a b s t r a c t

Background: Electric scooters (e-scooters) have become a widespread method of transportation. The purpose of this study is to provide risk stratification tools for modifiable risk factors associated with e-scooter injury morbidity.

Methods: Patients at an urban Level 1 Trauma center sustaining E-scooter injuries between November 2017 through March 2020 were identified. Primary outcomes of interest were major trauma, as defined by an Injury Severity Score (ISS) >15, or hospital admission.

Results: A total of 442 patients sustained orthopaedic (51%), facial (31%), cranial (13%), and chest/abdominal injuries (4.5%). Rate of helmet use was 2.5%, hospital admission was 40.7%, and intensive care was 3%. Patients with facial injuries were half as likely to sustain major trauma as compared to orthopaedic injuries (p < 0.05). Factors with higher likelihood of hospital admission included age > 40 years (OR 4.20, p < 0.01), alcohol or other substance intoxication (OR 4.14 and 9.87, p < 0.001), loss of consciousness (OR 2.72, p < 0.003), or trans- port to the hospital by ambulance (OR 4.47, p < 0.001).

Conclusions: There is a substantial proportion of major trauma within e-scooter injuries. Modifiable risk factors for hospital admission include use of head protection and substance use while riding e-scooters.

(C) 2021

1. Introduction

Rapidly increasing urban congestion and pollution have led to a High demand for micromobility solutions. In 2018, Americans collectively took 38.5 million trips on electric scooters (e-scooters) [1]. Electric powered transportation services continue to gain traction worldwide due to their ability to provide transportation options to broader geo- graphical regions, minimize the first/last mile problem, decrease the carbon footprint of commuters, and provide widely affordable transpor- tation. As a result, the number of cities incorporating e-scooters into public transit has increased at a faster rate than regulatory measures [2]. Similarly, e-scooter Injury incidence in the United States rose by an estimated 222% from 2017 to 2018 with growing public health

? This work has not been previously presented or published. There was no outside source of funding utilized for this study. The institutional ethics and quality research board reviewed and approved conduction of this study.

* Corresponding author at: UCSD Medical Center, Hillcrest, 200 West Arbor Drive, San Diego, CA 92103, USA.

E-mail addresses: [email protected] (O. Lavoie-Gagne), [email protected] (M. Siow), [email protected] (W. Harkin), [email protected] (A.R. Flores), [email protected] (P.J. Girard), [email protected] (A.K. Schwartz), [email protected] (W.T. Kent).

concerns amidst reports of increasingly devastating e-scooter injuries including permanent disabilities and death [3-11].

Literature on e-scooter injuries has been limited to patients present- ing to the emergency department [3,5-7,11-18], trauma registry re- views [4,13,19], or subspecialty specific patient cohorts [10,16,20-22]. To our knowledge, no study to date has comprehensively investigated e-scooter injuries across all clinical settings or objectively assessed clin- ical severity of injuries. To this end, we investigated all patients present- ing to the emergency department, trauma bay, or outpatient clinics at an urban level 1 trauma center over a consecutive 27-month period with the purpose to (1) characterize e-scooter injuries and (2) provide risk stratification tools for risk of major trauma and hospital admission. We hypothesized a significant subset of patients with e-scooter injuries would sustain major trauma requiring higher level of care and that these injuries with higher morbidity would be associated with modifi- able risk factors.

1. Material and methods
   1. Patient identification and data extraction

This investigation was conducted in accordance to STROBE guide- lines for cross-sectional studies using retrospective data. Following

