college students: a mental health survey in Jilin Province, China. J Affect Disord. 2019;246:166–73.10.1016/j.jad.2018.12.055 30583141

22.The Nippon FoundationThe 4th suicide awareness survey: report on the actual situation of suicide ideation and attempts 2021 TokyoThe Nippon FoundationThe Nippon Foundation. The 4th suicide awareness survey: report on the actual situation of suicide ideation and attempts. Tokyo: The Nippon Foundation; 2021.

23.Ministry of Health, Labour and Welfare. White paper on suicide prevention in Japan 2022. Tokyo: MHLW. 2022. www.mhlw.go.jp. Accessed 28 Dec 2025.

24.Owusu-Ansah FE Addae AA Peasah BO Oppong Asante K Osafo J Suicide among university students: prevalence, risks and protective factors Health Psychol Behav Med 2020 8 220 233 34040869 PMC8114407 10.1080/21642850.2020.1766978 Owusu-Ansah FE, Addae AA, Peasah BO, Oppong Asante K, Osafo J. Suicide among university students: prevalence, risks and protective factors. Health Psychol Behav Med. 2020;8:220–33.10.1080/21642850.2020.1766978 PMC8114407 34040869

25.Teismann T Joiner TE Robison M Brailovskaia J Self-burdensomeness, self-esteem and suicidal ideation Cognit Ther Res 2024 48 398 405 40832131 PMC12360707 10.1007/s10608-024-10477-x Teismann T, Joiner TE, Robison M, Brailovskaia J. Self-burdensomeness, self-esteem and suicidal ideation. Cognit Ther Res. 2024;48:398–405.10.1007/s10608-024-10477-x PMC12360707 40832131

26.Hiçdurmaz D Inci F Karahan S Predictors of mental health symptoms, automatic thoughts, and self-esteem among university students Psychol Rep 2017 120 650 669 28558541 10.1177/0033294117707945 Hiçdurmaz D, Inci F, Karahan S. Predictors of mental health symptoms, automatic thoughts, and self-esteem among university students. Psychol Rep. 2017;120:650–69.10.1177/0033294117707945 28558541

27.Luceño-Moreno L, Vázquez-Estévez D, Martín-García J, Talavera-Velasco B. Factors associated with suicidal ideation in college students of health sciences. Depress Anxiety. 2025;4397417:7. 10.1155/da/4397417 .PMC12321416 40761832

28.Mitsui N Asakura S Shimizu Y Fujii Y Toyomaki A Kako Y The association between suicide risk and self-esteem in Japanese university students with major depressive episodes of major depressive disorder Neuropsychiatr Dis Treat 2014 10 811 816 24868158 PMC4031242 10.2147/NDT.S59349 Mitsui N, Asakura S, Shimizu Y, Fujii Y, Toyomaki A, Kako Y, et al. The association between suicide risk and self-esteem in Japanese university students with major depressive episodes of major depressive disorder. Neuropsychiatr Dis Treat. 2014;10:811–6.10.2147/NDT.S59349 PMC4031242 24868158

29.Yamaguchi S Nagamine D Murofushi Y Matsuda K Negative interaction between emotional vulnerability and depressive symptoms may heighten suicidal ideation among Japanese university students: a cross-sectional study Front Psychiatry 2024 15 1383907 39421061 PMC11484067 10.3389/fpsyt.2024.1383907 Yamaguchi S, Nagamine D, Murofushi Y, Matsuda K. Negative interaction between emotional vulnerability and depressive symptoms may heighten suicidal ideation among Japanese university students: a cross-sectional study. Front Psychiatry. 2024;15:1383907.10.3389/fpsyt.2024.1383907 PMC11484067 39421061

30.Ali MH Hossain SB Mahmud MA Rahman KM Mridha MS Hassan MS Exploration of the factors associated with suicide ideation among students: a structural equation modeling (SEM) approach J Affect Disord Rep 2025 21 100936 10.1016/j.jadr.2025.100936 Ali MH, Hossain SB, Mahmud MA, Rahman KM, Mridha MS, Hassan MS, et al. Exploration of the factors associated with suicide ideation among students: a structural equation modeling (SEM) approach. J Affect Disord Rep. 2025;21:100936. 10.1016/j.jadr.2025.100936 .

31.Zhang J Liu X Fang L Combined effects of depression and anxiety on suicide: a case-control psychological autopsy study in rural China Psychiatry Res 2019 271 370 373 10.1016/j.psychres.2018.11.010 30529321 PMC6382523 Zhang J, Liu X, Fang L. Combined effects of depression and anxiety on suicide: a case-control psychological autopsy study in rural China. Psychiatry Res. 2019;271:370–3. 10.1016/j.psychres.2018.11.010 .PMC6382523 30529321

32.Dat NT Mitsui N Asakura S Watanabe S Takanobu K Fujii Y The mediating role of hopelessness in the relationship between self-esteem, social anxiety, and suicidal ideation among Japanese university students who visited a university health care center J Affect Disord Rep 2021 6 100192 Dat NT, Mitsui N, Asakura S, Watanabe S, Takanobu K, Fujii Y, et al. The mediating role of hopelessness in the relationship between self-esteem, social anxiety, and suicidal ideation among Japanese university students who visited a university health care center. J Affect Disord Rep. 2021;6:100192.

33.Nkwuda FCN, Ifeagwazi CM, Nwonyi SK, Oginyi RC. Suicidal ideation among undergraduate students: academic stress and self-esteem as predictive factors. Nigerian J Psychol Res. 2020;16:55–64. https://www.academia.edu/65942476/Suicidal_ideation_among_undergraduate_students_Academic_stress_and_self_esteem_as_predictive_factors. Accessed 28 Dec 2025.

34.Berardelli I Corigliano V Hawkins M Comparelli A Erbuto D Pompili M Lifestyle interventions and prevention of suicide Front Psychiatry 2018 9 567 10.3389/fpsyt.2018.00567 30459660 PMC6232529 Berardelli I, Corigliano V, Hawkins M, Comparelli A, Erbuto D, Pompili M. Lifestyle interventions and prevention of suicide. Front Psychiatry. 2018;9:567. 10.3389/fpsyt.2018.00567 .PMC6232529 30459660

35.Duffy A Keown-Stoneman C Goodday S Horrocks J Lowe M King N Predictors of mental health and academic outcomes in first-year university students: identifying prevention and early-intervention targets BJPsych Open 2020 6 e46 10.1192/bjo.2020.24 32381150 PMC7331085 Duffy A, Keown-Stoneman C, Goodday S, Horrocks J, Lowe M, King N, et al. Predictors of mental health and academic outcomes in first-year university students: identifying prevention and early-intervention targets. BJPsych Open. 2020;6:e46. 10.1192/bjo.2020.24 .PMC7331085 32381150

36.Arria AM O'Grady KE Caldeira KM Vincent KB Wilcox HC Wish ED Suicide ideation among college students: a multivariate analysis Arch Suicide Res 2009 13 230 246 19590997 PMC2709750 10.1080/13811110903044351 Arria AM, O’Grady KE, Caldeira KM, Vincent KB, Wilcox HC, Wish ED. Suicide ideation among college students: a multivariate analysis. Arch Suicide Res. 2009;13:230–46.10.1080/13811110903044351 PMC2709750 19590997

37.Kim MK. Relationship between negative emotions, family resilience, self-esteem and suicide ideation in university students. Korean J Fam Welfare. 2012;17: 61–83. http://journal.familywelfare.net/journal/article.php?code=25484. Accessed 28 Dec 2025.

