Measuring effects of safety-reminding interventions against risk habituation

Despite advances in automation of construction tasks and related technologies, worksites are still characterized by several labor-intensive works and numerous accidents. Various studies have focused on safety management to prevent accidents by encouraging safety behaviors within the workplace. Most studies focus on accident prevention as a relatively short-term event and do not consider risk habituation caused by familiarity. Thus, quantitative investigations on the effects of risk habituation and interventions are lacking. Because the level of risk habituation is relative and varies within individuals, identifying a practical solution suitable for the workplace will be challenging. This study measures the degree of incautiousness and its impact on risk habituation based on the workers' gait stability and provides effective intervention scenarios to delay risk habituation in the experimental setting. The safety intervention scenarios against risk habituation are designed as 1) installing fixed-repetitive intervention based on obstacle location, and 2) behavior-feedback intervention according to performed incautious behaviors; the effects are compared with trends in the average dynamic time warping (DTW) values of the subjects. The results show that behavior-feedback is more efficient than fixed-repetitive alarms against risk habituation. In addition, identifying the optimal intervention level (i.e., timing and frequency) that decreases or slows incautious behaviors over time is necessary. This study empirically reveals the impacts of risk habituation and provides managerial strategies for delivering interventions against negative familiarity in practice.