https://doi.org/10.1016/j.ajem.2021.02.019

0735-6757/(C) 2021

Institutional Review Board approval, the institutional Electronic Medical Record database was queried from November 1, 2017 to March 31, 2020 for all encounters with an associated diagnosis code pertaining to e-scooter injuries. Clinical contexts included presentation in an outpa- tient clinic at the institution, the emergency department (ED), or the trauma bay. Of note, the institution at which this study is organized with the trauma bay separate from the ED. The following International Classification of Diseases, Tenth Revision (ICD-10) codes were used for identification of e-scooter related injuries: V00.141A (Fall from scoter, initial), V00.141D (Fall from scooter, subsequent), V00.141S (Scooter acci- dent, sequelae), V00.142A (scooter colliding with stationary object), V00.148A (other scooter accident, initial), and V00.148D (other scooter accident, subsequent). Following removal of duplicates, records of clini- cal encounters with an associated e-scooter related injury diagnosis code were screened for the presence of any objective injury and injury mechanisms involving micromobility transportation e-scooter (i.e. ex- cluding assistive mobility devices for transient or permanent disability). Medical record review confirmed that all patients sustained one or more injury. Data of interest were explicitly defined prior to abstraction in a standardized guide. Full medical records were manually reviewed for demographic characteristics (gender, age, primary residence), con- text of injury (intoxication, mechanisms of injury, helmet use), clinical course (time of first medical evaluation, trauma protocol activation, ad- mission, surgery requirement, radioimaging), transport method (i.e. transported by paramedics versus self-presentation), injury diagnoses, laboratory results, and subspecialty consultation. Patients were then grouped into the following categories based on the body region with the highest Abbreviated Injury Scale score: orthopaedic, facial, cranial, or chest/abdominal injuries.

The primary outcome of this study consisted of major trauma, as de-

fined by injury severity scores (ISS) >15 [23]. The secondary outcome of this study was requirement of hospital admission for treatment of in- juries. At the institution of this study, patients receiving trauma protocol evaluation are routinely assigned an ISS for inclusion in the local trauma registry whereas those evaluated in the emergency department or out- patient clinic setting are not systematically encoded. Thus, ISS were ob- tained from the institutional trauma registry for patients initially evaluated in the trauma bay whereas those evaluated in the emergency department or outpatient clinic received independent ISS score assign- ments by two trained reviewers (*** and ***) using the same guidelines as for the trauma registry [24,25]. Disagreements on ISS scores assign- ments were resolved through mutual discussion with the senior author. Accurate and consistent ISS assignments were assessed via calculation of the kappa statistic for both inter and intra-rater reliability.

• 1. Statistical analysis

Descriptive data were summarized as counts and percentages. A Multivariable logistic regression models with major trauma as the pri- mary outcome was performed. Secondarily, a similar model was performed with hospital admission as the outcome. Independent variables included all patient characteristics that were present prior to clinical presentation. These consisted of gender, age, substance use, loss of consciousness (LOC) or concussion, transport to the institution via ambulance (BIBA), motor vehicle involvement, helmet use (as reported by patients, bystanders, or medical transportation personnel), and time of day during which the patient presented to the institution after e-scooter injury.

In order to provide further clinical guidance, the results of the logis- tic regressions were additionally presented as nomograms. Nomograms are a useful method of illustrating regression models for clinical risk stratification [26-29]. In this context, the prediction value from the no- mogram is the patient’s risk of experiencing one of the primary out- comes based on demographic and contextual factors [30,31]. The nomogram was calibrated using a bootstrapped approach with 500 iter- ations of testing and training datasets to calculate R-square and mean

standard error statistics to evaluate goodness of fit [32]. Calibration consisted of 500 bootstrap iterations to assess the Predictive ability of our model via mean absolute and mean squared error results. Nomo- gram validation is summarized in the supplementary materials. Statisti- cal analyses were conducted with two-tailed tests and significance was determined using alpha 0.05. All statistical analysis was performed using RStudio software version 3.6.2 (R Foundation for Statistical Com- puting, Vienna, Austria).

1. Results
   1. Patient identification and longitudinal trends

Database query yielded 490 unique records of which 442 met inclu- sion criteria (Fig. 1). There were 271 (61.3%) males and median age was

35.5 (IQR 27.0-47.8) years with a minority (6.3%) of patients >60 years old (Table 1). Incidence of injuries increased during the study period with a sharp decline during the last month of the study period due to so- cial distancing. Peak seasons of injury were during the spring and sum- mer months whereas peak times of initial clinical presentation were between 8 pm-10 pm.

Please see the Supplementary materials for additional epidemiolog- ical results.

