Answers are grounded in this paper and cite their sources.
The rapid expansion of shared e-scooter services has introduced a novel public health challenge, with rising injury rates reported globally. Existing epidemiological studies rely largely on retrospective hospital or police records, which lack data on pre-crash mechanisms. While naturalistic studies have begun to fill this gap, they have thus far relied on associational measures that may be biased by confounding. This study aims to quantify the causal effects of rider behavior and environmental factors on e-scooter safety–critical events (SCEs) using a formal causal inference framework. We conducted a case-control analysis within a naturalistic riding study of a public e-scooter service in Gothenburg, Sweden (2022–2023), drawing from 6868 trips. Our analysis included 61 SCEs (cases) and 183 randomly selected control events. We used directed acyclic graphs to identify confounders and applied the parametric g-formula weighted for exposure duration to estimate adjusted risk ratios (RRs) and risk differences. After adjustment, the risk of an SCE increased substantially for single-handed riding (RR = 5.5), riding through an intersection (RR = 3.8), and inexperience (RR = 2.8). Phone use was associated with an approximately threefold increase in risk (RR = 3.1), though this estimate was not statistically significant. Both rider behavior (e.g., inexperience, single-handed riding) and the built environment (e.g., intersections) are significant causal contributors to e-scooter crash risk. These findings provide evidence for a multi-level public health strategy that focuses on novice rider education, infrastructure modifications at intersections, and technological interventions to support riders during their trips.
No takes yet. Share an insight, caveat, or question.
Pai et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: