One of the biggest arguments that some legislators and anti-cycling advocates use to justify a push for regulating all electric-assisted mobility devices, is the “e-bike injury” argument.
It can be stated in various ways, but basically boils down to a statement similar to:
“Injuries and deaths on e-bikes are at an all-time high and only getting worse… more people die or are seriously injured on e-bikes than… (fill in the blank).”
But this argument both misstates the issue and relies on irrevocably flawed data.
The Problem: Misclassification & Bad Data
While I am a cyclist and a cycling advocate, by education and profession I’m a medical provider.
Because of this, I know how to read research studies. As such, let me be clear:
The data being used in the studies that make the “high injury/death rates on e-bicycles” claim, are all flawed.
I will gladly go on record directly challenging the conclusions of the studies I’ve listed below, because I will show that none of them have relied on accurate data.
But First, Definitions
In previous posts (here and here) as well as on my YouTube channel and on social media, I have exhaustively and conclusively defined the differences between an e-moto and a true e-bike. If you’re not sure of the differences or don’t know the Federal laws that define those differences, I’d encourage you to review that information first.
Device Classifications
E-bikes are Federally classified as consumer products, and e-motos are Federally classified as motor vehicles. This means that the classification and reporting of injuries is supposed to be entirely different.
Obviously, this requires that electric motorcycles are listed in states’ vehicle codes as the motor vehicles that they are, in accordance with existing Federal law. The Federal government in the U.S. does not write state vehicle codes nor do they administer driver’s licenses. States are responsible for that, which is why each state has its own driver’s license registry, and its own set of vehicle codes.
Consumer Product Injury Data
Injuries and deaths involving consumer products such as electric bicycles, fall under the system known as the National Electronic Injury Surveillance System (NEISS), managed by the Consumer Product Safety Commission (CPSC).
It’s a comprehensive system that uses codes corresponding to consumer products upon or by which a person can be injured. The system is used not only to describe what are called mechanisms of injury (MOI, an acronym that I’ll use throughout this article), but can also be used by the CPSC to determine if a consumer product needs to be recalled due to the number and/or types of injuries that occur.
In NEISS, “injury reporters” (police, EMS, and medical professionals) have the option to add narrative to the codes; that is, a self-customized further explanation of the code being used with respect to the specific injury. That’s not to say that the narrative option is always used, or is used accurately; but it’s available.
To be honest, the NEISS has suffered for years from being rather vague with some of the codes. For example, in earlier revisions of the system, electric bicycles were effectively lumped into 5035 (Minibikes, powered), or similar codes 5036 (off-road/dirt bikes), 5045 (powered scooters and skateboards), or 3215 (mopeds/power-assisted cycles). The 2024 revision corrected this by reassigning code 5045 to “E-bike (electric pedal bicycle)”, moving the code’s original designation to be now assigned to the “catch-all” code 5046 (power-assisted cycles, not elsewhere classified).
Motor Vehicle Injury Data
Injuries and deaths involving motor vehicles are reported in two ways.
Reports are tracked through the Fatality Analysis Reporting System (FARS) and the Crash Report Sampling System (CRSS), both administered by the National Highway Traffic Safety Administration (NHTSA). These two systems allow the government and researchers to track various metrics pertaining to vehicle accidents involving injuries and deaths, for a variety of purposes – not the least of which is contributing to the datasets that help determine if a vehicle or vehicle component requires recall.
Medical professionals have our own system of injury and death documentation, for a wide variety of injury types and mechanisms of injury. We document using the standard International Classification of Diseases (ICD) code system, now in its tenth revision. In our documentation, injuries or deaths related to motor vehicles as a mechanism of injury are in the “V” category in the currently-used ICD-10 system.
Identifying the Problem With Study Data
As clearly as can be stated, the problem for researchers and others trying to answer the question of injury or death rates on e-bicycles (e-bikes) is this:
The data being evaluated are completely inaccurate and thus invalid for what the researchers are trying to accomplish.
When a study claims that a high percentage of injuries are occurring on e-bikes, they are not making such claims from data that can be proved to be solely e-bike related. Rather, they’re basing their conclusions on data that mix various types of devices, including e-bikes, e-motos, and other power-assisted devices, but are not referencing data from all of the reporting systems that track specific devices.
This is what makes all of the current study data invalid, and is exactly why legislators and others concerned about the subject of “e-bike injuries” MUST NOT rely on such studies as a basis for decision-making.