• 1. Patient demographics and injuries

Patients with scooter injuries presenting to the ED or trauma bay accounted for 3% of the total trauma case volume during the study pe- riod. Of 442 patients with e-scooter-related injuries, 226 (51.1%) had or- thopaedic injuries, 138 (31.2%) had facial injuries, 58 (13.1%) had cranial injuries, and 20 (4.5%) had chest/abdominal injuries (Table 1). Modifiable mechanisms of injury included 28 (6.3%) scooter malfunctions, 116 (26.2%) riding on uneven ground surfaces, and 47 (10.6%) with motor vehicle involvement (Table 1). At the time of injury, only 11 (2.5%) of all included patients were wearing a helmet, 120 (27.1%) were under the influence of alcohol, and 32 (7.2%) were intox- icated with other recreational drugs (Table 1).

A total of 65 patients (14.7%) sustained major trauma [Median ISS: 5, IQR 2-9]. One patient deceased from injuries despite intensive care. This patient had experienced a diastatic occipital fracture with scattered sub- arachnoid and pontine hemorrhage followed by diffuse cerebral edema and cerebral herniation. Of a similar theme, almost one tenth of patients experienced some form of head injury with 18 (4.1%) having mild TBI (i.e. loss of consciousness or concussion), 18 (4.1%) having moderate TBI (i.e. head bleed treated non-operatively), and 5 (1.1%) having severe TBI (i.e. head bleed treated operatively and/or patient non-responsive or comatose). Almost half of the patients sustained a fracture (206, 46.6%) which consisted of 87 (19.7%) upper extremity fractures, 70

(15.8%) Facial fractures, 48 (10.9%) lower extremity fractures, and 13 (2.9%) spine or rib fractures (Table 1). A majority of patients received x-ray (75.3%) or non-contrast computed tomography (CT, 58.8%) as part of their clinical evaluation (Table 1).

Please see the Appendix for further details of injuries by body region.

• 1. Clinical course

Incidence of clinical presentation of e-scooter injuries was highest during evening (36%), followed by afternoon (28.1%), and early morn- ing (16.7%) periods of the day (Table 1). After initial evaluation, 180 (40.7%) patients were admitted for hospitalization with median length of stay of 1 day (IQR 1-3) and 14 (3.2%) patients requiring intensive care (Table 1). Seventy-two (16.3%) patients underwent surgical treat- ment for their injuries of which 11 required emergent surgical interven- tion (2.5%) (Table 1).

Most patients (192, 43.4%) received at least 1 subspecialty consulta- tion for management of their e-scooter related injuries. The most

Fig. 1. Flow diagram of included patients.

commonly consulted surgical subspecialty was trauma surgery (141, 31.9%) followed by orthopaedic trauma (62, 14.0%) and otolaryngology (52, 11.8%) (Table 1). Ophthalmology, neurosurgery, orthopaedic hand, and plastic surgery had consultation rates of approximately 10% (Table 1). Orthopaedic spine (1.6%) was the least consulted service (Table 1).

• 1. Risk of major trauma

Multivariable analysis revealed patients between 40 and 59 years of age and those transported to the institution by ambulance were more than twice as likely to sustain major trauma [Odds Ratio 2.19 (1.21-4.07) and 2.12, (1.17-3.84), respectively] (Table 2). On the other hand, factors not significantly associated with risk of major trauma included male gender [Odds Ratio 0.84 (0.50-1.43)], alcohol in- toxication [Odds Ratio 1.46 (0.80-2.65)], other intoxication [Odds Ratio

1.53 (0.64-3.49)], motor vehicle involvement [Odds Ratio 1.50 (0.70-3.12)], helmet use [Odds Ratio (0.09-3.03)], and time of day [Odds Ratio ranging from 0.40 to 1.35] (Table 2). Risk stratification no- mogram of major trauma revealed factors with the largest influence on risk for major trauma included age > 39, transport to the institution by ambulance, injuries sustained during early morning and late morn- ing times of day, intoxication, motor vehicle involvement, and lack of helmet use (Fig. 2).

• 1. Risk of hospital admission

Multivariable analysis revealed age 40-59 years [Odds Ratio 2.54 (1.34-4.88)], age > 60 years [Odds Ratio 4.20 (1.47-11.99)], alcohol

intoxication [Odds Ratio 4.14 (2.18-8.01)], other intoxication [Odds Ratio 9.87 (3.16-38.29), loss of consciousness [Odds Ratio 2.72 (1.39-5.32)], and transport by ambulance [Odds Ratio 4.47 (2.54-7.93)] to have significantly higher risk of hospital admission (Table 2). Please see the appendix for the nomogram depicting risk of hospital admission.

