Sleep duration as a mediator between daily working hours and next-day fatigue among platform delivery riders: a multilevel time-lagged analysis in South Korea
Long working hours and insufficient sleep are established contributors to fatigue-related crash risk. However, the within-person pathway from work duration to next-day fatigue via sleep loss remains uncharacterized in platform motorcycle delivery riders. We investigated whether sleep duration mediates the within-person association between daily working hours and next-day end-of-shift fatigue among platform-based food-delivery riders in South Korea. In this prospective 14-day ecological momentary assessment study, 55 riders underwent wrist actigraphy and completed pre- and end-of-shift surveys. Piecewise multilevel time-lagged structural equation modeling with a linear spline at 12 h was used to estimate the indirect effects of daily working hours on next-day fatigue through sleep duration, adjusting for delivery intensity and rest time as workload covariates. Riders worked a mean of 10.9 h per day and slept 5.4 h per night, with shorter sleep durations on work-days than on off-days. Subjective fatigue increased across 70.8% of work shifts. Each additional working hour reduced sleep by 3.6 min for shifts under 12 h (B = − 3.59, p = 0.005) and by 11.5 min beyond 12 h (B = − 11.52, p < 0.001); segmented regression located the breakpoint at 13 h (95% CI: 10–14 h), supporting a threshold region rather than a precisely determined cutoff. Shorter sleep duration predicted higher next-day end-of-shift fatigue (B = − 0.002, p = 0.005). The indirect effect through sleep duration was significant in both segments (< 12 h: B = 0.007, 95% CI 0.000–0.014; ≥ 12 h: B = 0.022, 95% CI 0.003–0.042; difference: 0.015, bootstrap 95% CI 0.002–0.041), while the direct effect was marginal for shorter shifts and absent for longer shifts. Delivery intensity independently predicted fatigue (B = 0.23, p = 0.003). These findings suggest that sleep loss is an important pathway linking prolonged work duration to next-day fatigue, with potential implications for fatigue-related crash risk among platform delivery riders. These findings may inform policies designed to support recovery opportunities after long shifts, including platform-level measures such as fatigue alerts or rest prompts, although such measures require validation against objective fatigue and safety outcomes.
Authors
- Ari Min (ORCID: https://orcid.org/0000-0002-5151-0559)
- Sookyung Kim (ORCID: https://orcid.org/0000-0001-6364-6907)
- Chang Gi Park (ORCID: https://orcid.org/0000-0003-3366-4279)
Institutions
- Soonchunhyang University (KR)
- University of Illinois Chicago (US)
- Chung-Ang University (KR)
Publication Details
- Journal
- BMC Public Health
- Published
- 2026-11-21
- DOI
- https://doi.org/10.1186/s12889-026-30229-8
- Primary Topic
- Sleep and Work-Related Fatigue
- Type
- article
- Field-Weighted Citation Impact
- 0.00