Temporal Analysis of a Japanese Online Forum for Suicide Risk Monitoring

Temporal Analysis of a Japanese Online Forum for Suicide Risk Monitoring | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Temporal Analysis of a Japanese Online Forum for Suicide Risk Monitoring Takahiro Arai, Hiroyuki Shinkai, Keita Yamauchi This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7442638/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 20 Nov, 2025 Read the published version in BMC Psychiatry → Version 1 posted 11 You are reading this latest preprint version Abstract Background : Suicide prevention can be significantly enhanced by time-sensitive surveillance using digital data sources like online forums. To inform more effective suicide prevention strategies, this study analyzes the temporal patterns of posts on a Japanese mental health forum—the NHK “Facing Suicide” forum—that may aid in early risk detection. Methods : We analyzed 63,046 posts from Japan’s national broadcaster (NHK) forum (Jan 1, 2008–Mar 31, 2025), stratified by sex and age (≤ 19, 20s, 30s, ≥ 40). Generalized additive models were used to model hourly, weekly, and monthly variations, with time included as a spline term. Results are presented as incidence rate ratios (IRRs) with 95% confidence intervals (CIs). Results : Females contributed 75.5% of posts, and the 20–29 age group was the most active (32.4%). Posting activity consistently peaked around 23:00 across all subgroups. A marked increase was observed among adolescents (≤ 19) in August (males: IRR = 1.30, 95% CI [1.05–1.61]; females: IRR = 1.55, 95% CI [1.43–1.69]), while adults showed decreases in January–February. Weekly patterns varied by subgroup; for instance, males aged 20–29 posted more on Mondays and Tuesdays (IRR = 1.17, 95% CI [1.05–1.30]). Conclusions : Online forum activity displays predictable temporal cycles with demographic-specific patterns. These findings provide an essential baseline for real-time systems to detect deviations that may signal elevated suicide risk. The August peak in adolescent posts aligns with back-to-school distress, and the late-night peaks underscore the need for 24-hour support services. These insights can guide the development of targeted, time-sensitive suicide prevention strategies. Japan online forum temporal patterns diurnal variation mental health suicide suicidal ideation generalized additive model Figures Figure 1 Figure 2 Introduction Suicide is a serious and complex public health challenge with serious consequences not only for individual lives but also for society as a whole. The establishment of robust surveillance systems is essential for the rapid detection and prevention of suicide, and this is one of the key focus areas of global health policy [1,2]. Extensive research has documented temporal patterns of suicide, showing that the occurrence varies by season, day of the week and time of day [3,4]. Such patterns suggest the influence of both biological and environmental factors, with increased occurrence of suicide noted in spring and autumn, on Mondays and in the early morning [5,6]. In Japan, analysis of national data from 1974 to 2014 revealed significant diurnal and weekly variations in suicide rates [7]. The study found that suicide peaked on Mondays for all age groups, with females and older males having a higher incidence during the day, and young men aged 20–39 years were at highest risk during the late night hours. Such insights highlight the need to tailor time-sensitive intervention strategies to the time of day when suicide risk is high. While traditional data sources have provided valuable insights into temporal suicidal tendencies, the need to integrate digital data sources such as internet search queries and social media activities into surveillance activities is being recognized [8,9,10]. Online behavior can act as an early warning sign for suicidal ideation, providing real-time data that can enhance predictive models [11,12,13,14]. Despite this potential, research that explores cyclical patterns and demographic‑specific risks in online environments remains limited. This study aims to fill this gap by analysing postings on an online forum established by the Japanese public broadcaster “NHK” (NHK Forum), investigating the cyclical nature of posting behavior across different demographic backgrounds. By utilising this data, this study seeks to advance the early detection of suicide risk and lay the foundation for more effective prevention strategies tailored to specific temporal and demographic patterns. Methods Data Data were extracted from the NHK Forum, hosted on NHK’s “Facing Suicide” website, where users anonymously discuss mental health concerns, including suicidal ideation [15]. Data were extracted from the NHK Forum, hosted on NHK’s “Facing Suicide” website, where users anonymously discuss mental health concerns, including suicidal ideation [15]. Posts often include expressions of despair (e.g., “I want to die” or “I want to kill myself”), as well as accounts of severe stress in daily life (e.g., at school or work). Accordingly, the posts analyzed in this study include both content related to suicidal ideation and content reflecting high‑stress states. Posts containing inappropriate language (e.g., discriminatory expressions; violations of privacy or honor; threats of harm to self or others) may be withheld in whole or in part. All published content is publicly viewable. When posting, users select a username and optionally indicate gender (recorded as male/female in the platform), age, and place of residence. Timestamps are recorded to the minute. We analyzed post counts on the forum, regardless of content. The dataset comprised 90,900 posts from 1 January 2008 to 31 March 2025. We used gender, age, and timestamp variables and created eight strata by crossing gender (male, female) with age (≤19, 20s, 30s, ≥40). Posts from 1 January 2008 to 31 December 2012 (n = 137) were excluded because they were structured as answers to specific questions rather than open-ended posts. In addition, cases where the gender or age of the poster was unspecified (n = 27,717) were excluded. After applying these criteria, the final analytic sample included 63,046 posts. The daily trend in posting increased from 1 April 2019 onward (Additional file 1). This increase suggests that the NHK forum was not widely known before this date. Therefore, we performed additional analysis with a time span from 01 April 2019 to 31 March 2025. To account for COVID‑19, we also examined whether key trends held during a pandemic period defined as 1 February 2020–31 January 2023—spanning the month of the first confirmed COVID‑19 fatality in Japan [16] through the month when the government announced its policy to reclassify COVID‑19 as a “Class 5” infectious disease [17]. Statistical Analysis We used a generalized additive model (GAM) to model variation in post counts, as GAMs flexibly capture non‑linear relationships, including cyclical patterns in time‑series data [18]. Following prior work on suicide‑related posting patterns on Reddit [19], the outcome was the number of posts by time of day. Explanatory variables included month and day‑of‑week dummies; time of day was modeled with a spline term. We specified a Poisson distribution with a log link, so that coefficients are interpretable as log incidence rate ratios (IRRs). Analyses were conducted in R (version 4.3.0) using the mgcv package. Two‑sided p < 0.05 denoted statistical significance. No ethics review was required since this study used publicly available data. Results A total of 63,046 posts to the NHK Forum were analyzed (Table 1). Table 1. Number of posts and relative frequency by gender, age, month, day of the week and time of day (Hour) (Table 1 about here) Females were the predominant contributors, accounting for 75.5% (N=47,602) of posts. The largest age group was 20–29 years (32.4%, N=20,418), followed by those aged 40 and over (28.5%) and 19 and under (20.3%). Monthly postings peaked in August (9.5%, N=5,994), with high activity also seen in May (8.8%) and July (8.5%). Weekly patterns showed a peak on Monday (15.5%, N=9,748) and Sunday (14.9%), with the lowest activity on Friday and Saturday. A clear diurnal pattern was observed, with posting activity peaking late at night. The highest number of posts occurred at 23:00 (9.2%, N=5,771), followed by 22:00 (8.5%, N=5,355) and midnight (00:00) (8.5%, N=5,333). Conversely, the early morning hours (05:00–07:00) showed the lowest activity. Figures 1 and 2 show the results of the GAM analysis, modeling the variation in posting frequency. (Figure 1 about here) Figure 1 shows the smooth term for time of day, confirming diurnal variation. Across subgroups, activity was minimal in the early morning (approximately 03:00–08:00), rose over the day, and peaked around midnight. This pattern was consistent across age and gender strata. (Figure 2 about here) Figure 2 shows monthly and weekly effects by gender and age, presented as IRRs (IRR = exp(β)). August saw a significant surge in posts among young people, males ≤19 years (IRR = 1.30, 95% CI [1.05, 1.61]) and females ≤19 years (IRR = 1.55, 95% CI [1.43, 1.69]). Conversely, January and February showed significant decreases across several adult subgroups. This included males aged 20–29 (Jan: IRR=0.76, 95% CI [0.66, 0.88]; Feb: IRR=0.84, 95% CI [0.73, 0.97]), 30–39 (Jan: IRR=0.79, 95% CI [0.66, 0.95]; Feb: IRR=0.81, 95% CI [0.68, 0.96]), and ≥40 (Jan: IRR=0.83, 95% CI [0.73, 0.94]; Feb: IRR=0.85, 95% CI [0.75, 0.96]), as well as females aged 30–39 (Jan: IRR=0.82, 95% CI [0.74,0.91]; Feb: IRR=0.86, 95% CI [0.77,0.95]) and ≥40 (Jan: IRR=0.83, 95% CI [0.73, 0.94]; Feb: IRR=0.85, 95% CI [0.75, 0.96]). In June, significant decreases were observed for males aged 20–29 (IRR = 0.83, 95% CI [0.72, 0.95]) and for females in the ≤19 (IRR = 0.90, 95% CI [0.82,0.99]), 30–39 (IRR = 0.88, 95% CI [0.79,0.97]), and ≥40 (IRR = 0.77, 95% CI [0.70,0.84]) age groups. In December, declines were seen among both males (IRR = 0.79, 95% CI [0.68, 0.91]) and females (IRR = 0.90, 95% CI [0.84, 0.98]) in the 20–29 age group. In terms of weekly patterns, there were significantly fewer posts on Fridays and Saturdays compared to Sundays for all age groups of females. This trend was also observed for males aged ≥ 40 years. In contrast, males aged 20–29 posted significantly more on Monday (IRR = 1.17, 95% CI [1.05, 1.30]) and Tuesday (IRR = 1.17, 95% CI [1.05, 1.30]). The robustness of our main findings was confirmed by an additional analysis (Additional file 2), which replicated key results, including the August increase for users ≤19 and the weekly patterns for males aged 20–29. A separate analysis of the COVID-19 period (Additional file 3) showed less consistent results, likely due to wider confidence intervals arising from fewer posts, although the August increase for users ≤19 was still observed as a similar trend. Discussion We analyzed 63,046 posts from an online forum dedicated to mental health and suicide‑related topics and identified temporal and demographic patterns relevant to suicide risk surveillance. These findings have implications for digitally enabled prevention strategies and for understanding online behavior during mental health crises. Demographic patterns revealed that the majority of forum users were female (75.5%). The finding that the predominant contributors were female is consistent with previous research suggesting that women are more likely to seek support and express distress [ 20 , 21 ]. This highlights the importance of creating a digital environment that responds to women’s needs and provides a safe space where mental health issues can be discussed openly [ 22 , 23 ]. The age distribution shows that users aged 20–29 are the most active, accounting for 32.4% of the sample. This age group is not only familiar with digital media but is also often experiencing major life transitions, such as career pressures and changing relationships, which may increase their vulnerability to mental health challenges [ 24 ]. A key finding was the pronounced August increase in online posts among adolescents (≤ 19 years) of both genders. This timing coincides with the end of Japan’s summer vacation, when anticipatory stress about returning to school may intensify. Consistent with prior evidence, Google search queries related to school avoidance (e.g., “I do not want to go to school”) also rise in late August [ 25 ]. The convergence of these signals suggests that August postings may be an early indicator of the post–summer-break increase in student suicides observed in early September [ 26 , 27 ]. A plausible psychosocial explanation is the “broken-promise effect.” This term describes the let-down felt when the relief of the vacation ends and is replaced by worries about returning to school. In this view, the forum activity captures this pre-back-to-school distress. Consequently, the NHK Forum data may provide important insights into the mental health state of students and could potentially be utilized as a supplementary tool for suicide risk surveillance related to school life. An increase in postings was observed among females in their 20s during May and July. The May peak may be partly related to the phenomenon colloquially known in Japan as “May illness,” in which fatigue, low motivation, and mild depressive symptoms tend to emerge after the excitement of a new academic or work year subsides [ 28 ]. This condition is not specific to young women, but may manifest more prominently in certain groups depending on social expectations and roles. For the July increase, that potential explanations include seasonal factors such as heat and disrupted sleep [ 29 ], or examination stress prior to summer vacation [ 30 ]. These hypotheses should be interpreted with caution, and further studies are needed to clarify the mechanisms underlying these seasonal fluctuations. Conversely, January and February showed significant decreases in postings in several subgroups, including males in their 20s and 30s and over 40s, and females in their 30s and over 40s. These decreases may be attributed to the positive psychological impact of the New Year as a protective factor by encouraging people to set new goals and feel better [ 31 , 32 ]. Additionally, the New Year period is often marked by family reunions, which can have positive effects on mental well-being. For many, these reunions provide a sense of belonging and emotional support, which might temporarily alleviate psychological distress. On the other hand, the psychological distress that should have been reflected in the posts may have been under-represented in January and February, when people were preparing for the start of new lives, such as employment, transfers or enrolment, and did not have the psychological space to post in online forums. The pattern of weekly variation showed increased posting on Monday (IRR = 1.17, 95% CI [1.05, 1.30]) and Tuesday (IRR = 1.17, 95% CI [1.05, 1.30]) for males aged 20–29. The results supported these previous studies, as it is known that suicide rates increase at the beginning of the week due to the 'Blue Monday effect' [ 33 , 34 ]. The peak of diurnal variation was at midnight, from 22:00 to around midnight, for all subgroups. These late night peaks suggest that they may have been brought about by groups with high levels of loneliness and isolation, highlighting the need for mental health resources to be available 24 hours a day [ 35 ]. Some existing support services are only available during the hours between morning and evening, leaving a critical gap at night, when users feel most vulnerable. Expanding support services to cover these hours could help care for people with poor mental health. These results are not necessarily in line with existing studies. Existing studies on social media indicate a peak posting time of 5am in the early morning, which differs from the results of our analysis [ 19 ]. This may be due to the difference with social media users used in existing studies. On the other hand, our study used data from NHK's online forum, which likely attracts a diverse and representative audience. Compared to other media, which tend to be skewed towards younger age groups, NHK's message boards appear to be used by a broader general audience. Therefore, it is likely that our data more reflected the wide variety population with suicidal ideation compared to online communities with specific user groups. Limitations The present study has several limitations. First, the gender and age data are self-reported and may not be accurate. Since users can freely select these attributes, their veracity cannot be confirmed. Second, this study does not take into account user-specific information, such as the geographical location of the user or whether the contributor is a repeat user. It was not possible to obtain unique user information from the data used in this research. Investigating these factors in future studies would help to gain more insight into behavioral patterns associated with suicide risk. Third, this study did not analyze the content of the posts. Analysing the textual content could quickly capture the nature of discussions related to suicide and could provide advantages for understanding and addressing suicidal ideation in real time [ 36 ]. Future research should incorporate textual analysis to identify the specific stressors (e.g., academic pressure, loneliness, bullying) that underlie the statistical patterns found in this study. Fourth, our study was conducted using data from open-access noticeboards and did not capture discussions that occurred in closed communities or on the dark web. On these platforms, where barriers to access are high, conversations may be taking place that promote suicidal behavior and cannot be easily monitored using traditional methods [ 37 , 38 ]. To develop a comprehensive monitoring system for suicide risk, we should aim to integrate data from a wide range of online sources across cyberspace. Such an approach would allow emerging trends and high-risk behaviors to be more clearly identified and more effectively addressed. Conclusion This study analyzed over 63,000 posts on the NHK forum, highlighting the potential of such platforms for suicide risk surveillance. The primary contribution of this research is the clarification of predictable baseline patterns of online expressions of suicidal ideation; we have identified the typical yearly, weekly, and daily rhythms in which different demographic groups disclose distress. This baseline understanding is the essential foundation for building a real-time monitoring system. Such a system would function by detecting