1. Discussion

In this series, we provide a comprehensive review of 442 patients presenting for e-scooter injuries in the emergency department, trauma bay, and outpatient settings during a 27-month period at an urban level 1 trauma center located in within the largest concentration of e-scooters in America [1]. We characterize the morbidity of e-scooter injuries; identify modifiable risk factors associated with major trauma and hospi- tal admission; and provide risk stratification tools for prediction of major trauma and hospital admission. Overall, a significant subset (40.7%) of patients required hospital admission for treatment of e-scooter injuries and a substantial proportion of patients (14.7%) sustained injuries consistent with major trauma. Factors with the larg- est risk of sustaining major trauma included age > 39, transport to the institution by ambulance, injuries sustained during early morning and late morning times of day, intoxication, motor vehicle involvement, and lack of helmet use. Identification of modifiable risk factors associ- ated with higher e-scooter injury morbidity may both guide clinical risk stratification and provide initial guidance for future Public health efforts in mitigating improved safety of e-scooters.

Injury incidence and patient demographics of this series’ co- hort are consistent with national representation. E-scooters’

Table 1

Cohort demographics, injury context, and clinical management.

Patient demographics and injury context	All patients	Not major trauma	Major trauma
	(N = 442)	(N = 345)	(N = 97)
Age (years)
<30	149 (33.7%)	125 (36.2%)	24 (24.7%)
30-39	113 (25.6%)	93 (27.0%)	20 (20.6%)
40-59	152 (34.4%)	107 (31.0%)	45 (46.4%)
>60	28 (6.3%)	20 (5.8%)	8 (8.3%)
Male	271 (61.3%)	208 (60.3%)	63 (64.9%)
Out of town visitor1	78 (17.6%)	65 (18.8%)	13 (13.4%)
Injury context
Primary injury
Orthopaedic	226 (51.1%)	111 (32.2%)	27 (27.8%)
Facial	138 (31.2%)	178 (51.6%)	48 (49.5%)
Cranial	58 (13.1%)	40 (11.6%)	18 (18.6%)
Chest/Abdominal	20 (4.5%)	16 (4.6%)	4 (4.1%)
Any intoxication2	134 (30.3%)	90 (26.1%)	44 (45.4%)
alcohol intoxication3	120 (27.1%)	79 (22.9%)	41 (42.3%)
Other intoxication4	32 (7.2%)	21 (6.1%)	11 (11.3%)
Scooter malfunction	28 (6.3%)	21 (6.1%)	7 (7.2%)
Uneven ground	116 (26.2%)	97 (28.1%)	19 (19.6%)
Motor vehicle involvement	47 (10.6%)	34 (9.9%)	13 (13.4%)
Helmet use5	11 (2.5%)	9 (2.6%)	2 (2.1%)
BIBA6	169 (38.2%)	112 (32.5%)	57 (58.8%)
Clinical course
Hospital admission	180 (40.7%)	100 (29.0%)	80 (82.5%)
Admission to ICU7	14 (3.2%)	5 (1.4%)	9 (9.3%)
Trauma protocol activation8	97 (21.9%)	63 (18.3%)	34 (35.1%)
Time of medical evaluation
Outpatient	21 (4.8%)	50 (14.5%)	24 (24.7%)
Morning (06:01-12:00)	64 (14.5%)	18 (5.2%)	3 (3.1%)
Afternoon (12:01-18:00)	124 (28.1%)	49 (14.2%)	15 (15.5%)
Evening (18:01-23:59)	159 (36.0%)	107 (31.0%)	17 (17.5%)
Early morning (00:00-06:00)	74 (16.7%)	121 (35.1%)	38 (39.2%)
Surgery	72 (16.3%)	20 (5.8%)	52 (53.6%)
Urgent surgery	11 (2.5%)	1 (0.3%)	10 (10.3%)
Injuries Major trauma9	97 (21.9%)	0 (0.0%)	97 (100.0%)
Mild TBI10	18 (4.1%)	18 (5.2%)	0 (0.0%)
Moderate TBI11	18 (4.1%)	6 (1.7%)	12 (12.4%)
Severe TBI12	5 (1.1%)	0 (0.0%)	5 (5.2%)
Any fracture13	206 (46.6%)	135 (39.1%)	71 (73.2%)
Lower extremity	48 (10.9%)	22 (6.4%)	26 (26.8%)
Upper extremity	87 (19.7%)	66 (19.1%)	21 (21.6%)
Facial	70 (15.8%)	47 (13.6%)	23 (23.7%)
Spine or Rib	13 (2.9%)	5 (1.4%)	8 (8.2%)
Radioimaging
X-ray	333 (75.3%)	249 (72.2%)	84 (86.6%)
non-contrast CT14	260 (58.8%)	184 (53.3%)	76 (78.4%)
Contrast CT14	18 (4.1%)	6 (1.7%)	12 (12.4%)
MRI15	9 (2.0%)	5 (1.4%)	4 (4.1%)
Subspecialty consultations
Trauma	141 (31.9%)	90 (26.1%)	51 (52.6%)
Orthopaedic trauma	62 (14.0%)	30 (8.7%)	32 (33.0%)
Otolaryngology	52 (11.8%)	29 (8.4%)	23 (23.7%)
Ophthalmology	46 (10.4%)	22 (6.4%)	24 (24.7%)
Neurosurgery	42 (9.5%)	16 (4.6%)	26 (26.8%)
Orthopaedic hand	42 (9.5%)	27 (7.8%)	15 (15.5%)
Plastics	39 (8.8%)	20 (5.8%)	19 (19.6%)
Orthopaedic spine	7 (1.6%)	1 (0.3%)	6 (6.2%)

