Developing Action Plans Based on Machine Learning Analysis to Prevent Sick Leave in a Manufacturing Plant. Issue 2 (8th February 2023)

• Record Type: Journal Article
• Authors: Kurisu, Ken; Song, You Hwi; Yoshiuchi, Kazuhiro
• Is Part Of: Journal of occupational and environmental medicine. Volume 65:Issue 2(2023)
• Page Start: 140
• Page End: 145
• Publication Date: 2023-02-08
• Subjects: Brief Job Stress Questionnaire -- machine learning -- mental disorders -- occupational health -- sick leave; Medicine, Industrial -- Periodicals; Environmental health -- Periodicals; Occupational Medicine -- Periodicals; Environmental Medicine -- Periodicals; Environmental health; Medicine, Industrial; Periodicals; Electronic journals; 616.9803
• DOI: 10.1097/JOM.0000000000002700
• Languages: English
• ISSNs: 1076-2752

Abstract

Abstract : This article describes a process of action plan development to prevent sick leave using a machine learning model at a Japanese manufacturing plant. Stress response scores in the Brief Job Stress Questionnaire, workers' age, and departments were important factors, and thus, we developed plans to address each of these factors. Abstract : Objective: We aimed to develop action plans for employees' health promotion based on a machine learning model to predict sick leave at a Japanese manufacturing plant. Methods: A random forest model was developed to predict sick leave. We developed plans for workers' health promotion based on variable importance and partial dependence plots. Results: The model showed an area under the receiving operating characteristic curve of 0.882. The higher scores on the Brief Job Stress Questionnaire stress response, younger age, and certain departments were important predictors for sick leave due to mental disorders. We proposed plans to effectively use the Brief Job Stress Questionnaire and provide more support for younger workers and managers of high-risk departments. Conclusions: We described a process of action plan development using a machine learning model, which may be beneficial for occupational health practitioners.