significant deviations from these established patterns, thereby signaling a potential surge in suicide risk. For instance, it could be crucial for estimating the Werther effect by tracking how posting patterns change immediately following media reports of a celebrity suicide, allowing for a rapid public health response. Future research, particularly the incorporation of textual content analysis, will further strengthen this system. By analyzing what is being said, not just when, the system could distinguish between general grief and high-risk suicidal ideation, enabling more targeted alerts. Ultimately, this line of research can help transform vast online data from a passive archive into an active tool for predictive suicide prevention. Abbreviations NHK: Japan Broadcasting Corporation; GAM: Generalized Additive Model; IRR: incidence rate ratio; CI: confidence interval. Declarations Ethics approval and consent to participate: This study was reviewed and approved by the Education Research Promotion Committee, Faculty of Management and Information, Tama University (Approval No. ERPC_R7_001). All procedures were conducted in accordance with the principles of the Declaration of Helsinki. As the study analyzed de-identified, publicly available data, the requirement for individual informed consent was waived by the committee. Consent for publication: Not applicable. The manuscript does not contain individual person’s data (including images, videos, or other identifying information). Availability of data and materials: The raw posts analyzed in this study are publicly available from the NHK “Facing Suicide” forum at: https://heart-net.nhk.or.jp/mukiau/message/ For transparency, the aggregated dataset used for the analyses (counts by gender, age, time, and date) and the R code are available from the corresponding author upon reasonable request. Competing interests: The authors declare that they have no competing interests. Funding: This work was supported by JSPS KAKENHI Grant Numbers JP21H04403, JP24K23761. The funding sources had no role in the study design, data collection, data analysis, data interpretation or preparation of the manuscript. Authors’ contributions: TA: conceptualization, methodology, data collection, formal analysis, visualization, writing—original draft. HS: contributed to the study concept and provided critical feedback on the interpretation of the analysis results. KY: provided critical feedback on the analysis results and substantially revised the discussion section. All authors read and approved the final manuscript. Acknowledgements : Not applicable. References Data and Surveillance Task Force of the National Action Alliance for Suicide Prevention. Improving national data systems for surveillance of suicide-related events. Am J Prev Med. 2014;47(3 Suppl 2):S122–9. doi:10.1016/j.amepre.2014.05.026. World Health Organization. LIVE LIFE: An implementation guide for suicide prevention in countries. 2021. https://www.who.int/publications/i/item/9789240026629. Accessed 8 May 2025. Christodoulou C, Douzenis A, Papadopoulos FC, et al. Suicide and seasonality. Acta Psychiatr Scand. 2012;125(2):127–46. doi:10.1111/j.1600-0447.2011.01750.x. Galvão P, Silva H, Silva C. Temporal distribution of suicide mortality: a systematic review. J Affect Disord. 2018;228:132–42. doi:10.1016/j.jad.2017.12.008. Bénard V, Geoffroy P, Bellivier F. Seasons, circadian rhythms, sleep and suicidal behaviors vulnerability. L’Encephale. 2015;41(4 Suppl 1):S29–37. doi:10.1016/S0013-7006(15)30004-X. Blenkiron P. The timing of deliberate self harm behaviour. Ir J Psychol Med. 2003;20:126–31. doi:10.1017/S079096670000793X. Boo J, Matsubayashi T, Ueda M. Diurnal variation in suicide timing by age and gender: evidence from Japan across 41 years. J Affect Disord. 2019;243:366–74. doi:10.1016/j.jad.2018.09.030. Lee KS, Lee H, Myung W, Song GY, Lee K, Kim H, Carroll BJ, Kim DK. Advanced daily prediction model for national suicide numbers with social media data. Psychiatry Investig. 2018;15(4):344–54. doi:10.30773/pi.2017.10.15. Castillo-Sánchez G, Marques G, Dorronzoro E, Rivera-Romero O, Franco-Martín M, Torre-Díez I. Suicide risk assessment using machine learning and social networks: a scoping review. J Med Syst. 2020;44:205. doi:10.1007/s10916-020-01669-5. Choi D, Sumner SA, Holland KM, Draper J, Murphy S, Bowen DA, et al. Development of a machine learning model using multiple, heterogeneous data sources to estimate weekly US suicide fatalities. JAMA Netw Open. 2020;3(12):e2030932. doi:10.1001/jamanetworkopen.2020.30932. Gunn JF, Lester D. Using Google searches on the internet to monitor suicidal behavior. J Affect Disord. 2013;148(2–3):411–2. doi:10.1016/j.jad.2012.11.004. Sueki H. Does the volume of internet searches using suicide-related search terms influence the suicide death rate? Data from 2004 to 2009 in Japan. Psychiatry Clin Neurosci. 2011;65(4):392–4. doi:10.1111/j.1440-1819.2011.02216.x. Taira K, Hosokawa R, Itatani T, Fujita S. Predicting the number of suicides in Japan using internet search queries: vector autoregression time series model. JMIR Public Health Surveill. 2021;7(12):e34016. doi:10.2196/34016. Michaels MS, Chu C, Silva C, Schulman BE, Joiner T. Considerations regarding online methods for suicide-related research and suicide risk assessment. Suicide Life Threat Behav. 2015;45(1):10–7. doi:10.1111/sltb.12105. NHK (Japan Broadcasting Corporation). Thinking about mental health. 2023. https://heart-net.nhk.or.jp/mukiau/. Accessed 1 April 2025. Prime Minister’s Office of Japan. Novel Coronavirus Expert Meeting (the 1st meeting). 2020. https://www.kantei.go.jp/jp/98_abe/actions/202002/16corona_sen.html. Accessed 1 April 2025. Prime Minister’s Office of Japan. Novel Coronavirus Response Headquarters (the 101st meeting). 2023. https://www.kantei.go.jp/jp/101_kishida/actions/202301/27corona.html. Accessed 1 April 2025. Wood SN. Generalized additive models: an introduction with R. 2nd ed. Boca Raton: Chapman and Hall/CRC; 2017. doi:10.1201/9781315370279. Dutta R, Gkotsis G, Velupillai S, Bakolis I, Stewart R. Temporal and diurnal variation in social media posts to a suicide support forum. BMC Psychiatry. 2021;21:259. doi:10.1186/s12888-021-03268-1. Li H, Sun J, Zhang Q, Wei D, Li W, Jackson T, Hitchman G, Qiu J. Neuroanatomical differences between men and women in help-seeking coping strategy. Sci Rep. 2014;4:5700. doi:10.1038/srep05700. Tamres L, Janicki D, Helgeson V. Sex differences in coping behavior: a meta-analytic review and an examination of relative coping. Pers Soc Psychol Rev. 2002;6:2–30. doi:10.1207/S15327957PSPR0601_1. Naslund JA, Aschbrenner KA, Marsch LA, Bartels SJ. The future of mental health care: peer-to-peer support and social media. Epidemiol Psychiatr Sci. 2016;25(2):113–22. doi:10.1017/S2045796015001067. Pretorius C, Chambers D, Coyle D. Young people’s online help-seeking and mental health difficulties: systematic narrative review. J Med Internet Res. 2019;21(11):e13873. doi:10.2196/13873. Khan A, Rehman A. Impact of difficulties faced by adolescents in making career decision on their mental health. J Educ Vocat Res. 2018;9(2):1–8. doi:10.22610/jevr.v9i2(V).2789. Arai T, Tsubaki H, Wakano A, Shimizu Y. Association between school-related Google Trends search volume and suicides among children and adolescents in Japan during 2016–2020: retrospective observational study with a time-series analysis. J Med Internet Res. 2024;26:e51710. doi:10.2196/51710. Matsubayashi T, Ueda M, Yoshikawa K. School and seasonality in youth suicide: evidence from Japan. J Epidemiol Community Health. 2016;70:1122–7. doi:10.1136/jech-2016-207583. Gabennesch H. When promises fail: a theory of temporal fluctuations in suicide. Soc Forces. 1988;67(1):129–45. doi:10.2307/2579103. Shiraishi N, Sakata M, Toyomoto R, Yoshida K, Luo Y, Nakagami Y, et al. Dynamics of depressive states among university students in Japan during the COVID-19 pandemic: an interrupted time series analysis. Ann Gen Psychiatry. 2023;22:38. doi:10.1186/s12991-023-00468-9. Fujii H, Fukuda S, Narumi D, Ihara T, Watanabe Y. Fatigue and sleep under large summer temperature differences. Environ Res. 2015;138:17–21. doi:10.1016/j.envres.2015.02.006. Zunhammer M, Eichhammer P, Busch V. Sleep quality during exam stress: the role of alcohol, caffeine, and nicotine. PLoS One. 2014;9(10):e109490. doi:10.1371/journal.pone.0109490. Dai H, Milkman K, Riis J. The fresh start effect: temporal landmarks motivate aspirational behavior. Manag Sci. 2014;60(10):2563–82. doi:10.1287/mnsc.2014.1901. Páez D, Bilbao M, Bobowik M, Campos M, Basabe N. Merry Christmas and Happy New Year! The impact of Christmas rituals on subjective well-being and family’s emotional climate. Rev Psicol Soc. 2011;26:373–86. doi:10.1174/021347411797361347. Maldonado G, Kraus J. Variation in suicide occurrence by time of day, day of the week, month, and lunar phase. Suicide Life Threat Behav. 1991;21(2):174–87. doi:10.1111/j.1943-278X.1991.tb00464.x. Kim E, Cho S, Na K, Jung H, Lee K, Cho S, Han D. Blue Monday is real for suicide: a case–control study of 188,601 suicides. Suicide Life Threat Behav. 2019;49:393–400. doi:10.1111/sltb.12429. Hawkley LC, Preacher KJ, Cacioppo JT. Loneliness impairs daytime functioning but not sleep duration. Health Psychol. 2010;29(2):124–9. doi:10.1037/a0018646. Franz PJ, Nook EC, Mair P, Nock MK. Using topic modeling to detect and describe self-injurious and related content on a large-scale digital platform. Suicide Life Threat Behav. 2020;50(1):5–18. doi:10.1111/sltb.12569. Liu J. Need to establish a new adolescent suicide prevention programme in South Korea. Gen Psychiatry. 2020;33(4):e100200. doi:10.1136/gpsych-2020-100200. Mörch CM, Côté LP, Corthésy-Blondin L, Plourde-Léveillé L, Dargis L, Mishara BL. The Darknet and suicide. J Affect Disord. 2018;241:127–32. doi:10.1016/j.jad.2018.08.028. Table Table 1 is available in the Supplementary Files section. Additional Declarations No competing interests reported. Supplementary Files Table1.xlsx Table1. Number of posts and relative frequency by gender, age, month, day of the week and time of day (Hour) Additionalfile1.docx Additionalfile2.docx Additionalfile3.docx Cite Share Download PDF Status: Published Journal Publication published 20 Nov, 2025 Read the published version in BMC Psychiatry → Version 1 posted Editorial decision: Revision requested 01 Oct, 2025 Reviews received at journal 29 Sep, 2025 Reviewers agreed at journal 21 Sep, 2025 Reviewers agreed at journal 21 Sep, 2025 Reviews received at journal 20 Sep, 2025 Reviewers agreed at journal 19 Sep, 2025 Reviewers invited by journal 19 Sep, 2025 Editor assigned by journal 19 Sep, 2025 Editor invited by journal 11 Sep, 2025 Submission checks completed at journal 11 Sep, 2025 First submitted to journal 11 Sep, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7442638","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":523059838,"identity":"937ada75-c161-401f-a10a-49b91115b479","order_by":0,"name":"Takahiro Arai","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAr0lEQVRIiWNgGAWjYJCCAwwMNgYMzECWBPFaEtJI1MLAkHDYgHjF/LMPPzzA+OO8MT878wEGyx1EaJE4l2YAdNhtM8lmtgQGyTPEWHOGAazFxuAwjwGDZBsROuTPsH8AajlnY0+0FoMzPCBbDpgZMBOrxfAMT8GBhLRkY4nDbAkHiPKL3Bn2zR8+2NgZ9vcfPvhYkpgQA4MEKH1YsoFYLTDA+JFkLaNgFIyCUTASAAARCTJHoW7VywAAAABJRU5ErkJggg==","orcid":"","institution":"Tama University","correspondingAuthor":true,"prefix":"","firstName":"Takahiro","middleName":"","lastName":"Arai","suffix":""},{"id":523059839,"identity":"a1298003-e7f0-43f9-9660-e0510f940ed1","order_by":1,"name":"Hiroyuki Shinkai","email":"","orcid":"","institution":"Kanagawa University","correspondingAuthor":false,"prefix":"","firstName":"Hiroyuki","middleName":"","lastName":"Shinkai","suffix":""},{"id":523059840,"identity":"5135482e-86e5-41e8-a0df-4e8c5027a7e7","order_by":2,"name":"Keita Yamauchi","email":"","orcid":"","institution":"Keio University","correspondingAuthor":false,"prefix":"","firstName":"Keita","middleName":"","lastName":"Yamauchi","suffix":""}],"badges":[],"createdAt":"2025-08-23 17:23:17","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7442638/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7442638/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12888-025-07539-z","type":"published","date":"2025-11-20T15:58:10+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":92595464,"identity":"1ab42a53-05ba-4be8-bf66-eff1ef529a72","added_by":"auto","created_at":"2025-10-01 13:06:58","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":54254,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscriptrevised.docx","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/eecc675009c2eb071f8ca7c9.docx"},{"id":92596101,"identity":"19ab505e-dbca-4fc8-bc9c-b4abcf12fde2","added_by":"auto","created_at":"2025-10-01 13:14:58","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11115,"visible":true,"origin":"","legend":"","description":"","filename":"Table1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/6e79e95cd5a3b79a4b899860.xlsx"},{"id":92594374,"identity":"56f3a004-e32d-4b1f-a065-b5803caf4bbe","added_by":"auto","created_at":"2025-10-01 12:50:58","extension":"json","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5625,"visible":true,"origin":"","legend":"","description":"","filename":"72fd5c49aa5e42e4a89342d69fa69a78.json","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/c9bdbbb2c83783a10e73927a.json"},{"id":92595274,"identity":"95206248-756a-47fb-8a1f-ecce6a5a5637","added_by":"auto","created_at":"2025-10-01 12:58:58","extension":"docx","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":63722,"visible":true,"origin":"","legend":"","description":"","filename":"Additionalfile1.docx","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/cc332d9e3ef4978c9ff8f147.docx"},{"id":92595277,"identity":"76b2e3c1-76f8-4c9e-99e5-b4bc33f63b4a","added_by":"auto","created_at":"2025-10-01 12:58:58","extension":"docx","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1285478,"visible":true,"origin":"","legend":"","description":"","filename":"Additionalfile2.docx","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/f1152959bb5127bd992aebff.docx"},{"id":92594388,"identity":"c9ea875c-73d6-45a6-982d-f583a87afb09","added_by":"auto","created_at":"2025-10-01 12:50:58","extension":"docx","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1328283,"visible":true,"origin":"","legend":"","description":"","filename":"Additionalfile3.docx","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/051745e60d74fcaccc030a36.docx"},{"id":92595466,"identity":"532fa513-33a9-49e7-bee7-6d1e6a4d7a9a","added_by":"auto","created_at":"2025-10-01 13:06:58","extension":"xml","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":86061,"visible":true,"origin":"","legend":"","description":"","filename":"72fd5c49aa5e42e4a89342d69fa69a781enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/2588208f0425c3bba74502ff.xml"},{"id":92594384,"identity":"27f9525a-b9ce-4cc7-b675-5ad293ac3588","added_by":"auto","created_at":"2025-10-01 12:50:58","extension":"pdf","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":324359,"visible":true,"origin":"","legend":"","description":"","filename":"Figure1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/c627560d0683c6deb93c1e8f.pdf"},{"id":92595278,"identity":"a76d55cf-81fc-45ca-aec3-eb81be764858","added_by":"auto","created_at":"2025-10-01 12:58:58","extension":"pdf","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":274474,"visible":true,"origin":"","legend":"","description":"","filename":"Figure2.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/db0bd8c150e0ce2daba274ca.pdf"},{"id":92594383,"identity":"4f12e99e-0717-446c-90c3-baba3aec6927","added_by":"auto","created_at":"2025-10-01 12:50:58","extension":"xml","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":84426,"visible":true,"origin":"","legend":"","description":"","filename":"72fd5c49aa5e42e4a89342d69fa69a781structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/2e32a403cc5b10dd4eba5e52.xml"},{"id":92594387,"identity":"84804713-f0a3-4a06-840e-bee646c639c4","added_by":"auto","created_at":"2025-10-01 12:50:58","extension":"html","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":93587,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/419441d3bf7017eddba22a43.html"},{"id":92594371,"identity":"82447f31-adaf-4a84-8d8c-53a86945796d","added_by":"auto","created_at":"2025-10-01 12:50:58","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":2293844,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSpline Estimates for Time Effects by Gender and Age Group\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis figure shows the estimated effects and 95% confidence intervals for time-of-day spline terms from a Generalized Additive Model (GAM). The analysis is segmented by gender and age group (≤19, 20s, 30s, ≥40). Estimates reflect the smoothed influence of time-of-day variations on the response variable.\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/f91c8f845c25b4b1edfb58d2.png"},{"id":92594380,"identity":"a550a969-861b-434d-b6d1-d4a2673442aa","added_by":"auto","created_at":"2025-10-01 12:50:58","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":2061458,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eEstimates for Monthly and Weekday Dummies by Gender and Age Group\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis figureshows estimated coefficients and 95% confidence intervals for monthly and weeklydummies from a Generalized Additive Model (GAM), stratified by gender and age group (≤19, 20s, 30s, ≥40). The reference categories are March for months and Sunday for day of the week. The results are presented as incidence rate ratios (IRRs). Positive estimates indicate higher values relative to the reference, while negative estimates indicate lower values. Confidence intervals excluding 1.0 denote statistical significance at the 5% level.