a b s t r a c t

Introduction: There has been a spike in recent news regarding motorized scooter injuries due to the expansion of scooter sharing companies. Given the paucity of literature on this topic, the purpose of our study was to describe and quantify emergency department encounters associated with motorized scooter related injuries.

Methods: The National Electronic Injury surveillance System (NEISS) was queried for motorized scooter related injuries from 2013 to 2017. Patient demographics, diagnosis, injury location, narrative description of incident, and disposition data were collected from emergency department encounters.

Results: There were an estimated 32,400 motorized scooter injuries from 2013 to 2017. The estimated incidence did not change significantly over time with 1.9 cases per 100,000 in 2013 and 2.6 cases per 100,000 in 2017. A 77.0% increase in scooter injuries was noted for millennials from 2016 to 2017. Head injuries were the most com- mon body area injured (27.6%). Fractures or dislocations (25.9%) were the most common diagnosis. The most common site of fracture was the wrist and lower arm (35.4%). There were no deaths. Major orthopaedic injury and concussion were the strongest independent predictors of hospital admission.

Conclusions: Head injuries were the most commonly injured body part, while fractures or dislocations were the most common diagnosis. These results highlight the importance of using protective equipment while riding mo- torized scooters, and lay a foundation for future policies requiring helmet use.

(C) 2019

Introduction

First invented in 1915 by Autoped [1], the motorized kick scooter has become a controversial Mode of transportation. Scooter share compa- nies have existed since 2012, however the two most popular companies, Lime and Bird, began operations in July and September 2017, respec- tively. These scooters have been proposed as a “last mile” solution to existing public transportation systems [2]. Their services have grown enormously, with the two main companies each reporting N10 million rides since launch [3]. Multiple news articles have described a spike in injuries and even a reported death associated with motorized scooters since the popularization of scooter-share companies [4-6]. Based on these reports, riders have filed a class action lawsuit against scooter- share companies [7]. Despite substantial media attention, there is little

? This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

* Corresponding author at: 59 Executive Park South, Suite 2000, Atlanta, GA 30029, United States of America.

E-mail address: [email protected] (E.R. Wagner).

epidemiological data describing scooter related injuries in the United States.

The National Electronic Injury Surveillance System is a rep- resentative sample of Emergency Department encounters in the United States. Epidemiologic studies using this database have examined injury trends and patterns associated with other popular transportation mo- dalities, such as hoverboards [8,9], bicycles [10], all-terrain-vehicles, and dirt bikes [11]. Using this database, we aimed to describe trends in incidence, injury patterns, and Patient dispositions for motorized scooter related injuries in recent years.

Methods

Data for this study was obtained from the United States Con- sumer Product Safety Commission, National Electronic Injury Sur- veillance System (NEISS). The NEISS collects patient information on emergency department (ED) visits associated with a consumer product. Approximately 100 hospitals are sampled annually, includ- ing children’s hospitals, to create a nationally representative proba- bility sample of product related injuries occurring in the United

https://doi.org/10.1016/j.ajem.2019.03.049 0735-6757/(C) 2019

Fig. 1. Estimated incidence of motorized scooter injuries (2013-2017). *95% confidence interval bands around estimates.

Estimated Incidence of Motorized Scooter Injuries in the United States (2013-2017)

3.5

3

2.5

2

1.5

1

0.5

2013 2014

2015

Year

2016

2017

Esitmated Injuries Per 100,000 People

States [12]. In addition to patient demographics, incident date, ED diagnosis, injury location, and patient disposition, the NEISS con- tains a brief narrative describing incident scenarios. This narrative can be queried to elucidate data that would not otherwise be cap- tured with coding provided by the NEISS.

As an analysis of publicly available de-identified data, this study is exempt from our Institutional Review Board. We queried the NEISS for years 2013-2017 using product code 5042 (“Scooters/skateboards, powered”) and identified 2675 unweighted emergency department en- counters. The search term “scooter” was subsequently used to identify motorized scooter related injuries from the patient narrative. 827 un- weighted cases were identified using this criterion. Hoverboard and

Estimated Motorized Scooter Injuries for Millenials (2013-2017)

3000

2500

2000

1500

1000

500

0

2013

2014

2015

Year

2016

2017

powered skateboard injuries were excluded with the use of the word “hover”, “board”, and “skate”. A search and exclusion of patients whose narrative included terms “wheelchair,” “motorbike,” and “mo- torcycle” was performed to further eliminate any miscoded patients in the dataset. After our search and exclusion was performed, 820 un- weighted motorized scooter injuries were included in the study.