Compounding The Problem
While the different reporting systems noted earlier are straightforward enough in their own spheres of influence, there are several issues that compound the problem of trying to gather accurate injury data in practice.
Confusing Motor Vehicles with Consumer Products
The largest issue arises when electric motorcycles (motor vehicles) are confused in any number of ways with electric bicycles (consumer products). I’ll explain how this happens later.
The Federal legal definitions are clear: e-bicycles and e-motorcycles are two entirely different types of devices. Again, I’ve explained the differences and the laws behind them in the previous article, Defining the Differences: E-Bikes and E-Motos.
Even a cursory look on social media, let alone news outlets and legislative communications, make it all too clear that the confusion between the two types of devices is at near-pandemic level. Part of this is due to our American society that, just like referring to any brand of facial tissue as “Kleenex”, also refers to any two-wheeled vehicle as a “bike”, despite the fact that there are a variety of devices out there that don’t technically fit the definition of “bicycle” aka “bike.”
It’s bad enough when the general public make this error; it’s devastating to injury statistics when police, EMS, or medical professionals commit the same error.
Using The Wrong Reporting System
When an injury or death occurs on an electric motorcycle, it is supposed to be reported under the NHTSA-administered tracking systems. But when the mechanism of injury is misidentified as a e-bike (consumer product) and not an e-moto (motor vehicle), injury reporters erroneously use the NEISS to code the injury, a system that was not designed for such a purpose.
E-motos are not classified in NEISS, nor should they, because they’re not consumer products. But even when properly identified as e-motos and not e-bikes, many injury reporters will often make the huge error of not classifying in CRSS/FARS, and instead classifying in NEISS using the closest code that they believe might apply.
Even with the 2024 code revisions, injury reporters can easily get confused about the actual MOI, simply because it’s very unlikely that they will be told truthfully if the MOI was an e-bike or an e-moto (I’ll give a good example below).
Muddying The Waters
This “problem compounding” only muddies the statistical waters when trying to determine how many injuries or deaths actually occur on e-bikes or e-motos with the highest degree of accuracy, .
With the recent re-assignment of NEISS code 5045 to be electric bicycles and only electric bicycles, it would seem as if statistical research can finally accurately determine how many true e-bike injuries or deaths occur.
But remember that this change only occurred in 2024, meaning that any data prior to that date are nonspecific to e-bikes, which makes nearly every study done in recent years on e-bike injuries completely invalid.
The Studies
When the subject of “e-bike injuries” arises, four studies are frequently cited. They are:
- Fernandez AM, Li KD, Patel HV, Allen IE, Ghaffar U, Hakam N, Breyer BN. Electric Bicycle Injuries and Hospitalizations. JAMA Surg. 2024 May 1;159(5):586-588
- Burford KG, Itzkowitz NG, Rundle AG, DiMaggio C, Mooney SJ. The Burden of Injuries Associated With E-Bikes, Powered Scooters, Hoverboards, and Bicycles in the United States: 2019‒2022. Am J Public Health. 2024 Dec;114(12):1365-1374.
- DiMaggio CJ, Bukur M, Wall SP, Frangos SG, Wen AY. Injuries associated with electric-powered bikes and scooters: analysis of US consumer product data. Inj Prev. 2020 Dec;26(6):524-528.
- Younes H. Comparing injuries from e-scooters, e-bikes, and bicycles in the United States. J Cycling Micromob Res. 2025 Jun;100061:2950-1059.
The Problems With Current Research
These and other similar research studies, while well-intended, all falter with the same consistent problem: they all use inaccurate, outdated, and vague data as a basis for their findings.
Outdated Data, Old Codes
When researchers rely on the NEISS codes compiled prior to the code changes in 2024, they are dooming their research to total inaccuracy. Prior to the code change, there was no way to “split” dataset information between true e-bikes, e-motos, or even dirt bikes, minibikes, and scooters, so the data compiled for research can’t determine accurately enough if the numbers truly represented e-bikes, or something else.
Reporting Inaccuracies
NEISS and FARS/CRSS injury data come primarily from those responsible for reporting injuries; namely, law enforcement and medical professionals.
But because representing e-bikes in NEISS codes had been so confusing for so long, and because many e-moto injuries have been regularly mis-coded using the NEISS system instead of the FARS or CRSS systems, it’s literally impossible to accurately determine just by the datasets alone, if a reported “e-bike” injury was truly on an e-bike, rather than on a motor vehicle.