¹ As reported by patient or clinical plan for follow-up visits.

² Intoxication as defined by clinical evaluation, positive BAL, or urine toxicology.

³ Blood alcohol level (BAL) > 11 millimol/L or clinical intoxication.

⁴ Including positive urine toxicology.

⁵ As reported by patient, bystanders, and/or medical transport staff.

⁶ Brought in by ambulance (BIBA).

⁷ Intensive Care Unit (ICU).

⁸ Trauma Bay evaluation (trauma bay is separate from emergency department at the representative institution.

⁹ Defined as injury severity score > 15.

¹⁰ Mild Traumatic brain injury including loss of consciousness or concussion.

¹¹ Moderate TBI including brain bleeds treated non-operatively.

¹² Severe TBI including patients with brain bleeds requiring surgery or non-responsive/comatose.

¹³ Patients experiencing any fracture. Patients may be represented more than once in subgroup categories (lower extremity, upper extremity, facial, and spine/rib).

¹⁴ Computed Tomography (CT).

¹⁵ Magnetic Resonance Imaging (MRI).

Table 2

multivariable regressions of characteristics associated with higher odds of major trauma (ISS > 15) or hospital admission.

	Odds ratio	95% CI?
Major trauma
Male	0.84	0.50-1.43
Age (years)
<30	Ref.
30-39	0.98	0.49-1.95
40-59	2.19	1.21-4.07
>60	2.29	0.81-6.13
Primary injury
Alcohol intoxication	1.46	0.80-2.65
Other intoxication	1.53	0.64-3.49
LOC or concussion+	1.14	0.62-2.07
Brought in by ambulance	2.12	1.17-3.84
Motor vehicle involvement	1.50	0.70-3.12
Helmet use	0.68	0.09-3.03
Time of first medical contact
Outpatient	Ref.
Early morning (00:00-06:00)	1.35	0.37-6.52
Morning (06:01-12:00)	1.08	0.29-5.20
Afternoon (12:01-18:00)	0.40	0.11-1.96
Evening (18:01-23:59)	0.91	0.27-4.20
Admission to hospital
Male	0.69	0.39-1.18
Age (years)
<30	Ref.
30-39	1.71	0.87-3.41
40-59	2.54	1.34-4.88
>60	4.20	1.47-11.99
Primary injury Orthopaedic	Ref.
Facial	0.70	0.37-1.28
Cranial	1.73	0.67-4.56
Chest/Abdominal	1.17	0.33-3.81
Alcohol intoxication	4.14	2.18-8.01
Other intoxication	9.87	3.16-38.29
LOC or concussion+	2.72	1.39-5.32
Brought in by ambulance	4.47	2.54-7.93
Motor vehicle involvement	1.60	0.68-3.77
Helmet use	0.35	0.05-2.09
Time of first medical contact
Outpatient	Ref.
Early morning (00:00-06:00)	1.49	0.38-7.50
Morning (06:01-12:00)	0.93	0.24-4.62
Afternoon (12:01-18:00)	0.78	0.21-3.77
Evening (18:01-23:59)	1.15	0.33-5.44

* 95% Confidence Intervals.