\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/ccfb5a84bef449c6e62163e0.png"},{"id":96651091,"identity":"e8973079-13a4-4856-808f-c6dc1004d884","added_by":"auto","created_at":"2025-11-24 16:13:49","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4783173,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/b4cd3168-46fd-48a8-8766-12ce6b857379.pdf"},{"id":92595273,"identity":"3b79852f-b871-4110-9ab8-634c566abc27","added_by":"auto","created_at":"2025-10-01 12:58:58","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":11115,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTable1. Number of posts and relative frequency by gender, age, month, day of the week and time of day (Hour)\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Table1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/2bbf35f3a581f63129c3c44e.xlsx"},{"id":92594372,"identity":"2d61a9d6-7d08-4283-8420-79f6cd172519","added_by":"auto","created_at":"2025-10-01 12:50:58","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":63722,"visible":true,"origin":"","legend":"","description":"","filename":"Additionalfile1.docx","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/0823b4027c6d164fca37eb84.docx"},{"id":92594379,"identity":"27eb089e-8330-49ee-9f3b-4e28d6fa4781","added_by":"auto","created_at":"2025-10-01 12:50:58","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":1285478,"visible":true,"origin":"","legend":"","description":"","filename":"Additionalfile2.docx","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/c82654a409326c0b8e311b9a.docx"},{"id":92594381,"identity":"9f289fff-0e6b-43d6-8966-558a65a6f057","added_by":"auto","created_at":"2025-10-01 12:50:58","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":1328283,"visible":true,"origin":"","legend":"","description":"","filename":"Additionalfile3.docx","url":"https://assets-eu.researchsquare.com/files/rs-7442638/v1/72df0a90f036107312849775.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Temporal Analysis of a Japanese Online Forum for Suicide Risk Monitoring","fulltext":[{"header":"Introduction","content":"\u003cp\u003eSuicide is a serious and complex public health challenge with serious consequences not only for individual lives but also for society as a whole. The establishment of robust surveillance systems is essential for the rapid detection and prevention of suicide, and this is one of the key focus areas of global health policy [1,2]. Extensive research has documented temporal patterns of suicide, showing that the occurrence varies by season, day of the week and time of day [3,4]. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSuch patterns suggest the influence of both biological and environmental factors, with increased occurrence of suicide noted in spring and autumn, on Mondays and in the early morning [5,6]. In Japan, analysis of national data from 1974 to 2014 revealed significant diurnal and weekly variations in suicide rates [7]. The study found that suicide peaked on Mondays for all age groups, with females and older males having a higher incidence during the day, and young men aged 20\u0026ndash;39 years were at highest risk during the late night hours. Such insights highlight the need to tailor time-sensitive intervention strategies to the time of day when suicide risk is high.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWhile traditional data sources have provided valuable insights into temporal suicidal tendencies, the need to integrate digital data sources such as internet search queries and social media activities into surveillance activities is being recognized [8,9,10].\u003c/p\u003e\n\u003cp\u003eOnline behavior can act as an early warning sign for suicidal ideation, providing real-time data that can enhance predictive models [11,12,13,14]. Despite this potential, research that explores cyclical patterns and demographic‑specific risks in online environments remains limited.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThis study aims to fill this gap by analysing postings on an online forum established by the Japanese public broadcaster \u0026ldquo;NHK\u0026rdquo; (NHK Forum), investigating the cyclical nature of posting behavior across different demographic backgrounds. By utilising this data, this study seeks to advance the early detection of suicide risk and lay the foundation for more effective prevention strategies tailored to specific temporal and demographic patterns.\u003c/p\u003e"},{"header":"Methods","content":"\u003ch2\u003eData\u003c/h2\u003e\n\u003cp\u003eData were extracted from the NHK Forum, hosted on NHK\u0026rsquo;s \u0026ldquo;Facing Suicide\u0026rdquo; website, where users anonymously discuss mental health concerns, including suicidal ideation [15].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eData were extracted from the NHK Forum, hosted on NHK\u0026rsquo;s \u0026ldquo;Facing Suicide\u0026rdquo; website, where users anonymously discuss mental health concerns, including suicidal ideation [15]. Posts often include expressions of despair (e.g., \u0026ldquo;I want to die\u0026rdquo; or \u0026ldquo;I want to kill myself\u0026rdquo;), as well as accounts of severe stress in daily life (e.g., at school or work). Accordingly, the posts analyzed in this study include both content related to suicidal ideation and content reflecting high‑stress states. Posts containing inappropriate language (e.g., discriminatory expressions; violations of privacy or honor; threats of harm to self or others) may be withheld in whole or in part. All published content is publicly viewable. When posting, users select a username and optionally indicate gender (recorded as male/female in the platform), age, and place of residence. Timestamps are recorded to the minute.\u003c/p\u003e\n\u003cp\u003eWe analyzed post counts on the forum, regardless of content.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe dataset comprised 90,900 posts from 1 January 2008 to 31 March 2025. We used gender, age, and timestamp variables and created eight strata by crossing gender (male, female) with age (\u0026le;19, 20s, 30s, \u0026ge;40).\u003c/p\u003e\n\u003cp\u003ePosts from 1 January 2008 to 31 December 2012 (n = 137) were excluded because they were structured as answers to specific questions rather than open-ended posts. In addition, cases where the gender or age of the poster was unspecified (n = 27,717) were excluded. After applying these criteria, the final analytic sample included 63,046 posts. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe daily trend in posting increased from 1 April 2019 onward (Additional file 1). This increase suggests that the NHK forum was not widely known before this date. Therefore, we performed additional analysis with a time span from 01 April 2019 to 31 March 2025.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo account for COVID‑19, we also examined whether key trends held during a pandemic period defined as 1 February 2020\u0026ndash;31 January 2023\u0026mdash;spanning the month of the first confirmed COVID‑19 fatality in Japan [16] through the month when the government announced its policy to reclassify COVID‑19 as a \u0026ldquo;Class 5\u0026rdquo; infectious disease [17].\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eStatistical Analysis\u003c/h2\u003e\n\u003cp\u003eWe used a generalized additive model (GAM) to model variation in post counts, as GAMs flexibly capture non‑linear relationships, including cyclical patterns in time‑series data [18]. Following prior work on suicide‑related posting patterns on Reddit [19], the outcome was the number of posts by time of day. Explanatory variables included month and day‑of‑week dummies; time of day was modeled with a spline term. We specified a Poisson distribution with a log link, so that coefficients are interpretable as log incidence rate ratios (IRRs). Analyses were conducted in R (version 4.3.0) using the mgcv package. Two‑sided p \u0026lt; 0.05 denoted statistical significance.\u003c/p\u003e\n\u003cp\u003eNo ethics review was required since this study used publicly available data.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eA total of 63,046 posts to the NHK Forum were analyzed (Table 1).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 1. Number of posts and relative frequency by gender, age, month, day of the week and time of day (Hour)\u003c/p\u003e\n\u003cp\u003e(Table 1 about here)\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFemales were the predominant contributors, accounting for 75.5% (N=47,602) of posts. The largest age group was 20\u0026ndash;29 years (32.4%, N=20,418), followed by those aged 40 and over (28.5%) and 19 and under (20.3%).\u003c/p\u003e\n\u003cp\u003eMonthly postings peaked in August (9.5%, N=5,994), with high activity also seen in May (8.8%) and July (8.5%). Weekly patterns showed a peak on Monday (15.5%, N=9,748) and Sunday (14.9%), with the lowest activity on Friday and Saturday. A clear diurnal pattern was observed, with posting activity peaking late at night. The highest number of posts occurred at 23:00 (9.2%, N=5,771), followed by 22:00 (8.5%, N=5,355) and midnight (00:00) (8.5%, N=5,333). Conversely, the early morning hours (05:00\u0026ndash;07:00) showed the lowest activity.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFigures 1 and 2 show the results of the GAM analysis, modeling the variation in posting frequency.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(Figure 1 about here)\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFigure 1 shows the smooth term for time of day, confirming diurnal variation. Across subgroups, activity was minimal in the early morning (approximately 03:00\u0026ndash;08:00), rose over the day, and peaked around midnight. This pattern was consistent across age and gender strata.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(Figure 2 about here)\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFigure 2 shows monthly and weekly effects by gender and age, presented as IRRs (IRR = exp(\u0026beta;)).\u003c/p\u003e\n\u003cp\u003eAugust saw a significant surge in posts among young people, males \u0026le;19 years (IRR = 1.30, 95% CI [1.05, 1.61]) and females \u0026le;19 years (IRR = 1.55, 95% CI [1.43, 1.69]).