Study Method Errors
Researchers attempting to use the pre-2024 data as a basis, run into these massive walls of error. To try to get around them, researchers have often tried to “grab everything”: manually searching narratives for many different spellings of the same words, like e-bike, e bike, ebike, etc., in an effort to grab whatever may be available. Sometimes this works in research; but when attempting to drill down on datasets that are inaccurate enough to begin with, it’s unfortunately a hopeless effort.
The most egregiously error-filled study was Fernandez, who (as shown above) cited using NEISS codes 5035 (Minibikes, powered), and 3215 (mopeds or power-assisted cycles). This massive error is compounded by the fact that Fernandez specified that the study was excluding injuries from motor bikes and minibikes, while at the same time incorporating the minibikes NEISS code (5035) among the datasets used in the study.
Burford and DiMaggio appear to be committing the same errors as Fernandez, but by relying solely on emergency department reports of injury through NEISS codes alone, and not cross-referencing reports through other methods (police reports, chart records, etc.). This type of cross-referencing could be time-consuming, but it would result in potentially clearer data. Burford even makes note of the years of its research datasets as 2019-2022, far earlier than the current NEISS system revision.
Younes appears to do the same thing as the others. But I will also strongly criticize her publicly for being openly narcissistic in the study cited above. She consistently references her own prior work, instead of engaging in fresh, original research. This is a sign of a bad – and biased – researcher. We don’t need any more of them.
After berating these researchers for the many unrecoverable errors in their studies, I will grant them all one “pass”: since they’re all relying solely on NEISS codes as reported in injury data, it’s no wonder that they’ve come up with inaccurate percentages of e-bike injuries, because of how the injuries are coded in the first place.
Confusing E-Bikes and E-Motos: How This Happens
Consider the unpleasant but all too common fact that when a death or injury incident occurs, first-responders (police, EMS, ED staff) face the task of not only describing the injury or injuries and the person(s) involved, they also must identify and document the mechanism(s) of the injury.
But when these first responders don’t know the difference between and e-bike and and e-moto, they’ve fallen at the very first hurdle: misidentifying the mechanism of injury.
Misidentifying The Mechanism
As I’ve said countless times, e-motos are not e-bikes, and everyone needs to understand that. And when it comes to accurately reporting injuries and/or deaths on e-motos or e-bikes, understanding the differences isn’t just helpful, it’s essential.
I contend that if you show 500,000 first responders two pictures – one of a true e-bike and one of an e-moto – I’d be willing to wager that at least 60% of them (if not more) will misidentify the e-moto picture as an “e-bike.”
Consider this fact also: when an incident involving an e-bike or e-moto occurs, it is not guaranteed that police or EMS would need to be involved. Some injuries may not be serious enough to require police or EMS intervention.
Therefore, while police officers would be more likely than EMS professionals to correctly identify an e-moto as an e-moto and not an e-bike, the problem is this: if they don’t need to respond to an incident, who’s going to tell the emergency department that the MOI was an e-moto?
These misidentifications make it less likely that the injury or death report would be routed to the appropriate reporting system, either NEISS for e-bikes or CRSS/FARS for e-motos.
Misrepresenting The Mechanism
When it comes to injury data compiled and used by the studies such as are listed above, the reporting from emergency departments is by far the worst offender – but not deliberately.
Let me use an example here to illustrate exactly how the MOI can easily be misrepresented, leading to the misidentification of the MOI, and subsequently the wholesale mis-coding of the data.
“I Was Just Riding My E-Bike”
Joe Teenager was riding his Surron e-moto on the street, popping wheelies with his friends, riding four abreast and impeding traffic. Suddenly, an irate driver gets a good case of “road rage”, and speeds past this “rat pack” of kids, passing them on a two-lane public road and swerving back into his proper lane at the last minute before an oncoming vehicle would have hit him. While finishing his wheelie, Joe gets spooked by the car cutting him off, and he clips the back-end of the passing car, losing control and crashing.
The other kids in his pack all scatter because they realize what they were doing, and don’t want to be accosted by the police when they arrive. The police are called, and EMS also attend Joe, who is shaken, but conscious. He’s sustained brush-burn injuries to the arm and leg that impacted the pavement, and hit his head because he wasn’t wearing a helmet like his parents gently asked him to. He’s transported by EMS to the local hospital.
