Nearly Half of Fatal E-Scooter Crashes Involved Alcohol, Swedish Data Shows

Electric scooters were sold as a greener, easier way to move around a city. A new look at nearly a decade of fatal crash data from Sweden shows that the people dying on e-scooters have almost nothing in common with the people dying on regular bikes. Treating both groups with the same safety rules may be missing who is actually at risk, and why.

Researchers at Chalmers University of Technology in Gothenburg, Sweden, went through a national database of every fatal road crash in the country from 2016 to 2024 that involved a conventional bicycle, an electric bicycle, or an electric scooter. They found three groups of riders are dying in different circumstances, at different times of day, and for different reasons.

Published in the Journal of Safety Research, the study covers 204 deaths: 152 conventional bicyclists, 34 e-bike riders, and 18 e-scooter riders. Eighteen e-scooter cases is a small number, but the pattern behind it is consistent enough to be worth a close look.

Private Scooters, Not Rentals, Drive Fatal E-Scooter Deaths

A common assumption is that dangerous e-scooter riding is mostly a rental problem: tourists and young people grab a shared scooter off the sidewalk, ride recklessly, and get hurt. Regulate the rental companies, keep scooters off the streets after midnight, and the danger mostly goes away.

Sweden’s fatal-crash numbers point the other way. Of the 18 e-scooter deaths, 66.7% involved privately owned scooters rather than rentals. Among e-scooter riders who had been drinking, the figure was higher still, at 87.5%. Researches argue safety measures aimed only at shared fleets would leave private riders out of the picture entirely.

Alcohol and Late Nights: A Deadly Mix

Alcohol stands out most sharply. It was a factor in 44.4% of fatal e-scooter crashes, compared with 26.5% for e-bike riders and 12.5% for conventional bicyclists. Among riders who had been drinking, the levels were high. The authors cite earlier research showing that cyclists become seriously impaired at a blood alcohol level around 0.8 grams per liter. Every intoxicated e-scooter rider in this study was above that mark, and nearly two-thirds of all intoxicated riders were past 1.4 grams per liter, the point where that earlier research found no impaired rider could match a sober one.

Sweden, unlike its Nordic neighbors, sets no legal blood alcohol limit for cyclists or e-scooter riders. Norway, Denmark, Finland, and Iceland all limit e-scooter riders. The authors say the severity of intoxication in their data supports considering a similar rule, though they stop short of demanding one.

Timing tells its own story. More than half of fatal e-scooter crashes (55.6%) happened at night, between 8 p.m. and 6 a.m., and more than half fell on weekends. Conventional bicycle deaths clustered on weekdays and in daylight, the rhythm of commuting and errands.

E-scooter crashes were also far more likely to be solo. Two-thirds involved no other vehicle at all: the rider fell, swerved, or lost control with no car or truck nearby. Part of the reason may be how the machines are built. Small wheels, a short frame, and a high center of gravity can make e-scooters less stable, especially at speed or when a rider’s coordination is off.

Not One Rider Who Died Wore a Helmet

Head injuries were the leading cause of death in every group, at well over half of cases across bicyclists, e-bike riders, and e-scooter riders alike. Yet helmet use was low, and among e-scooter riders who died, it was zero.

Not one of the 18 e-scooter riders who died was wearing a helmet. Roughly one in four conventional bicyclists and e-bike riders had one on when they were killed, a rate already below the 38 to 47% helmet use reported for Swedish riders overall. That gap hints that riders who skip the helmet may face a higher chance of a crash turning deadly.

Helmets are no cure-all. Some riders in the study wore one and still died of head injuries, a limit of protective gear in the worst crashes. Even so, the complete absence of helmets among e-scooter deaths is an obvious target for public health messaging. Sweden requires helmets only for riders under 15, and the authors say the data backs encouraging adults to wear them too.

The Age Gap Behind Fatal E-Scooter Crashes

Age may be the sharpest divide of all. Conventional bicyclists who died had a median age of 71, and e-bike riders 71.5. E-scooter riders who died had a median age of 47.5, more than two decades younger. That’s still older than the typical e-scooter user, around 30, a sign that middle-aged riders are overrepresented among the deaths.

Men made up the large majority of deaths in every group, close to 82% of bicyclist and e-bike deaths and 88.9% of e-scooter deaths. But the two profiles diverge everywhere else. Most fatal bicycle crashes involved a collision with a car or truck during a weekday ride, dangers tied to the vulnerabilities of older age and time spent in mixed traffic.

Those two profiles are the central finding of the research. A 71-year-old biking to the market on a Tuesday morning and a middle-aged man riding his own scooter home from a bar on a Saturday night, no helmet on his head, face completely different dangers. Researchers argue that rules built around one of them will not protect the other, and that Sweden’s crash database is detailed enough to build safety measures around who is actually getting hurt, rather than a single blanket approach applied to every rider.

Disclaimer: This article summarizes findings from a peer-reviewed study for a general audience and is provided for informational purposes only. It is not safety, medical, or legal advice. The research is based on 18 fatal e-scooter cases in Sweden and lacks exposure data (how many total miles each type of rider travels), so it describes patterns among deaths rather than calculating true crash rates. Intoxication data were missing for roughly 20 to 25% of cases. Statements about impairment at specific blood alcohol levels come from earlier research (Hartung et al., 2015) cited in the study, not from the Swedish data itself. Figures and interpretations reflect the original paper by Pai, Fredriksson, and Dozza.

---