\u003c/p\u003e\n\u003cp\u003eConversely, January and February showed significant decreases across several adult subgroups. This included males aged 20\u0026ndash;29 (Jan: IRR=0.76, 95% CI [0.66, 0.88]; Feb: IRR=0.84, 95% CI [0.73, 0.97]), 30\u0026ndash;39 (Jan: IRR=0.79, 95% CI [0.66, 0.95]; Feb: IRR=0.81, 95% CI [0.68, 0.96]), and \u0026ge;40 (Jan: IRR=0.83, 95% CI [0.73, 0.94]; Feb: IRR=0.85, 95% CI [0.75, 0.96]), as well as females aged 30\u0026ndash;39 (Jan: IRR=0.82, 95% \u0026nbsp;CI [0.74,0.91]; Feb: IRR=0.86, 95% CI [0.77,0.95]) and \u0026ge;40 (Jan: IRR=0.83, 95% CI [0.73, 0.94]; Feb: IRR=0.85, 95% CI [0.75, 0.96]).\u003c/p\u003e\n\u003cp\u003eIn June, significant decreases were observed for males aged 20\u0026ndash;29 (IRR = 0.83, 95% CI [0.72, 0.95]) and for females in the \u0026le;19 (IRR = 0.90, 95% CI [0.82,0.99]), 30\u0026ndash;39 (IRR = 0.88, 95% CI [0.79,0.97]), and \u0026ge;40 (IRR = 0.77, 95% CI [0.70,0.84]) age groups.\u003c/p\u003e\n\u003cp\u003eIn December, declines were seen among both males (IRR = 0.79, 95% CI [0.68, 0.91]) and females (IRR = 0.90, 95% CI [0.84, 0.98]) in the 20\u0026ndash;29 age group.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn terms of weekly patterns, there were significantly fewer posts on Fridays and Saturdays compared to Sundays for all age groups of females. This trend was also observed for males aged \u0026ge; 40 years. In contrast, males aged 20\u0026ndash;29 posted significantly more on Monday (IRR = 1.17, 95% CI [1.05, 1.30]) and Tuesday (IRR = 1.17, 95% CI [1.05, 1.30]).\u003c/p\u003e\n\u003cp\u003eThe robustness of our main findings was confirmed by an additional analysis (Additional file 2), which replicated key results, including the August increase for users \u0026le;19 and the weekly patterns for males aged 20\u0026ndash;29. A separate analysis of the COVID-19 period (Additional file 3) showed less consistent results, likely due to wider confidence intervals arising from fewer posts, although the August increase for users \u0026le;19 was still observed as a similar trend.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eWe analyzed 63,046 posts from an online forum dedicated to mental health and suicide‑related topics and identified temporal and demographic patterns relevant to suicide risk surveillance. These findings have implications for digitally enabled prevention strategies and for understanding online behavior during mental health crises.\u003c/p\u003e\u003cp\u003eDemographic patterns revealed that the majority of forum users were female (75.5%). The finding that the predominant contributors were female is consistent with previous research suggesting that women are more likely to seek support and express distress [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. This highlights the importance of creating a digital environment that responds to women\u0026rsquo;s needs and provides a safe space where mental health issues can be discussed openly [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eThe age distribution shows that users aged 20\u0026ndash;29 are the most active, accounting for 32.4% of the sample. This age group is not only familiar with digital media but is also often experiencing major life transitions, such as career pressures and changing relationships, which may increase their vulnerability to mental health challenges [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eA key finding was the pronounced August increase in online posts among adolescents (\u0026le;\u0026thinsp;19 years) of both genders. This timing coincides with the end of Japan\u0026rsquo;s summer vacation, when anticipatory stress about returning to school may intensify. Consistent with prior evidence, Google search queries related to school avoidance (e.g., \u0026ldquo;I do not want to go to school\u0026rdquo;) also rise in late August [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. The convergence of these signals suggests that August postings may be an early indicator of the post\u0026ndash;summer-break increase in student suicides observed in early September [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. A plausible psychosocial explanation is the \u0026ldquo;broken-promise effect.\u0026rdquo; This term describes the let-down felt when the relief of the vacation ends and is replaced by worries about returning to school. In this view, the forum activity captures this pre-back-to-school distress. Consequently, the NHK Forum data may provide important insights into the mental health state of students and could potentially be utilized as a supplementary tool for suicide risk surveillance related to school life.\u003c/p\u003e\u003cp\u003eAn increase in postings was observed among females in their 20s during May and July. The May peak may be partly related to the phenomenon colloquially known in Japan as \u0026ldquo;May illness,\u0026rdquo; in which fatigue, low motivation, and mild depressive symptoms tend to emerge after the excitement of a new academic or work year subsides [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. This condition is not specific to young women, but may manifest more prominently in certain groups depending on social expectations and roles. For the July increase, that potential explanations include seasonal factors such as heat and disrupted sleep [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e], or examination stress prior to summer vacation [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. These hypotheses should be interpreted with caution, and further studies are needed to clarify the mechanisms underlying these seasonal fluctuations.\u003c/p\u003e\u003cp\u003eConversely, January and February showed significant decreases in postings in several subgroups, including males in their 20s and 30s and over 40s, and females in their 30s and over 40s. These decreases may be attributed to the positive psychological impact of the New Year as a protective factor by encouraging people to set new goals and feel better [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. Additionally, the New Year period is often marked by family reunions, which can have positive effects on mental well-being. For many, these reunions provide a sense of belonging and emotional support, which might temporarily alleviate psychological distress. On the other hand, the psychological distress that should have been reflected in the posts may have been under-represented in January and February, when people were preparing for the start of new lives, such as employment, transfers or enrolment, and did not have the psychological space to post in online forums.\u003c/p\u003e\u003cp\u003eThe pattern of weekly variation showed increased posting on Monday (IRR\u0026thinsp;=\u0026thinsp;1.17, 95% CI [1.05, 1.30]) and Tuesday (IRR\u0026thinsp;=\u0026thinsp;1.17, 95% CI [1.05, 1.30]) for males aged 20\u0026ndash;29. The results supported these previous studies, as it is known that suicide rates increase at the beginning of the week due to the 'Blue Monday effect' [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eThe peak of diurnal variation was at midnight, from 22:00 to around midnight, for all subgroups. These late night peaks suggest that they may have been brought about by groups with high levels of loneliness and isolation, highlighting the need for mental health resources to be available 24 hours a day [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eSome existing support services are only available during the hours between morning and evening, leaving a critical gap at night, when users feel most vulnerable. Expanding support services to cover these hours could help care for people with poor mental health.\u003c/p\u003e\u003cp\u003eThese results are not necessarily in line with existing studies. Existing studies on social media indicate a peak posting time of 5am in the early morning, which differs from the results of our analysis [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. This may be due to the difference with social media users used in existing studies. On the other hand, our study used data from NHK's online forum, which likely attracts a diverse and representative audience. Compared to other media, which tend to be skewed towards younger age groups, NHK's message boards appear to be used by a broader general audience. Therefore, it is likely that our data more reflected the wide variety population with suicidal ideation compared to online communities with specific user groups.\u003c/p\u003e\n\u003ch3\u003eLimitations\u003c/h3\u003e\n\u003cp\u003eThe present study has several limitations.\u003c/p\u003e\u003cp\u003eFirst, the gender and age data are self-reported and may not be accurate. Since users can freely select these attributes, their veracity cannot be confirmed.\u003c/p\u003e\u003cp\u003eSecond, this study does not take into account user-specific information, such as the geographical location of the user or whether the contributor is a repeat user. It was not possible to obtain unique user information from the data used in this research. Investigating these factors in future studies would help to gain more insight into behavioral patterns associated with suicide risk.\u003c/p\u003e\u003cp\u003eThird, this study did not analyze the content of the posts. Analysing the textual content could quickly capture the nature of discussions related to suicide and could provide advantages for understanding and addressing suicidal ideation in real time [\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. Future research should incorporate textual analysis to identify the specific stressors (e.g., academic pressure, loneliness, bullying) that underlie the statistical patterns found in this study.\u003c/p\u003e\u003cp\u003eFourth, our study was conducted using data from open-access noticeboards and did not capture discussions that occurred in closed communities or on the dark web. On these platforms, where barriers to access are high, conversations may be taking place that promote suicidal behavior and cannot be easily monitored using traditional methods [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. To develop a comprehensive monitoring system for suicide risk, we should aim to integrate data from a wide range of online sources across cyberspace. Such an approach would allow emerging trends and high-risk behaviors to be more clearly identified and more effectively addressed.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study analyzed over 63,000 posts on the NHK forum, highlighting the potential of such platforms for suicide risk surveillance. The primary contribution of this research is the clarification of predictable baseline patterns of online expressions of suicidal ideation; we have identified the typical yearly, weekly, and daily rhythms in which different demographic groups disclose distress.\u003c/p\u003e\u003cp\u003eThis baseline understanding is the essential foundation for building a real-time monitoring system. Such a system would function by detecting significant deviations from these established patterns, thereby signaling a potential surge in suicide risk. For instance, it could be crucial for estimating the Werther effect by tracking how posting patterns change immediately following media reports of a celebrity suicide, allowing for a rapid public health response. Future research, particularly the incorporation of textual content analysis, will further strengthen this system. By analyzing what is being said, not just when, the system could distinguish between general grief and high-risk suicidal ideation, enabling more targeted alerts. Ultimately, this line of research can help transform vast online data from a passive archive into an active tool for predictive suicide prevention.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eNHK: Japan Broadcasting Corporation; GAM: Generalized Additive Model; IRR: incidence rate ratio; CI: confidence interval.\u003c/p\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate:\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was reviewed and approved by the Education Research Promotion Committee, Faculty of Management and Information, Tama University (Approval No. ERPC_R7_001). All procedures were conducted in accordance with the principles of the Declaration of Helsinki. As the study analyzed de-identified, publicly available data, the requirement for individual informed consent was waived by the committee. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication:\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNot applicable. The manuscript does not contain individual person\u0026rsquo;s data (including images, videos, or other identifying information).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe raw posts analyzed in this study are publicly available from the NHK \u0026ldquo;Facing Suicide\u0026rdquo; forum at: https://heart-net.nhk.or.jp/mukiau/message/\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFor transparency, the aggregated dataset used for the analyses (counts by gender, age, time, and date) and the R code are available from the corresponding author upon reasonable request.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests:\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThis work was supported by JSPS KAKENHI Grant Numbers JP21H04403, JP24K23761.\u0026nbsp;The funding sources had no role in the study design, data collection, data analysis, data interpretation or preparation of the manuscript.\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026rsquo; contributions:\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTA: conceptualization, methodology, data collection, formal analysis, visualization, writing\u0026mdash;original draft.\u003c/p\u003e\n\u003cp\u003eHS: contributed to the study concept and provided critical feedback on the interpretation of the analysis results.\u003c/p\u003e\n\u003cp\u003eKY: provided critical feedback on the analysis results and substantially revised the discussion section.\u003c/p\u003e\n\u003cp\u003eAll authors read and approved the final manuscript.\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e:\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eData and Surveillance Task Force of the National Action Alliance for Suicide Prevention. Improving national data systems for surveillance of suicide-related events. Am J Prev Med. 2014;47(3 Suppl 2):S122\u0026ndash;9. doi:10.1016/j.amepre.2014.05.026.\u003c/li\u003e\n\u003cli\u003eWorld Health Organization. LIVE LIFE: An implementation guide for suicide prevention in countries. 2021. https://www.who.int/publications/i/item/9789240026629. Accessed 8 May 2025.\u003c/li\u003e\n\u003cli\u003eChristodoulou C, Douzenis A, Papadopoulos FC, et al. Suicide and seasonality. Acta Psychiatr Scand. 2012;125(2):127\u0026ndash;46. doi:10.1111/j.1600-0447.2011.01750.x.\u003c/li\u003e\n\u003cli\u003eGalv\u0026atilde;o P, Silva H, Silva C. Temporal distribution of suicide mortality: a systematic review. J Affect Disord. 2018;228:132\u0026ndash;42. doi:10.1016/j.jad.2017.12.008.\u003c/li\u003e\n\u003cli\u003eB\u0026eacute;nard V, Geoffroy P, Bellivier F. Seasons, circadian rhythms, sleep and suicidal behaviors vulnerability. L\u0026rsquo;Encephale. 2015;41(4 Suppl 1):S29\u0026ndash;37. doi:10.1016/S0013-7006(15)30004-X.\u003c/li\u003e\n\u003cli\u003eBlenkiron P. The timing of deliberate self harm behaviour. Ir J Psychol Med. 2003;20:126\u0026ndash;31. doi:10.1017/S079096670000793X.\u003c/li\u003e\n\u003cli\u003eBoo J, Matsubayashi T, Ueda M. Diurnal variation in suicide timing by age and gender: evidence from Japan across 41 years. J Affect Disord. 2019;243:366\u0026ndash;74. doi:10.1016/j.jad.2018.09.030.\u003c/li\u003e\n\u003cli\u003eLee KS, Lee H, Myung W, Song GY, Lee K, Kim H, Carroll BJ, Kim DK. Advanced daily prediction model for national suicide numbers with social media data. Psychiatry Investig. 2018;15(4):344\u0026ndash;54. doi:10.30773/pi.2017.10.15.\u003c/li\u003e\n\u003cli\u003eCastillo-S\u0026aacute;nchez G, Marques G, Dorronzoro E, Rivera-Romero O, Franco-Mart\u0026iacute;n M, Torre-D\u0026iacute;ez I. Suicide risk assessment using machine learning and social networks: a scoping review. J Med Syst. 2020;44:205. doi:10.1007/s10916-020-01669-5.\u003c/li\u003e\n\u003cli\u003eChoi D, Sumner SA, Holland KM, Draper J, Murphy S, Bowen DA, et al. Development of a machine learning model using multiple, heterogeneous data sources to estimate weekly US suicide fatalities. JAMA Netw Open. 2020;3(12):e2030932. doi:10.1001/jamanetworkopen.2020.30932.\u003c/li\u003e\n\u003cli\u003eGunn JF, Lester D. Using Google searches on the internet to monitor suicidal behavior. J Affect Disord. 2013;148(2\u0026ndash;3):411\u0026ndash;2. doi:10.1016/j.jad.2012.11.004.\u003c/li\u003e\n\u003cli\u003eSueki H. Does the volume of internet searches using suicide-related search terms influence the suicide death rate? Data from 2004 to 2009 in Japan. Psychiatry Clin Neurosci. 2011;65(4):392\u0026ndash;4. doi:10.1111/j.1440-1819.2011.02216.x.\u003c/li\u003e\n\u003cli\u003eTaira K, Hosokawa R, Itatani T, Fujita S. Predicting the number of suicides in Japan using internet search queries: vector autoregression time series model. JMIR Public Health Surveill. 2021;7(12):e34016. doi:10.2196/34016.\u003c/li\u003e\n\u003cli\u003eMichaels MS, Chu C, Silva C, Schulman BE, Joiner T. Considerations regarding online methods for suicide-related research and suicide risk assessment. Suicide Life Threat Behav. 2015;45(1):10\u0026ndash;7. doi:10.1111/sltb.12105.\u003c/li\u003e\n\u003cli\u003eNHK (Japan Broadcasting Corporation). Thinking about mental health. 2023. https://heart-net.nhk.or.jp/mukiau/. Accessed 1 April 2025.\u003c/li\u003e\n\u003cli\u003ePrime Minister\u0026rsquo;s Office of Japan. Novel Coronavirus Expert Meeting (the 1st meeting). 2020. https://www.kantei.go.jp/jp/98_abe/actions/202002/16corona_sen.html. Accessed 1 April 2025.\u003c/li\u003e\n\u003cli\u003ePrime Minister\u0026rsquo;s Office of Japan. Novel Coronavirus Response Headquarters (the 101st meeting). 2023. https://www.kantei.go.jp/jp/101_kishida/actions/202301/27corona.html. Accessed 1 April 2025.\u003c/li\u003e\n\u003cli\u003eWood SN. Generalized additive models: an introduction with R. 2nd ed. Boca Raton: Chapman and Hall/CRC; 2017. doi:10.1201/9781315370279.\u003c/li\u003e\n\u003cli\u003eDutta R, Gkotsis G, Velupillai S, Bakolis I, Stewart R. Temporal and diurnal variation in social media posts to a suicide support forum. BMC Psychiatry. 2021;21:259. doi:10.1186/s12888-021-03268-1.\u003c/li\u003e\n\u003cli\u003eLi H, Sun J, Zhang Q, Wei D, Li W, Jackson T, Hitchman G, Qiu J. Neuroanatomical differences between men and women in help-seeking coping strategy. Sci Rep. 2014;4:5700. doi:10.1038/srep05700.\u003c/li\u003e\n\u003cli\u003eTamres L, Janicki D, Helgeson V. Sex differences in coping behavior: a meta-analytic review and an examination of relative coping. Pers Soc Psychol Rev. 2002;6:2\u0026ndash;30. doi:10.1207/S15327957PSPR0601_1.\u003c/li\u003e\n\u003cli\u003eNaslund JA, Aschbrenner KA, Marsch LA, Bartels SJ. The future of mental health care: peer-to-peer support and social media. Epidemiol Psychiatr Sci. 2016;25(2):113\u0026ndash;22. doi:10.1017/S2045796015001067.\u003c/li\u003e\n\u003cli\u003ePretorius C, Chambers D, Coyle D. Young people\u0026rsquo;s online help-seeking and mental health difficulties: systematic narrative review. J Med Internet Res. 2019;21(11):e13873. doi:10.2196/13873.\u003c/li\u003e\n\u003cli\u003eKhan A, Rehman A. Impact of difficulties faced by adolescents in making career decision on their mental health. J Educ Vocat Res. 2018;9(2):1\u0026ndash;8. doi:10.22610/jevr.v9i2(V).2789.\u003c/li\u003e\n\u003cli\u003eArai T, Tsubaki H, Wakano A, Shimizu Y. Association between school-related Google Trends search volume and suicides among children and adolescents in Japan during 2016\u0026ndash;2020: retrospective observational study with a time-series analysis. J Med Internet Res. 2024;26:e51710. doi:10.2196/51710.\u003c/li\u003e\n\u003cli\u003eMatsubayashi T, Ueda M, Yoshikawa K. School and seasonality in youth suicide: evidence from Japan. J Epidemiol Community Health. 2016;70:1122\u0026ndash;7. doi:10.1136/jech-2016-207583.\u003c/li\u003e\n\u003cli\u003eGabennesch H. When promises fail: a theory of temporal fluctuations in suicide. Soc Forces. 1988;67(1):129\u0026ndash;45. doi:10.2307/2579103.\u003c/li\u003e\n\u003cli\u003eShiraishi N, Sakata M, Toyomoto R, Yoshida K, Luo Y, Nakagami Y, et al. Dynamics of depressive states among university students in Japan during the COVID-19 pandemic: an interrupted time series analysis. Ann Gen Psychiatry. 2023;22:38. doi:10.1186/s12991-023-00468-9. \u003c/li\u003e\n\u003cli\u003eFujii H, Fukuda S, Narumi D, Ihara T, Watanabe Y. Fatigue and sleep under large summer temperature differences. Environ Res. 2015;138:17\u0026ndash;21. doi:10.1016/j.envres.2015.02.006.\u003c/li\u003e\n\u003cli\u003eZunhammer M, Eichhammer P, Busch V. Sleep quality during exam stress: the role of alcohol, caffeine, and nicotine. PLoS One. 2014;9(10):e109490. doi:10.1371/journal.pone.0109490.\u003c/li\u003e\n\u003cli\u003eDai H, Milkman K, Riis J. The fresh start effect: temporal landmarks motivate aspirational behavior. Manag Sci. 2014;60(10):2563\u0026ndash;82. doi:10.1287/mnsc.2014.1901.\u003c/li\u003e\n\u003cli\u003eP\u0026aacute;ez D, Bilbao M, Bobowik M, Campos M, Basabe N. Merry Christmas and Happy New Year! The impact of Christmas rituals on subjective well-being and family\u0026rsquo;s emotional climate. Rev Psicol Soc. 2011;26:373\u0026ndash;86. doi:10.1174/021347411797361347.\u003c/li\u003e\n\u003cli\u003eMaldonado G, Kraus J. Variation in suicide occurrence by time of day, day of the week, month, and lunar phase. Suicide Life Threat Behav. 1991;21(2):174\u0026ndash;87. doi:10.1111/j.1943-278X.1991.tb00464.x.\u003c/li\u003e\n\u003cli\u003eKim E, Cho S, Na K, Jung H, Lee K, Cho S, Han D. Blue Monday is real for suicide: a case\u0026ndash;control study of 188,601 suicides. Suicide Life Threat Behav. 2019;49:393\u0026ndash;400. doi:10.1111/sltb.12429.\u003c/li\u003e\n\u003cli\u003eHawkley LC, Preacher KJ, Cacioppo JT. Loneliness impairs daytime functioning but not sleep duration. Health Psychol. 2010;29(2):124\u0026ndash;9. doi:10.1037/a0018646.\u003c/li\u003e\n\u003cli\u003eFranz PJ, Nook EC, Mair P, Nock MK. Using topic modeling to detect and describe self-injurious and related content on a large-scale digital platform. Suicide Life Threat Behav. 2020;50(1):5\u0026ndash;18. doi:10.1111/sltb.12569.\u003c/li\u003e\n\u003cli\u003eLiu J. Need to establish a new adolescent suicide prevention programme in South Korea. Gen Psychiatry. 2020;33(4):e100200. doi:10.1136/gpsych-2020-100200.\u003c/li\u003e\n\u003cli\u003eM\u0026ouml;rch CM, C\u0026ocirc;t\u0026eacute; LP, Corth\u0026eacute;sy-Blondin L, Plourde-L\u0026eacute;veill\u0026eacute; L, Dargis L, Mishara BL. The Darknet and suicide. J Affect Disord. 2018;241:127\u0026ndash;32. doi:10.1016/j.jad.2018.08.028.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Table","content":"\u003cp\u003eTable 1 is available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-psychiatry","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bpsy","sideBox":"Learn more about [BMC Psychiatry](http://bmcpsychiatry.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bpsy/default.aspx","title":"BMC Psychiatry","twitterHandle":"@BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Japan, online forum, temporal patterns, diurnal variation, mental health, suicide, suicidal ideation, generalized additive model","lastPublishedDoi":"10.21203/rs.3.rs-7442638/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7442638/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground\u003c/strong\u003e: Suicide prevention can be significantly enhanced by time-sensitive surveillance using digital data sources like online forums. To inform more effective suicide prevention strategies, this study analyzes the temporal patterns of posts on a Japanese mental health forum—the NHK “Facing Suicide” forum—that may aid in early risk detection.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e: We analyzed 63,046 posts from Japan’s national broadcaster (NHK) forum (Jan 1, 2008–Mar 31, 2025), stratified by sex and age (≤ 19, 20s, 30s, ≥ 40). Generalized additive models were used to model hourly, weekly, and monthly variations, with time included as a spline term. Results are presented as incidence rate ratios (IRRs) with 95% confidence intervals (CIs).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e: Females contributed 75.5% of posts, and the 20–29 age group was the most active (32.4%). Posting activity consistently peaked around 23:00 across all subgroups. A marked increase was observed among adolescents (≤ 19) in August (males: IRR = 1.30, 95% CI [1.05–1.61]; females: IRR = 1.55, 95% CI [1.43–1.69]), while adults showed decreases in January–February. Weekly patterns varied by subgroup; for instance, males aged 20–29 posted more on Mondays and Tuesdays (IRR = 1.17, 95% CI [1.05–1.30]).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions\u003c/strong\u003e: Online forum activity displays predictable temporal cycles with demographic-specific patterns. These findings provide an essential baseline for real-time systems to detect deviations that may signal elevated suicide risk. The August peak in adolescent posts aligns with back-to-school distress, and the late-night peaks underscore the need for 24-hour support services. These insights can guide the development of targeted, time-sensitive suicide prevention strategies.\u003c/p\u003e","manuscriptTitle":"Temporal Analysis of a Japanese Online Forum for Suicide Risk Monitoring","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-01 12:50:53","doi":"10.21203/rs.3.rs-7442638/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-10-01T07:32:11+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-09-30T01:42:23+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"325811947975987209929037105104982279936","date":"2025-09-22T00:22:06+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"316041145053385540503871357157168609818","date":"2025-09-21T22:23:24+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-09-20T16:38:48+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"16340381354431932147750120729734129598","date":"2025-09-19T13:10:57+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-19T13:09:20+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-19T13:04:54+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-09-11T08:15:32+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-09-11T08:11:00+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Psychiatry","date":"2025-09-11T08:05:00+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-psychiatry","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bpsy","sideBox":"Learn more about [BMC Psychiatry](http://bmcpsychiatry.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bpsy/default.aspx","title":"BMC Psychiatry","twitterHandle":"@BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"38f78051-8e8a-46f4-be8c-893ac5eb1913","owner":[],"postedDate":"October 1st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-11-24T16:11:14+00:00","versionOfRecord":{"articleIdentity":"rs-7442638","link":"https://doi.org/10.1186/s12888-025-07539-z","journal":{"identity":"bmc-psychiatry","isVorOnly":false,"title":"BMC Psychiatry"},"publishedOn":"2025-11-20 15:58:10","publishedOnDateReadable":"November 20th, 2025"},"versionCreatedAt":"2025-10-01 12:50:53","video":"","vorDoi":"10.1186/s12888-025-07539-z","vorDoiUrl":"https://doi.org/10.1186/s12888-025-07539-z","workflowStages":[]},"version":"v1","identity":"rs-7442638","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7442638","identity":"rs-7442638","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}