38.Landa-Blanco M Romero K Caballero I Gálvez-Pineda E Fúnes-Henríquez MJ Romero R Exploring suicide ideation in university students: sleep quality, social media, self-esteem, and barriers to seeking psychological help Front Psychiatry 2024 15 1352889 38645419 PMC11027559 10.3389/fpsyt.2024.1352889 Landa-Blanco M, Romero K, Caballero I, Gálvez-Pineda E, Fúnes-Henríquez MJ, Romero R. Exploring suicide ideation in university students: sleep quality, social media, self-esteem, and barriers to seeking psychological help. Front Psychiatry. 2024;15:1352889.10.3389/fpsyt.2024.1352889 PMC11027559 38645419

39.Yamaguchi S Matsuda K Nagamine D Association between psychological vulnerability, depressive symptoms, and suicidal ideation among university students Jpn J Appl Psychol 2025 51 12 20 Yamaguchi S, Matsuda K, Nagamine D. Association between psychological vulnerability, depressive symptoms, and suicidal ideation among university students. Jpn J Appl Psychol. 2025;51:12–20.

40.Zou H Kong X Chen Z Zhang W Stressful life events among college students: direct and indirect relationships with insomnia, depression, self-esteem, and suicide ideation Psychol Sch 2024 61 3626 3644 10.1002/pits.23247 Zou H, Kong X, Chen Z, Zhang W. Stressful life events among college students: direct and indirect relationships with insomnia, depression, self-esteem, and suicide ideation. Psychol Sch. 2024;61:3626–44.

41.Rudd MD The prevalence of suicidal ideation among college students Suicide Life Threat Behav 1989 19 173 183 2749860 10.1111/j.1943-278X.1989.tb01031.x Rudd MD. The prevalence of suicidal ideation among college students. Suicide Life Threat Behav. 1989;19:173–83.10.1111/j.1943-278x.1989.tb01031.x 2749860

42.Lamis DA Lester D Gender differences in risk and protective factors for suicidal ideation among college students J Coll Stud Psychother 2013 27 62 77 10.1080/87568225.2013.739035 Lamis DA, Lester D. Gender differences in risk and protective factors for suicidal ideation among college students. J Coll Stud Psychother. 2013;27:62–77.

43.Mai AS Wan YM Tan BJW Tan EK Solving the gender paradox in adolescent suicide: challenges and directions Front Psychiatry. 2024 15 1386153 38962059 PMC11219559 10.3389/fpsyt.2024.1386153 Mai AS, Wan YM, Tan BJW, Tan EK. Solving the gender paradox in adolescent suicide: challenges and directions. Front Psychiatry. 2024;15:1386153.10.3389/fpsyt.2024.1386153 PMC11219559 38962059

44.Carretta RF McKee SA Rhee TG Gender differences in risks of suicide and suicidal behaviors in the USA: a narrative review Curr Psychiatry Rep 2023 25 809 824 10.1007/s11920-023-01473-1 37930559 PMC11225381 Carretta RF, McKee SA, Rhee TG. Gender differences in risks of suicide and suicidal behaviors in the USA: a narrative review. Curr Psychiatry Rep. 2023;25:809–24.10.1007/s11920-023-01473-1 PMC11225381 37930559

45.English I Campbell DG Prevalence and characteristics of universal depression screening in U.S. college health centers Fam Syst Health 2019 37 131 149 31058525 10.1037/fsh0000411 English I, Campbell DG. Prevalence and characteristics of universal depression screening in U.S. college health centers. Fam Syst Health. 2019;37:131–49.10.1037/fsh0000411 31058525

46.Yamada Y Ishitsuka M Ohmachi T Kawahara K Tanaka A Tachibana H Psychological distress and self-esteem as predictors of suicidal ideation among university students Campus Health 2025 62 176 177 Yamada Y, Ishitsuka M, Ohmachi T, Kawahara K, Tanaka A, Tachibana H, et al. Psychological distress and self-esteem as predictors of suicidal ideation among university students. Campus Health. 2025;62:176–7.

47.Ebert DD Mortier P Kaehlke F Bruffaerts R Baumeister H Auerbach RP Barriers of mental health treatment utilization among first-year college students: first cross-national results from the WHO World Mental Health International College Student Initiative Int J Methods Psychiatr Res 2019 28 e1782 31069905 PMC6522323 10.1002/mpr.1782 Ebert DD, Mortier P, Kaehlke F, Bruffaerts R, Baumeister H, Auerbach RP, et al. Barriers of mental health treatment utilization among first-year college students: first cross-national results from the WHO World Mental Health International College Student Initiative. Int J Methods Psychiatr Res. 2019;28:e1782.10.1002/mpr.1782 PMC6522323 31069905

48.Chen S-P Chang W-P Stuart H Self-reflection and screening mental health on Canadian university students: a randomized controlled trial BMC Psychol 2020 8 99 10.1186/s40359-020-00446-w PMC7391623 32727614 Chen S-P, Chang W-P, Stuart H. Self-reflection and screening mental health on Canadian university students: a randomized controlled trial. BMC Psychol. 2020;8:99.10.1186/s40359-020-00446-w PMC7391623 32727614

49.Kählke F Hasking P Küchler A-M Baumeister H Mental health services for German university students: acceptance of intervention targets and preference for delivery modes Front Digit Health 2024 6 1284661 38426046 PMC10903098 10.3389/fdgth.2024.1284661 Kählke F, Hasking P, Küchler A-M, Baumeister H. Mental health services for German university students: acceptance of intervention targets and preference for delivery modes. Front Digit Health. 2024;6:1284661.10.3389/fdgth.2024.1284661 PMC10903098 38426046

50.Fujimoto M Availability and validity of using Kessler 10 (K10) for University new students’ mental health survey: relation among outpatient treatment records, prescription and medication Bull Tokyo Mirai Univ 2014 7 147 155 Fujimoto M. Availability and validity of using Kessler 10 (K10) for University new students’ mental health survey: relation among outpatient treatment records, prescription and medication. Bull Tokyo Mirai Univ. 2014;7:147–55.

51.Kessler RC Andrews G Colpe LJ Hiripi E Mroczek DK Normand SLT Short screening scales to monitor population prevalences and trends in non-specific psychological distress Psychol Med 2002 32 959 976 12214795 10.1017/S0033291702006074 Kessler RC, Andrews G, Colpe LJ, Hiripi E, Mroczek DK, Normand SLT, et al. Short screening scales to monitor population prevalences and trends in non-specific psychological distress. Psychol Med. 2002;32:959–76.10.1017/s0033291702006074 12214795

52.Chamberlain P Goldney R Delfabbro P Gill T Dal Grande L Suicidal ideation: the clinical utility of the K10 Crisis 2009 30 39 42 19261567 10.1027/0227-5910.30.1.39 Chamberlain P, Goldney R, Delfabbro P, Gill T, Dal Grande L. Suicidal ideation: the clinical utility of the K10. Crisis. 2009;30:39–42.10.1027/0227-5910.30.1.39 19261567

53.Sakai W Noguchi H Comparison of tests of mental health for student counseling: formation of a common measure Jpn J Educ Psychol 2015 63 111 120 10.5926/jjep.63.111 Sakai W, Noguchi H. Comparison of tests of mental health for student counseling: formation of a common measure. Jpn J Educ Psychol. 2015;63:111–20.

54.Kroenke K Spitzer RL Williams JB The PHQ-9: validity of a brief depression severity measure J Gen Intern Med 2001 16 606 613 10.1046/j.1525-1497.2001.016009606.x 11556941 PMC1495268 Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. 2001;16:606–13.10.1046/j.1525-1497.2001.016009606.x PMC1495268 11556941

55.Searle AK Van Hooff M McFarlane AC Davies CE Tran T Hodson SE Screening for depression and psychological distress in a currently serving military population: the diagnostic accuracy of the K10 and the PHQ9 Assessment 2019 26 1411 1426 29192508 10.1177/1073191117745124 Searle AK, Van Hooff M, McFarlane AC, Davies CE, Tran T, Hodson SE, et al. Screening for depression and psychological distress in a currently serving military population: the diagnostic accuracy of the K10 and the PHQ9. Assessment. 2019;26:1411–26.10.1177/1073191117745124 29192508

56.Mitsui N, Sawagashira R, Ishihara K, Asakura S. Daigakusei o taisho toshita K10 ni yoru jisatsu nenryo no yosoku no kokoromi [Attempt to predict suicidal ideation among university students using K10] [Abstract of conference presentation]. In: The 63rd Annual Meeting of Japan University Health Association, Koriyama, Fukushima, Japan. 2025. p. 91.