⁺ Loss of Consciousness (LOC).

overtake of shared micromobility in 2018 led to a national in- crease in e-scooter injury rates, mirrored within the trends of the presently included cohort during the study period [1]. Recent National Electronic Injury surveillance System (NEISS) data dem- onstrated a predominance of injuries within the 18 to 34 age range with 64% of those injured being male [17]. Similar institu- tional reviews in Southern California and Texas, currently the two largest centers of e-scooter share systems in the country with over 10,000 available e-scooters, have reported comparable demographics to this current cohort of e-scooter injuries [5,6].

In this series, e-scooter injury morbidity encompassed a wide range from minor abrasions to ISS of 75 for a patient deceased secondary to in- juries sustained. While several studies have reported injury rates based on injury type and body part [3,5,6,10,17,18,20], only a few subspecialty specific reviews address e-scooter injury morbidity [3,6,10,20]. The presently reported range of injury severity is not unique to this patient cohort. Previously published extremity fracture rates range from 27 to 31.7% [6,17], equivocal to our overall extremity fracture rate of 31.6%. In- terestingly, this cohort’s facial fracture rate of 15.8% is half of the single other reported rate in the literature despite our inclusion of higher acu- ity patients [6]. This may be due to our concomitant inclusion of

outpatient visits for e-scooter injuries and thus may be more represen- tative of national incidence.

Regarding severity of head injuries, the range of injuries reported in the present cohort is consistent with current literature. A case series of neurosurgical consults for e-scooter injuries reports on several patients with intracranial bleeds (subdural hemorrhage, subarachnoid hemor- rhage, deep intraparenchymal punctate hemorrhage), TBI, spinal com- pression fractures, and one death [20]. Another study investigating craniofacial trauma reported 11.5% of their cohort experiencing either subarachnoid hemorrhage or subdural hemorrhage [5]. Overall re- ported head injury rates in the literature range from 20 to 40.4% with varying criteria for definitions of head injuries, making it challenging to compile evidence against the perceived relative safety of riding e-scooters without head protection [3,5,6,17].

In terms of e-scooter injury context, the presently reported alcohol intoxication rate of 27.1% is much higher than previously reported rates of 5.2-18.0% in American studies [5,6], but similar to European and Australian reports of 27-41.9% [11-13,19]. In addition to this series’ decreased selection bias due to inclusion of patients presenting in all clinical contexts rather than just the emergency department or trauma registry, these variances may also be due to different institutional med- ical record documentation practices. The presently reported 2.5% hel- met use falls within the consistently low rates around the globe ranging from 0 to 5.6% [3,5,6,12,15], with the exception of Australia’s impressive 46% of riders wearing helmets at the time of injury [19]. Fi- nally, the significant involvement of uneven ground (26.2%), scooter malfunctions (6.3%), and motor vehicles (10.6%) in the present cohort’s injury mechanism is corroborated by the current literature [6,12].

Risk stratification of e-scooter injury morbidity is imperative to both clinical evaluation of individual patients and public health efforts to im- prove regulatory Safety measures surrounding e-scooter use. E-scooters have largely replaced pedal bicycles and success of the United States for- mal e-scooter share pilots launched in summer 2018 will likely lead to continued expansion of the industry [1,33]. In many states, e-scooters are not defined in motor vehicle codes and thus are not explicitly cov- ered in local operational law. Recent state-specific legislation proposals by e-scooter companies to legalize e-scooter use have incorporated clauses preempting local city authorities from regulating shared micromobility services [1]. The amounting global literature reporting increasingly morbid e-scooter injuries is clear evidence of the need for improved public safety regulations [3,4,6,11-13,15,17-19,22]. The issue at hand is complex with corporate social responsibility, medical ethics, public health ethics, and government stewardship at play [34-37]. Nonetheless, this study identifies several targetable risk factors associ- ated with increased e-scooter injury morbidity, a pivotal step in improv- ing the minimal public health safety measures currently in place.

