Development of a Prediction Model for Predicting 10‐year Incidence of Type 2 Diabetes in Japanese People; Panasonic Cohort Study 7

ABSTRACT Aims In Japan, several prediction models and scoring systems for type 2 diabetes have been reported; however, none have high utility. We developed a new clinical prediction model for the onset of type 2 diabetes. Materials and Methods The development dataset was obtained from 72,124 Japanese employees who participated in a health check‐up programme conducted by Panasonic Corporation (Osaka, Japan), were aged 40 years or older, were diabetes‐free at baseline, and followed‐up for up to 10 years. The external validation dataset was obtained from 12,885 participants of the NAGALA (Gifu City, Gifu Prefecture Longitudinal Analysis) cohort. A prediction model was developed to predict the 10‐year risk of developing diabetes using information from the health checkup programme. The developed model was internally validated, and externally validated using the NAGALA cohort. Results Using information on age, sex, body mass index, systolic blood pressure, log‐triglycerides, high‐density lipoprotein, log‐alanine aminotransferase, fasting plasma glucose, weight gain, and smoking status obtained from a health checkup programme, we developed a novel, highly sensitive, and specific model for predicting the 10‐year risk of developing diabetes. The prediction model showed excellent performance, with an optimism‐corrected c‐index of 0.877 and a c‐index of 0.882 in the external validation cohort. Conclusion We developed a noninvasive diabetes risk‐prediction model for the Japanese population and confirmed its utility for identifying individuals at high risk of type 2 diabetes over time.