Incident date, patient demographics, diagnosis, injury location, nar- rative description of incident, and disposition data were collected from emergency department encounters. Population data for incidence estimates were acquired from the United States Census Bureau [13].

National estimates were created using weights provided by the NEISS. Statistical analyses were performed using IBM SPSS Statistics, Version 25.0 (Armonk, NY: IBM Corp). The complex samples function of SPSS was used to create 95% confidence intervals (95% C.I.). All data and analyses in this study are weighted estimates except when noted. Descriptive analyses, chi-square for categorical variables, and Mann U Whitney tests for continuous variables were employed. Patient diagno- ses and demographic factors were entered into a binomial logistic re- gression to elucidate independent predictors of inpatient admission. 95% confidence intervals were used to determine statistical significance.

Results

The incidence of motorized scooter injuries

There were an estimated 32,400 (95% C.I. 25,403-39,396; 820 un- weighted) motorized scooter injuries in the United States between 2013 and 2017. The estimated yearly incidence of motorized scooter in- juries appeared to rise over time, from 1.9 cases per 100,000 (6230 total cases) in 2013 to 2.6 cases per 100,000 (8525 total cases) in 2017, how- ever this trend did not reach statistical significance (Fig. 1). A spike in

Esitmated Injury Count

+77.0%

+6.1%

-28.4%

+52.1%

Fig. 2. Estimated motorized scooter injuries for millennials. *95% confidence interval bands around estimates.

Table 1

Demographics of patients with motorized scooter injuries.

Table 2

Patient diagnoses and disposition after presentation to emergency department with mo- torized scooter injury.

Estimated patients (95% C.I.) Percentage of total

Age	Infant/toddler (0-5) School age (6-12)	1628 (990-2264) 11,199 (8136-14,261)	5.0% 34.6%			Number of patients (95% C.I.)	Percentage of total
	Adolescent (13-22)	5186 (3164-7208)	16.0%	Diagnosis	Fracture/dislocation	8405 (6346-10,305)	25.9%
	Millennial (23-39)	5161 (2750-7573)	15.9%		Contusion	8366 (5632-11,101)	25.8%
	Adult (40-64)	6209 (4160-8259)	19.2%		Laceration/avulsion	4912 (3623-6201)	15.2%
	Elderly (65+)	3017 (1509-4525)	9.3%		Internal organ injury	3526 (2327-4725)	10.9%
Gender	Male	19,425 (14,682-24,168)	60.0%		Other/unreported	3308 (1889-4727)	10.2%
	Female	12,975 (9188-16,761)	40.0%		Sprain or strain	2367 (1288-3445)	7.3%
Race	White	17,741 (12,945-22,538)	54.8%		Concussion	824 (364-1283)	2.5%
	Black	3534 (1814-5253)	10.9%		Dental injury	253 (18-488)	0.8%
	Asian	294 (38-550)	0.9%		Burn	275 (11-539)	0.8%
	American Indian	153 (0-368)	0.5%		Hematoma	143 (0-345)	0.4%
	Unknown/other	10,678 (5125-16,231)	33.0%		Amputation Crush	14 (0-43) 10 (0-22)	b0.1% b0.1%
				Disposition	Treated & released	29,426 (23,368-35,484)	90.8%

scooter injuries was noted for millennials (aged 22 to 39) from 2016 to 2017 (+77.0%, Fig. 2). Injury trends over time, stratified by age group, are displayed in Supplemental Table 1.

Risk factors for admission

Demographic characteristics of scooter injury patients are described in Table 1. The majority of patients were white (54.8%) and male (60.0%). Scooter injuries occurred more frequently on weekend days (Fig. 3). Patient dispositions are described in Table 2. 90.8% of patients were discharged from the ED after treatment, 5.6% were admitted, 1.5% were transferred to another facility, 0.8% were held for observation, and 1.3% Left without being seen. There were no fatalities. Univariate analyses of factors associated with admission to the hospital are displayed in Table 4. Binomial logistic regression analysis (Table 5) re- vealed major orthopaedic injury (fracture/dislocation, amputation, Crush injury; OR = 17.75, 95% C.I. 14.57-21.61, p b 0.001), internal

organ injury (OR = 13.83, 95% C.I. 11.01-17.38, p b 0.001) concussion

(OR 4.06, 95% C.I. 2.10-7.82, p b 0.001), male gender (OR 2.60, 95% C.I.

2.26-2.99, p b 0.001), weekday presentation (OR = 2.51, 95% C.I.

2.15-2.93, p b 0.001), white race (OR = 2.26, 95% C.I. 1.89-2.69, p b

Admitted to inpatient 1815 (972-2657) 5.6%

Treated & transferred 472 (137-807) 1.5%

Left without being seen 416 (0-1023) 1.3%

Held for observation 271 (0-624) 0.8%

0.001), and each additional year of older age (OR = 1.044, 95% C.I. 1.041-1.046, p b 0.001) were independent predictors of hospital admis- sion after motorized scooter injury.

Distribution of injury

The anatomic distribution of motorized scooter injuries is demon- strated in Fig. 4. Head injuries were most common, representing 27.6% of all injuries. Patient diagnoses are described in Table 2. The most com- mon injuries were fracture or dislocation (25.9%), contusion (25.8%), and laceration (15.2%). Of the 8405 (95% C.I. 6346-10,305) fracture/dis- locations, 78 (95% C.I. 0-182) were dislocations. There were an esti- mated 14 (95% C.I. 0-43) (b0.1%) amputations reported in the last five years. The anatomic locations of fracture/dislocations are further de- scribed in Table 3. The wrist and lower arm constituted 35.4% of frac- tures or dislocations.

Fig. 3. Estimated motor scooter injury frequency by day of the week. *95% confidence interval bands around estimates.

Sunday

Saturday

Friday

Thursday

Wednesday

Tuesday

Monday

0

1000

2000

3000

4000

5000

6000

7000

8000

Estimated Motor Scooter Injury Frequency by Day of Week

Estimated Number of Injuries with 95% Confidence Intervals

Fig. 4. Anatomic distribution of motorized scooter injuries.

Discussion

Epidemiology of motorized scooter injuries

Motorized scooter use has become exceedingly popular since the mainstream launch of scooter share companies in 2017, with Bird and Lime each reporting N10 million rides since launch [3]. Although we were unable to detect a significant difference in incidence between 2013 and 2017, it is reasonable to anticipate a significant increase in the incidence of motorized scooter related injuries as their use con- tinues to rise. This is particularly relevant for millennials, for whom we noted a discrete uptick in injuries between 2016 and 2017. This study is likely underestimating the true burden of motorized scooter in- juries, as our data ends in 2017 and scooter share companies did not be- come widely popular until 2018 [3]. Despite this limitation, we believe it is important to characterize patterns of use, Injury profiles, and out- comes for motorized scooter riders.

The majority of injuries occurred in school age children (34.6%), who likely represent the majority of riders prior to scooter-shares. Bird and Lime prohibit children from riding their devices [14,15], however, media reports of underaged users remain prevalent [16]. While several states have helmet regulations for young bicycle riders, these regula- tions may not extend to electric kick scooter riders. If states do not have laws protecting young users of electric-scooters, state legislation should be considered.

The literature suggests that alternative Transportation methods are getting safer, or at least utilization is decreasing. Skateboard injuries continue to decline [8]. Bicycle injury severity may be decreasing [17]. In 2017, Motorcycle accidents decreased by 5.6% from the previous year [18] and motorcycle related injuries decreased 3% from 2014 to 2015 [19]. These safety improvement trends highlight the urgency in studying motorized scooters, a new and potentially dangerous form of transportation.

