Temporal analysis of posts on a Japanese online message board for suicide risk monitoring
Authors:
Open Access
- 20 November 2025
- journal article
- research article
- Published by Springer Nature in BMC Psychiatry
- Vol. 25 (1) , 1111
Abstract
Background: Suicide prevention can be significantly enhanced by time-sensitive surveillance using digital data sources like online message board. To inform more effective suicide prevention strategies, this study analyzes the temporal patterns of posts on a Japanese mental health message board—NHK’s “Facing Suicide” website—that may aid in early risk detection. Methods: We analyzed 63,046 posts from Japan’s national broadcaster (NHK) message board (1 Jan, 2008–31 Mar, 2025), stratified by gender and age (≤19, 20s, 30s, ≥40). Generalized additive models were used to model hourly, weekly, and monthly variations, with time of day 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 years) 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 message board activity displays predictable temporal cycles with demographic-specific patterns. These findings provide an essential baseline for real-time monitoring 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 to provide 24-hour support services and implement automated multi-layered online intervention strategies directly within message boards. These insights can guide the development of targeted, time-sensitive suicide prevention strategies.Keywords
All Related Versions
This publication has 45 references indexed in Scilit:
- Development of a Machine Learning Model Using Multiple, Heterogeneous Data Sources to Estimate Weekly US Suicide FatalitiesJAMA Network Open, 2020
- Suicide Risk Assessment Using Machine Learning and Social Networks: a Scoping ReviewJournal of Medical Systems, 2020
- Diurnal variation in suicide timing by age and gender: Evidence from Japan across 41 yearsJournal of Affective Disorders, 2019
- The Darknet and suicideJournal of Affective Disorders, 2018
- Temporal distribution of suicide mortality: A systematic reviewJournal of Affective Disorders, 2018
- Saisons, rythmes circadiens, sommeil et vulnérabilité aux conduites suicidairesL'Encéphale, 2015
- Fatigue and sleep under large summer temperature differencesEnvironmental Research, 2015
- Improving National Data Systems for Surveillance of Suicide-Related EventsAmerican Journal of Preventive Medicine, 2014
- EPA-0404 – Correlation and causality between the use of online health forums and suicidality: evidence from a german representative survey and a large-scale online panel surveyEuropean Psychiatry, 2014
- Using google searches on the internet to monitor suicidal behaviorJournal of Affective Disorders, 2013