The responding police officer’s knowledge of e-motos and e-bikes is marginal at best, and when he writes his report of the incident, he decides that the MOI is an e-bike; but, as his report gets coded, NEISS code 5036 is used because the coder reading the report considered it an electric dirt bike. The EMS staff calls it an “e-bike”, and this gets reported to the receiving emergency department.
The hospital staff attend Joe, and when the ED provider arrives to assess him, Joe is asked how the injury came about.
Joe says “I was just riding my e-bike.”
Joe chooses not to tell the ED provider that he was engaging in reckless riding behavior on a public street with other e-motos, and how the crash actually came about. He withholds this information because he didn’t want to implicate his friends (who abandoned him at the scene) and wants to make it appear that the driver of the car who sped past them was 100% at fault.
Because it’s a busy day in the hospital, the ED provider probes no further, and quickly documents “e-bike” as the MOI, which is then coded using the NEISS code for e-bike (5045) because of what Joe said; even though what Joe was actually riding wasn’t an e-bike at all, it was an e-moto.
The injury data from two different sources (police and ED), using two different NEISS codes, get compiled into the NEISS system as both “e-bike” and “off-road bike”, and thus become part of the millions of erroneous and inaccurate data points that get reviewed by researchers looking to determine the percentage of injuries that occur on e-bikes – all because those who were in a perfect position to clarify exactly what the precise mechanism of injury was, didn’t.
So…
THERE IS THE SOURCE OF THE PROBLEM, RIGHT THERE.
The MOI is misunderstood, misidentified, mis-coded, and no one ever clarifies anything. It’s no wonder then, that all of the studies are wrong.
The Seemingly Impossible Task
If you think that my example was just some mythical story, think again.
It may sound like I’m berating my medical colleagues who work in emergency medicine – and to some degree, I am.
But even though I work in family medicine, I understand the pressures that emergency department providers are under. I have a personal friend who has been an ED physician for over 25 years, and he has reported to me how often things like the example above actually do happen every day in emergency departments across the country.
It isn’t because ED providers are incompetent or uncaring. Rather, it comes down to several factors that make their job of accurately defining the MOI a seemingly impossible task.
1. The patient will frequently misidentify the mechanism of injury as a “bike”.
This is mostly for the reason I mentioned earlier about our American language being “lazy” and calling anything with 2 wheels a “bike.”
For decades, the term “bike” has been so constantly misapplied to devices that don’t qualify for the term. This means that any attempt by injury reporters to get the truth out of an injured person (particularly a youth) is nearly impossible without more external information (like from police or actual photographs) and a knowledge of the differences between e-bikes and e-motos.
It also doesn’t help that many e-moto manufacturers are deliberately and deceptively marketing their machines as “e-bikes” when they know that their machines don’t legally qualify for that name.
2. Emergency departments are often understaffed and overworked.
My friend, who works in a very busy ED in a major metropolitan area, has confirmed how frequently emergency departments are short-staffed or under stress due to the volume of cases arriving during any given shift. This is not to say that ED providers won’t try their best to be as accurate as possible in their documentation, however.
But an overworked provider may be responsible for 5 or more urgent cases at one time. They may simply expect the intake or nursing staff to gather appropriate MOI documentation (even though the nursing staff is also frequently understaffed and very busy). Or, they may find it easier and faster to simply take the patient’s “word for it” regarding the MOI.
So the end result is the same: when a patient says “e-bike”, and the staff do not ask more questions like what brand of e-bike, what did it look like, or other questions that would clearly identify the exact type of device involved in an incident, the true MOI gets misrepresented and improperly documented.
ED staff need to ask probing questions, but when they feel that they don’t have the time, or the patient will refuse to answer further questions, inaccurate reporting of injury is inevitable.
3. Deaths involving e-motos rely on police for accurate coding.
Since no ED provider can possibly ask a dead person what type of device he/she was involved with, coding deaths on these devices rely on the police to use the correct coding systems for the MOI. This brings up the same “misunderstanding” conundrum that befalls many in the general public.
We would typically expect law enforcement officers, who deal with incidents involving different types of mobility devices, including cars, trucks, motorcycles, etc., to know the difference between an e-moto and an e-bike so they can use the correct reporting system for the MOI.
But that’s not always the case. Again, it comes down to education.
The Nexus Of Error
If first responders don’t have a thorough knowledge of the different types of devices out there, and if they don’t ask the right questions and get the most accurate information, then injuries on e-motos will certainly continue to get misclassified as occurring on an “e-bike” using NEISS instead of where they should be reported.