Risk factors for hospitalization

Multivariate analysis (Table 4) revealed older age, white race, and male gender were demographic factors that independently predicted hospitalization after motorized scooter injury. older people have been shown to have more severe injuries in other trauma settings [20]. older individuals who choose to ride motorized scooters should exer- cise particular caution as they are at increased risk of serious injury. Al- though weekend injuries were more common (Fig. 3), weekday presentation was associated with increased odds of hospitalization. Similarly, weekday Motor vehicle collisions have been previously asso- ciated with increased hospitalization [21], however, the underlying cause of these variations is not clear.

Head injuries and concussions

Head injuries constituted 27.6% of all injuries and were the most common site of injury. This striking number has the potential to serve as a foundation for future legislation requiring head protection for mo- torized scooter riders. Furthermore, concussions were diagnosed in 2.5% of patients, with an associated four-fold increased risk of hospital ad- mission. Concussions are notoriously difficult to diagnose [26] and are routinely underdiagnosed in emergency settings [27]. Therefore, our data likely grossly underestimates the true concussion incidence in this cohort.

Helmet laws are already an area of controversy for scooter share companies. California recently passed a state law removing helmet reg- ulations for electric scooter riders over the age of 18 [28]. This is

Table 3

Anatomic distribution of fractures and dislocations.

		Number of patients	Percentage of
		(n = 8405)	total
Head/neck	Skull	40 (0-82)	0.5%
	Facial	497 (142-852)	5.9%
	Neck	156 (0-377)	1.9%
Trunk	Upper trunk	401 (104-696)	4.8%
	Lower trunk	468 (101-835)	5.6%
Upper extremity	Shoulder	359 (70-649)	4.3%
	Upper arm	184 (0-408)	2.2%
	Elbow	277 (0-625)	3.3%
	Lower arm	1688 (890-2386)	20.1%
	Wrist	1290 (665-11,916)	15.3%
	Hand	108 (0-272)	1.3%
	Finger	176 (0-354)	2.1%
	Upper leg	94 (0-197)	1.1%
	Knee	57 (0-145)	0.7%
Lower extremity	Lower leg	1293 (643-1943)	15.4%
	Ankle	592 (267-916)	7.0%
	Foot	429 (110-748)	5.1%
	Toe	296 (26-566)	3.5%

orthopaedic injuries and preventati”>Table 4

Univariate analysis of factors associated with inpatient admission following motorized scooter injury.

Factor		Not admitted (n = 30,585)			Inpatient admission (n = 1815)
		Percentage	95% C.I.		Percentage	95% C.I.
Gender	Male	93.4% (18,147)	91.1-95.2%		6.6% (1279)	4.8-8.9%
	Female	95.9% (12,439)	90.1-98.3%		4.1% (536)	1.7-9.9%
Race	Non-White	96.6% (5100)	90.5-98.9%		3.4% (177)	1.8-9.5%
	White	92.4% (16,400)	88.5-95.1%		7.6% (1341)	1.6-4.9%
Time	Weekday	93.0% (19,393)	89.9-95.2%		7.0% (1456)	4.8-10.1%
Age	Weekend	96.9% (11,192) 26.8 (16) yearsa	93.6-98.5% 23.0-30.5 years		5.6% (1815) 49.1 (49) yearsa	1.5-6.4% 38.8-59.5 years
a Mean with median in	parentheses.

particularly concerning, given legislation requiring helmet use in other modes of transportation, including motorcycles and mopeds, has been shown to reduce the burden of maxillofacial injury and traumatic brain injury [29]. In our study, 6.4% of all fractures involved facial bones or the skull. These injuries could have potentially been averted with helmet use. Additionally, universal helmet laws for motorcycles and mopeds, such as those proposed in California and Rome, Italy, have been shown to be effective in increasing the proportion of helmet users from b20% to over 96% [30]. This study highlights the risk of head injuries associated with motorized scooter use. regulatory bodies should consider the data in this study as part of the scientific foundation supporting helmet regulation for motorized scooter riders.

