Last May, I wrote an article discussing the safety of e-scooters, as Seattle was beginning to discuss a scooter-share program modeled after its existing bike-share program, and as other cities aggressively rolled out scooter share. My conclusion was that there simply was not enough data to draw a definitive answer on e-scooter safety, though the early results presented reasons to be concerned.
Now, as Seattle is well on its way to rolling out scooter-share later this spring, a new report on scooter injuries nationwide in 2018 brings us a bigger data set — and fresh reasons to be concerned.
The report, published this week in the highly respected Journal of the American Medical Association (JAMA), pulls data from the National Electronic Injury Surveillance System (NEISS), which “provides national estimates of injuries that present to emergency departments across the United States.” The research team queried for injuries related to e-scooters from 2014-2018, then crunched the numbers.
They found a dramatic increase in e-scooter related injuries over those five years: from 4,582 in 2014 to 8,016 in 2017, with a big jump up to 14,651 in 2018. That alone is not entirely surprising given the rollout of scooter-share programs across the country, and the raw numbers alone don’t tell us much about the safety of e-scooters. We need to put those numbers in the context of the overall amount of scooter-miles driven, and compare them to other modes of transportation.
According to a NACTO study, in 2018 there were 38.5 million e-scooter trips taken, with an average trip length of 1.2 miles. That gives us a total of 46.2 million scooter-miles driven — a much larger sample size than just the few published studies of scooter-share pilots that we previously had to work with. With the 14,651 injuries recorded in the NEISS database, e-scooters have a rate of 3.17 injuries per 10,000 miles driven. That’s somewhat higher than the figures reported from previous studies in Austin, TX and Portland, OR, which were about 2.2 injuries per 10,000 miles.
Let’s compare that to motor vehicles. According to the NHTSA, in 2018 there were 3.24 trillion miles driven in motor vehicles, with 2,710,000 injuries. That equates to .008 injuries per 10,000 miles — almost 400 times lower than the reported injury rate for scooters.
We can also look at the severity of the injuries, starting with the most severe: fatalities. There were 36,560 motor vehicle fatalities reported in 2018, a rate of 1.13 per 100 million miles — or .522 fatalities per 46.2 million miles (the number of scooter-miles driven in 2018). In other words, if you took a random sample of 46.2 million miles of the 3.24 trillion driven in motor vehicles in 2018, there is a 52% chance that a fatality would have occurred.
There is no directly comparable fatality report for e-scooters. The NEISS data set lists one fatality in 2018 (a man drove his scooter into a tree), but news reports suggest that there were more that year, and there are additional reports for 2019 as well. But even the scant reports are enough to conclude that the fatality rate for e-scooters (2+ in 46.2 million miles) is at least as high as for motor vehicles, and possibly higher.
The details of the e-scooter injury reports explain why: 32% of the scooter injuries were to the head. Studies have shown that a very low percentage of e-scooter riders wear helmets, even if they are required by law.
So the picture is now becoming very clear: e-scooters are as dangerous as motor vehicles, if not significantly more dangerous, and are far less regulated.
Last month SDOT published an initial SEPA analysis of its proposed scooter-share program, which gives us a preview of what we might see on Seatte streets in the coming months. In the report, it said that it plans to issue permits for up to 20,000 scooters. That’s a high bar; bike-share has never reached nearly that level of deployment, though scooters are cheaper to buy and maintain (one of the reasons that Jump and Lime are pushing to move from bikes to scooters quickly). So let’s pare that back and conservatively assume, at least for the first few years, 10,000 is a more reasonable expectation for the number of scooters deployed in Seattle. In Portland’s 2018 pilot, it saw an average of 2.9 trips per day, per scooter. At an average trip length of 1.2 miles, that would be 12.7 million scooter-miles in Seattle annually. At that level, we can expect about 4000 scooter injuries per year — 11 per day. That would include 1280 head injuries, and at least a 50% chance of a scooter-related fatality. If the scooter-share companies deploy more than 10,000 scooters (and the trips-per-bike ratio stays the same), then the injury statistics could be much higher.
This presents some very difficult policy questions for city officials. The SEPA report says that SDOT will ask the City Council to allow scooters in roadways, bike lanes and multimodal trails, but not on sidewalks. But the city is under-invested in bike lanes currently, and Portland’s scooter-share study showed that scooter riders move to the sidewalk when they don’t feel safe riding on the street because of the speed of motor vehicles or the absence of a bike lane. In addition, the low rate of riders wearing helmets, and the high rate of intoxication of scooter riders in collisions, suggests that altogether it would take an enormous enforcement effort by the city to make scooter-share safer in Seattle than what other cities have experienced.
San Diego is a case in point: it has chosen to ban scooters along its waterfront boardwalks because of the safety issues . In response to this and other efforts by the city to regulate scooters, just this week Lime decided to pull out of San Diego entirely.
SDOT’s SEPA report dismisses the injury and enforcement issues, claiming “The SSP is not anticipated to create a significant increase in the need for fire, police, or health care services.” Largely it bases this upon its own experience with bike-share, and with a tortured reading of the results of the Portland scooter-share pilot. SDOT doesn’t even bother to estimate the number of collisions and injuries likely to result from a scooter-share program at the size it plans to permit, let alone the impact on first-responders, healthcare facilities, and traffic enforcement personnel. The SEPA “Determination of Non-significance” has been appealed to the Hearing Examiner, by former WSDOT Director Doug MacDonald.
There is currently no reason to believe that Seattle’s experience with scooter-share will be any improvement over San Diego, Portland, Austin, or the other U.S. cities that have rolled out scooter-share and seen an epidemic of injuries. And yet SDOT is planning to launch its program this spring, with the encouragement of some City Council members. What can possibly go wrong?