57.Hasking PA Robinson K McEvoy P Melvin G Bruffaerts R Boyes ME Development and evaluation of a predictive algorithm and telehealth intervention to reduce suicidal behavior among university students Psychol Med 2024 54 971 979 37732419 10.1017/S0033291723002714 PMC10939946 Hasking PA, Robinson K, McEvoy P, Melvin G, Bruffaerts R, Boyes ME, et al. Development and evaluation of a predictive algorithm and telehealth intervention to reduce suicidal behavior among university students. Psychol Med. 2024;54:971–9.10.1017/S0033291723002714 PMC10939946 37732419

58.Cañón Buitrago SC Pérez Agudelo JM Narváez Marín M Montoya Hurtado OL Bermúdez Jaimes GI Predictive model of suicide risk in Colombian university students: quantitative analysis of associated factors Front Psychiatry 2024 15 1291299 38855643 PMC11157033 10.3389/fpsyt.2024.1291299 Cañón Buitrago SC, Pérez Agudelo JM, Narváez Marín M, Montoya Hurtado OL, Bermúdez Jaimes GI. Predictive model of suicide risk in Colombian university students: quantitative analysis of associated factors. Front Psychiatry. 2024;15:1291299.10.3389/fpsyt.2024.1291299 PMC11157033 38855643

59.Ballı M Doğan AE Şenol SH Yapıcı EH Machine learning based identification of suicidal ideation using non-suicidal predictors in a university mental health clinic Sci Rep 2025 15 13843 10.1038/s41598-025-97387-4 40263508 PMC12015358 Ballı M, Doğan AE, Şenol SH, Yapıcı EH. Machine learning based identification of suicidal ideation using non-suicidal predictors in a university mental health clinic. Sci Rep. 2025;15:13843.10.1038/s41598-025-97387-4 PMC12015358 40263508

60.Brock S Update on suicide statistics from the youth risk behavior survey system Communiqué (NASP) 2025 54 20 23 Brock S. Update on suicide statistics from the youth risk behavior survey system. Communiqué (NASP). 2025;54:20–3.

61.Ohno Y (Principal Investigator). Utsubyou no sukuriningu to sono tenki toshite no jisatsu no yobou sisutemu kouchiku ni kansuru kenkyuu (soukatsu kenkyuu houkokusho) [Screening for depressive states and development of a suicide prevention system (Comprehensive research report)]. Tokyo: Ministry of Health, Labour and Welfare; 2002. https://mhlw-grants.niph.go.jp/project/4186. Accessed 26 Dec 2025.

62.Rosenberg M Society and the adolescent self-image 1965 Princeton (NJ)Princeton University Press10.1515/9781400876136 Rosenberg M. Society and the adolescent self-image. Princeton (NJ): Princeton University Press; 1965.

63.Furukawa T, Ohno Y, Uda H, Nakane Y. Ippan jinkou chuu no seishinshikkan no kanbenna skuriningu ni kansuru kenkyuu [Study on simple screening for mental disorders in the general population]. In: Kawakami N, editor. Kokoro no kenkou mondai to taisaku kiban no jittai ni kansuru kenkyuu Kenkyuu houkokusho [Report of research co-operation: study on the actual condition of mental health problems and countermeasure infrastructure]. Tokyo: Ministry of Health, Labour and Welfare; 2003; p. 127–30. https://mhlw-grants.niph.go.jp/system/files/2002/000202/200200097A/200200097A0007.pdf. Accessed 26 Dec 2025.