The result of these errors is a vast negative skew of the data.
This makes it absolutely impossible for any researcher to obtain valid, quality data regarding injuries on e-bikes or e-motos, thus invalidating any conclusions they draw based on those inaccurate datasets.
These errors only serve to compound the research and statistical problems.
These errors are also why it is wrong for legislators to rely on such inaccurate research data as a nexus for regulatory decision-making.
Huge errors in determining injury data on e-bikes and e-motos will persist until the problems are resolved.
So What Do We Do?
In order to solve this problem, we need to embark on a multiple-pronged approach.
The CPSC Needs To Clean Up The Codes
At the top of this list is the Consumer Product Safety Commission.
While the CPSC has made a positive step forward by re-assigning 5045 to be “E-bike (electric pedal bicycle)” in the 2024 revision, they need to go further.
The CPSC should separate the “catch-all” codes for the other types of mobility devices that are frequently confused with e-bikes, and permanently remove codes relating to any device that the Department of Transportation and NHTSA consider a motor vehicle, including gas- or electric-powered minibikes, off-road/dirt bikes, mopeds, and scooters.
By doing this, it removes the vagueness and confusion about whether a minibike, dirt bike, or scooter should be coded in NEISS since the codes wouldn’t be able to be used going forward. The CPSC can still retain 5046 as a catch-all code for any power-assisted device that doesn’t clearly fit as any of the aforementioned products.
Obviously this would require the CPSC to embark on re-educating injury reporters that certain devices cannot be coded in NEISS. This one step could go a long way to minimize the confusion and reduce dataset corruption.
Injury Reporters Need Educated The Right Way
Clarifying the codes in NEISS is a great first step. But it’s only as good as the people who use them.
NEISS still heavily depends on police and ED staff descriptions, the “narratives”, and the reliability of the human coders to interpret vague information. NEISS assigns codes using as much detail as the data allow.
But when those data are vague or misleading, the accuracy of NEISS injury datasets falls apart completely.
So, everyone involved in first-response actions – including police, EMS, and ED staff – absolutely must document and code accurately and thoroughly.
Emergency Departments & EMS
I’ll call out my medical colleagues in EDs across the country with this: it is completely unacceptable to simply “take the word” of a patient injured on a mobility device.
Everyone – from the providers (physicians, nurse practitioners, and physician assistants) to the nurses, aides, and intake staff – need to be re-trained to ask the right questions, and not take “I was on my e-bike” for a legitimate answer – especially if the injured person isn’t an adult. When assessing injury and assigning a plan of care, one thing that we need to know the exact MOI. Bicycles and motorcycles can cause different, or different degrees of, injury. Knowing what the cause was, helps us to develop a plan of care that has the best chance for success with the least chance for complications.
Law Enforcement
Police responding to such incidents as we’re discussing here, need to report exactly what type of device was involved, and not just use “e-bike” as an answer, especially if the device involved is actually an e-moto. Police department administrators across the country need to make absolutely certain that their officers clearly understand the difference between an e-bike and an e-moto. Many of them do, and I personally applaud all of you for that. But many others still don’t, and that error needs fixed.
Journalists/News Organizations
News reporters who respond to accident scenes need to stop using the term “e-bike” in their reporting if the device was an e-moto. Journalism is supposed to be all about the truthful and impartial reporting of events. But when news reporters and organizations post a story involving an “e-bike” but the picture from the scene is obviously an e-moto, their credibility takes a massive nose-dive. Just like everyone else, journalists need to be educated on the differences between e-bikes and e-motos, and stop using terms that don’t apply to the situation.
Working Together
The mandate is clear: everyone needs to use the proper terms, use accurate coding, and stop producing research based on outdated and inaccurate data.
When we work together to accomplish these goals, we will start to see what the observational data have shown for years: the mechanisms of injury involved in the greater percentage of injuries and deaths aren’t electric bicycles, but high-powered electric motorcycles.
Only when legislators realize this, can sensible legislation be enacted to help keep people as safe as possible. Some states have already done this (like California’s SB 1271), or are in the process. Now others need to catch on.
There is a logical solution. In the next article, I’ll lay out a design for an e-bike and e-moto law that can be that can be used by every state to not only follow Federal laws regarding both types of devices, but will treat everyone fairly and keep everyone safe.
