Predictors of prolonged sickness absence after work injuries: a retrospective study using Tukey’s Fence in 66,658 Hong Kong workers

Objectives: To identify key predictors of prolonged sickness absence among workers with occupational injuries in Hong Kong, with a focus on demographic, occupational, and injury-related factors. Methods: We conducted a retrospective analysis of 66,658 occupational injury cases managed by a major rehabilitation coordination provider in Hong Kong, representing approximately 20% of all such cases in the region between 2003 and 2020. Prolonged absence outliers were defined as cases of sickness absence exceeding 129 days, based on Tukey’s Fence, a non-parametric method based on interquartile range. Multivariable logistic regression was used to examine associations between outlier status and the following predictors: age, gender, year of accident, physical demand, industry sector, skill level, injury type, and injured body region. All covariates were mutually adjusted in a single multivariable model. Results: Outliers constituted 16.0% of the total sample. Older age and female gender were significant individual-level predictors. Fractures were strongly associated with extended sickness absence, whereas soft tissue injuries were linked to shorter recovery durations. Torso injuries—particularly to the back and pelvis—carried higher risks of prolonged absence, while lower limb injuries were associated with earlier return to work. Workers in physically demanding roles, especially within the construction and transportation industries, were associated with elevated risk. Notably, semiskilled workers had higher odds of prolonged absence than unskilled workers. Conclusion: This study identifies patterns that may influence prolonged sickness absence, suggesting that intersecting demographic, occupational, and injury-related factors contribute to variation in recovery outcomes. The present findings may help inform future efforts to prioritize tailored interventions for high-risk subgroups, including older employees, women, semiskilled laborers, and those engaged in physically strenuous jobs. Future research should also incorporate psychosocial variables to better capture additional drivers of delayed recovery.