Orthopaedic injuries and Preventative measures

Orthopaedic injuries were very common. The most common diagno- sis was a fracture or dislocation (Table 2). Upper extremity fractures or dislocations accounted for nearly half of all orthopaedic injuries, with “wrist” and “lower arm” being the most common sites of injury. Fur- thermore, major orthopaedic injuries were an independent risk factor for admission to the hospital after motorized scooter injury, perhaps im- plying that these patients required surgical treatment.

Wrist guards have been demonstrated to reduce the risk of fractures in biomechanical cadaveric studies [22-24], and in the skiing/snow- boarding literature, another medium-energy mechanism [25]. Given the frequency of Upper extremity fractures in our study, we strongly ad- vocate the use of wrist guards while riding motorized scooters until more specific research evaluating their efficacy becomes available. In addition to wrist guards, any reduction in speed would decrease the en- ergy of collisions and likely decrease the risk of serious orthopaedic in- jury for motorized scooter riders. Further research efforts should examine potentially safer scooter designs, such as more rugged wheels or three-wheeled scooters.

Limitations

This study examining data from the NEISS has inherent limitations. The NEISS does not include nuanced pre-hospital data, comprehensive clinical data, or detailed outcomes after initial triage in the ED. This

Table 5

Multivariate analysis of independent predictors of inpatient admission following motor- ized scooter injury.

Factor Odds ratio (OR) p-Value 95% C.I. Major orthopaedic injury^a,b 17.75 b0.001 14.57-21.61

Internal organ injury^a 13.83 b0.001 11.01-17.38

Concussion^a 4.06 b0.001 2.10-7.82

Male gender 2.6 b0.001 2.26-2.99

Weekday presentation 2.51 b0.001 2.15-2.93

White race 2.26 b0.001 1.89-2.69

Age (years) 1.044 b0.001 1.041-1.046

^a Reference category is soft tissue injury (hematoma, contusion, sprain and strain).

^b Includes fracture, dislocation, amputation, and Crush injuries.

dataset does not include injuries that did not present to an emergency department and thus likely underestimates the true incidence of motor- ized scooter injuries. As previously stated, it is likely that motorized scooter injuries have increased substantially in the last year, alongside the increase in scooter use, but 2018 NEISS data is not yet available [3]. The NEISS underreports mortality as it excludes pre-hospital and post-hospital deaths. NEISS narratives are imprecisely transcribed. Data on race was unknown for approximately one-third of the patients in the study, limiting our ability to meaningfully comment on racial dif- ferences in motorized scooter injuries. Additionally, based on review of the narratives, our data may include some injuries from mechanisms other than motorized scooters. The NEISS has distinct codes for moped (3215), dirt bike (5036) and minibikes (5035). The NEISS coding man- ual further specifies that powered wheelchair and motorcycle injuries should not be included in the dataset [31]. On review of the narratives, however, it was clear that some motorbike and wheelchair related inju- ries were misclassified and appeared in the initial database. To counter this potential confounder, we performed a search and exclusion of po- tentially miscoded patients as described in the methods. Despite these efforts, the data likely includes some misclassified wheelchair injuries. These potential inclusions are likely of minimal significance as a small fraction of the patients in the study are elderly. Lastly, we were unable to determine if patients in the study were intoxicated. Scooter use while under the influence of alcohol is an important safety concern. Fu- ture studies should consider the additional risks of scooter operation while intoxicated, as many individuals who would otherwise use a taxi, or walk, may be riding scooters.

Conclusions

Head injuries are the most common site of injury for motorized scooter riders presenting to the ED. A significant proportion of these patients had fractures and/or concussions. Extremity fractures or dislocations were the most common diagnosis. Through a better understanding of the injury pattern associated with motorized scooter use, stronger Injury prevention efforts can be devised and fu- ture Serious injuries can be averted. Future legislation to improve the safety of motorized scooters could consider strategies such as helmet use, extremity pads, and speed limits to protect riders from poten- tially devastating injuries during motorized scooter use.

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2019.03.049.

References

1. Hanlon. M., (The original) 100 year-old motorscooter up for sale. New Atlas; 2016.
2. Sipe N, Pojani D. Can e-scooters solve the ‘last mile’ problem? They’ll need to avoid the fate of dockless bikes, in The Conversation; 2018.
3. Hawkins, A., The electric scooter craze is officially one year old - what’s next.
4. Azad A. That electric scooter might be fun. It might also be deadly, in CNN; 2018.
5. Abcarian R. Bird scooters - so much fun, so damn dangerous, in Los Angeles Times; 2018.
6. Holley P. Police: Man dies after apparent electric scooter accident, in The Washington Post; 2018.
7. Walker A. Bird, Lime named in class-action lawsuit, in Curbed; 2018.
8. McIlvain C, et al. Injuries associated with hoverboard use: a review of the National Electronic Injury Surveillance System. Am J Emerg Med Mar 2019;37(3):472-7.
9. Siracuse BL, et al. Hoverboards: a new cause of pediatric morbidity. Injury 2017;48 (6):1110-4.
10. McAdams RJ, et al. Bicycle-related injuries among children treated in US emergency departments, 2006-2015. Accid Anal Prev 2018;118:11-7.
11. Lombardo DJ, et al. Extremity fractures associated with ATVs and dirt bikes: a 10- year national epidemiologic study. Musculoskelet Surg 2017;101(2):145-51.
12. National Electronic Injury Surveillance System . Available from https://www.

    cpsc.gov/Research-Statistics/NEISS-Injury-Data; November 17, 2018.

    Bureau, U.S.C. U.S. and World Population Clock 2018; Available from: https://www. census.gov/popclock/.

13. Lime. Lime-S Etiquette. Available from: https://www.li.me/electric-scooter.
14. Bird. Bird Rental Agreement, Waiver of Liability and Release. Available from: https:// www.bird.co/agreement/.
15. Kubzansky W. The secret life of teen scooter outlaws. Available from https://www. theverge.com/2018/9/23/17882996/teens-electric-scooter-age-requirement-bird- lime.
16. Kim YJ, et al. Trends in the incidence and outcomes of bicycle-related injury in the emergency department: a nationwide population-based study in South Korea, 2012-2014. PLoS One 2017;12(7):e0181362.
17. Motorcyclist Traffic Fatalities by State: 2017 Preliminary Data, G.H.S. Association, Ed-

    itor. 2017.

    Traffic Safety Facts: 2015 Data, N.H.T.S. Administration, Editor. 2017.

18. Bonne, S. and D.J. Schuerer, Trauma in the older adult: epidemiology and evolving geriatric trauma principles. Clin Geriatr Med, 2013. 29(1): p. 137-50.
19. Tiruneh A, et al. Are injury admissions on weekends and weeknights different from weekday admissions? Eur J Trauma Emerg Surg 2018. https://doi.org/10.1007/ s00068-018-1022-8.
20. Maurel ML, et al. Biomechanical study of the efficacy of a new design of wrist guard. Clin Biomech (Bristol, Avon) 2013;28(5):509-13.
21. Lewis LM, et al. Do wrist guards protect against fractures? Ann Emerg Med 1997;29 (6):766-9.
22. Choi WJ, Robinovitch SN. Pressure distribution over the palm region during forward falls on the outstretched hands. J Biomech 2011;44(3):532-9.
23. Kim, S. and S.K. Lee, Snowboard wrist guards-use, efficacy, and design. A systematic review. Bull NYU Hosp Jt Dis, 2011. 69(2): p. 149-57.
24. Sharp DJ, Jenkins PO. Concussion is confusing us all. Pract Neurol 2015;15(3): 172-86.
25. Rowe BH, et al. Concussion diagnoses among adults presenting to three Canadian emergency departments: missed opportunities. Am J Emerg Med 2018;36(12): 2144-51.
26. Chiland E. California removes helmet requirement for Electric scooters, in Curbed; 2018.
27. Adams NS, et al. The effects of motorcycle helmet legislation on craniomaxillofacial injuries. Plast Reconstr Surg 2017;139(6):1453-7.
28. La Torre G, et al. Head injury resulting from scooter accidents in Rome: differences before and after implementing a universal helmet law. Eur J Public Health 2007;17 (6):607-11.
29. Commission, U.S.C.P.S.. NEISS coding manual. www.cpsc.gov; 2017.