Working hour patterns and risk of occupational accidents. An optimal matching analysis in a hospital employee cohort

Abstract

BackgroundHypothesis-free studies applying advanced statistical analysis of objective working hour patterns and occupational accidents are lacking. This study aimed to identify patterns of working hours among hospital employees and to investigate the associations between the identified patterns and the risk of an occupational accident. MethodIn this cohort study of 4419 hospital employees, we collected electronic payroll-based working hour data (i.e., timing and duration) for each participant and linked them to records of occupational accident register between 2008 and 2018. We used optimal matching to assess similarity between individual working hour patterns for a period of 7 days preceding an accident or, for employees without an accident, a random pseudo-accident date. Using cluster analysis, we categorized employees into working hour pattern clusters. Log-binomial regression was used to examine risk ratios (RR) with 95 % confidence intervals (CI) of an occupational accident between cluster memberships. Results1626 participants experienced an occupational accident which took place either at the workplace (65 %) or while commuting (35 %). Six clusters of working hour patterns were identified. Compared to the cluster with the fewest accidents, clusters with a higher proportion of accidents were characterized by late work shifts and a high proportion of quick returns (<11-hour shift interval,) and long work shifts (>12-hour shift), RR 1.31, 95 %CI 1.13–1.52 for the cluster with the most accidents. ConclusionsThis data-driven study suggests that working late and long with insufficient rest is associated with increased probability of occupational accidents. Working hour arrangements in 24/7 care of hospital merit attention to regularity and sufficient rest to support occupational safety.