Speaker: Prof Andrew Timming, RMIT
Title: Using Artificial Neural Networks to Predict Suicidal Ideation: Implications for Algorithmic HRM
Date: Thurs, 13th Oct 2022
Time: 1pm – 2:30pm
Venue: Building 24, Copland, Seminar Room 1106
Online meeting link: https://anu.zoom.us/j/88595382044?pwd=RW8xejA2SU9yVVFaT2hLWlBidUpUUT09
Meeting ID: 885 9538 2044
Prof Giles Hirst is the host of this visit
Using data from the National Survey on Drug Use and Health, this study employs artificial neural networks to predict suicidal ideation among employees using a set of HR-related inputs. Several multilayer perceptrons are designed, optimized, and subsequently validated using post-hoc logistic regression analyses. The final algorithm successfully predicts suicidal ideation at 216 percent above the baseline using the following predictors: employee gender, age, previous drug use, history of in-patient mental health treatment, history of out-patient mental health treatment, whether the employee has previously served in the armed forces, sexual orientation, whether the employee is married, and the employee’s K6 (psychosocial distress) score. Given the shifting boundaries of medical privacy in light of COVID-19, it is increasingly conceivable that HR—through occupational health—will have access to these employee data. Leveraging them can potentially identify employees at risk of suicide, enabling HR to provide life-saving support. The study has important implications for algorithmic HRM and occupational health and safety.
Andrew R. Timming is Professor of Human Resource Management and Deputy Dean Research & Innovation in the School of Management at RMIT University. He holds a PhD from Cambridge University. He is the inaugural Registered Reports Editor at Human Resource Management Journal. These days, he mostly researches mental health/ illness in the workplace and tattoos.