Thanks for reading to the end! Articles like this one take days to research and write. If you find my reporting valuable, please consider making a contribution to support my work.
Excellent analysis. I am conducting a similar such analysis for San Diego. If possible I suggest you make a comparison between scooter injury and death rates with those for bicycles as often scooters are held to be “just the same as bicycles” but I believe the data is now showing that they clearly are not. In part, as explained by cycling accident expert John Schubert in our Facebook group (SafeWalkwaysSD), this is because of the differences in design of scooters versus bikes, the former inducing a much higher rate of head injury than the latter.
Although Lime claimed the Boardwalk ban was a contributory cause of their pull out from SD in fact it has not yet come into effect and their ridership had been dropping significantly over the previous months (https://www.10news.com/news/local-news/san-diego-news/scooter-ridership-drops-off-dramatically-in-san-diego). Factors contributing to that include the increased price per minute they charged, cooler weather, fewer tourists and possibly both a diminished novelty factor and greater awareness of injuries.
FYI A colleague and I published a paper in January 2019 debunking the false claim the scooter companies made that they were helping SD reduce carbon emissions. In fact they are increasing them.
If I can be of help please don’t hesitate to contact me.
Thanks. Last year I tried to compare to bicycles, but I was unable to track down reliable national data on annual bicycle-miles travelled and bicycle injuries and fatalities. If you have reliable data, I would love to see it.
I don’t have national data, only local. National data can be easily dismissed by local decision-makers on the grounds that “that’s interesting but we’re different here”. Nonetheless, your national estimates provide useful background.
What I have is number of scooters and bicycle accidents for the last two years (obtained via public records requests), number of permitted scooters and estimated number of bicycles in the the city of SD, and am working on the estimated ave trip length for scooters and bicycles and estimated frequency of usage for both, from which one can construct estimates of mileage. So my analysis of injury rates per 10,0000 miles is not yet complete. Like you I am having trouble getting bike data.
For Seattle you should be able to access city data on bike accidents but you don’t have any yet for scooters I assume. Bike groups should be able to provide data on bikes and numerous sources provide ave trip length for scooters. The Portland 2019/2019 pilots should help understand ave trip length for cities with rainy climates.
That would enable a retrospective comparison of injury rates for bikes and scooters. However that’s not a lot of use to you right now as you’d have to wait for scooters to be launched for the injury/accident data to be gathered.
One thing to check for the City of Seattle is whether the State of Washington confers power to cities in the state to require helmets to be worn by scooter drivers. I don’t know the regs in Washington and idk whether helmets are required at all there. However there is a case to be made to the City of Seattle to require them if it can.
My initial analysis for SD simply takes the number of accidents and adjusts for the size of the population of vehicles available, so instead of comparing the number of people injured involving bikes Vs scooters, which would of course be misleading, as you rightly point out in your article, I compared the number of people injured per 20,000 vehicles. That resulted in an injury rate for scooters seven times that for bikes.
One then needs to adjust for trip length and frequency of use.
However, I am certain that for both Seattle as San Diego the ave trip length of bikes is longer than for scooters. If so that increases the estimate.
One also needs allow for the frequency of use. Using the NACTO report of 2018 and local data I expect my estimates of frequency to show that the estimated scooter frequency of use is lower than that for bikes. That again then increases the estimate.
The conclusion of the initial analysis is that the injury rate of scooters versus bicycles per 20,000 vehicles in SD is multiples higher … AT LEAST SEVEN TIMES HIGHER … for scooters than bikes, but probably far higher than that.
This simply confirms what we are seeing in the multiple studies being made by trauma centers nation-wide.
Btw in our FB group we have a local activist from Portland, OR. If you are not connected with him you might be able to via the group.
Thanks for your thoughts on this. I worry about using the injury statistics from just one city, because there isn’t enough data to reach reliable statistical conclusions. Particularly for fatalities: overall the motor vehicle, bicycle, and pedestrian fatality rates in Seattle are very low, to the point where annual variation of even 1-2 fatalities looks like a significant change. There are also demographic issues: are we comparing potential scooter-share users’ injury rates to bike-share users, or to seasoned bicyclists who are more likely to wear proper equipment, know the safest routes, and have better skills? I’m trying to avoid making questionable inferences; I’d rather say “we don’t have enough data,” as I did in my first article, than draw a conclusion that won’t stand up to scrutiny.
It’s my experience that the locals, at least here, don’t say “we’re different here.” Most of the time they say just the opposite: they also point to national trends and statistics to make the case for a local change by assuming Seattle would follow suit.
Helmets are required here, though it is not enforced and compliance is very low.
I agree with you on the small number argument regarding fatalities. I am not looking at fatalities. My data is of injuries that were recorded by the police department for both 2018 and 2019. Single digit variations do not change the results radically.
The comparison is between records of injuries incurred by scooter drivers versus bicyclists. There is no bike share in San Diego so the stats compare shared mobility devices which comprise 99.9% of the scooter population with bicyclists. Cycling is not particularly “seasonal” here due to the mild climate. Though there will be some seasonal variation it will not be as great as for Seattle.
That cyclists may wear helmets, be more experienced or know safer routes are all explanatory factors that may explain differences in injury rates. They are not factors that preclude the comparison.
My reference to dismissal of national data was to local politicians rather than the general population.
My calculations indicate a lower injury rate for scooters than yours by factors of 10 to 15. However they also indicate a higher injury rate for scooters versus bikes in the order of 30 to 70 times (2018 / 2019). In 2019 the number of scooter injuries increased significantly here while the number of bike injuries decreased slightly.