Analysis of Relationships Between Suicide and Time of Life and Lifestyle Behaviors in Japan Before and After the COVID-19 Pandemic and Use of Generative AI for EBPM

This study aims to explore regional trends in metropolitan suicide rates in Tokyo and the metropolitan area around Tokyo before and after the COVID-19 pandemic, using generative AI to provide evidence for policy-making. Data for one municipality in Tokyo in 2016 and 2021, extracted from the Ministry of Health, Labour and Welfare's Basic Data on Suicide in the Region, will be used. The data will include 71 variables, including age, living together, occupation, location, method, time of day, day of the week, reason for suicide and history of suicide attempts. The survey method visualizes suicide data for the years 2016 and 2021. Then, spatial correlation is detected using suicide data in the metropolitan area. In addition, data on usual health, working conditions and lifestyle behaviors such as volunteering and hobbies are integrated with the suicide data from the living time and lifestyle data. Here, the data is expanded to all municipalities in Japan, and the relationships are tested using linear regression and vector generalized linear models. Another research approach uses ChatGPT (version 4), a generative AI. Fuchu City, Tokyo, where the author lives, is the focus of this study, comparing annual changes in Fuchu City before and after the COVID-19 pandemic (2016 and 2021) and contrasting data for Tokyo as a whole and Fuchu City in 2021. Here, a comparison of Fuchu City (municipality) and Tokyo (prefecture) averages is prompted into ChatGPT using 71 variables of data, allowing it to detect differences and generate a document summarizing the results; the accuracy of the AI-generated content is assessed to determine its suitability for evidence-based policy-making.