Modifiable factors independently associated with increased injury morbidity included substance use while riding e-scooters and loss of consciousness. Most intuitive is the regulation of adequate head protec- tion [38]. Helmet use while riding e-scooters may provide equivalent benefits to use while on bicycles with an up to 80% reduction in odds of head injury [19,38,39]. Unfortunately, e-scooter company social media promotion is reported to portray riders wearing protective gear in a mere 6.79% of content with completely absent written content re- garding protective gear [40,41]. Kinetic and finite element analysis of e-scooter crashes has shown worsened brain injuries with increasing speeds of scooters for ground impact and similar results for increasing speeds of e-scooter and motor vehicles in e-scooter versus auto colli- sions [42,43].

Although e-scooters have been shown to provide better impact pro- tection for riders as compared to pedestrian versus vehicle and bicycle versus vehicle collisions [43], this is likely negated when considering e-scooter riders’ more risky behaviors [44], 50% higher risk of traffic crashes [45], predominant use for recreational transport on weekend nights [1], and ability to reach unregulated speeds of up to 30 miles per hour [2]. Compounded with the impaired judgment under the

Fig. 2. Risk nomogram of major trauma following electric scooter accident. Risk of major trauma can be predicted by drawing a straight line from each characteristic to the points scale at the top of the nomogram. Points from each row should be added and the sum of the total points located on the “Total Points” scale at the bottom of the nomogram. A vertical straight line can then be visualized from the appropriate location on the total points scale down to the “Major Trauma” scale to predict the risk of major trauma.

influence of alcohol or other recreational drugs, it is unsurprising to note the high proportion of high-energy e-scooter injuries, higher incidence of injuries in the evening, and higher risk of major trauma for those pa- tients presenting in the evening or early morning. Finally, it is notewor- thy that the present cohort included 2 pedestrians involved in e-scooter collisions. Pedestrians with vision or hearing impairment, mobile device distraction, young children, or the elderly are most at risk for potentially devastating injuries secondary e-scooter collisions [46]. Regulation of head protection, driving under the influence, and designated riding lanes could each significantly decrease the severity of e-scooter injuries seen by medical professionals.

• 1. Limitations

This study has several strengths and limitations. Use of ICD-10 diag- nosis codes for identification of our patient population may have incor- porated a selection bias towards more severe injuries. While this method successfully included most patients evaluated for scooter- related injuries, it relies on accurate assignment of Diagnosis codes by providers. Providers may not be aware of the available ICD-10 code specification for e-scooter injuries or may not code injuries according to mechanism of injury. Additionally, patients with very minor injuries may not have disclosed to their provider the association of e-scooters with their injury. Thus, our patient population may be skewed towards more severe injuries that warranted patient interview and documenta- tion by multiple providers. Secondly, abstraction of helmet use prior to injury relied on medical provider query of helmet use and subsequent medical record documentation. This may have lead to underrepresenta- tion of helmet use within the present cohort. On the other hand, incor- poration of patients’ self-reported helmet use portends the risk of overrepresentation. Overall, the risk of over or under-representation of helmet use were likely minimized as the currently reported rates are similar to prior observational studies [6]. the majority of patients had medical record documentation from more than one provider

(increasing the likelihood of more inclusive documentation within the medical record), and the majority of patients had explicit documenta- tion of answering negatively to the question of ‘were you wearing a hel- met when you were injured’. Finally, as a retrospective review this study does not encompass long-term outcomes of consequences secondary to e-scooter injuries. Nonetheless, this review reports a valuable investiga- tion of modifiable risk factors for future public health advocacy of safety measures targets to minimize morbid e-scooter injuries.

As e-scooters flood the shared micromobility industry with minimal public safety regulations, the incidence of both minor and major e- scooter injuries is likely to continue to rise. There is a substantial propor- tion of major trauma within e-scooter injuries with continued low rates of helmet use. Modifiable risk factors for hospital admission include use of head protection and substance use while riding e-scooters. Future in- vestigations may provide further insight into effective preventative Public health measures to decrease risk of morbid e-scooter injuries.

Declarations

This work has not been previously presented or published. There were no outside sources of funding used in this study and the authors have no conflicts of interest to disclose. The institutional ethics and quality research board reviewed and approved conduction of this study with a waiver of informed consent due to the retrospective nature of the study (UCSD Human Research Protections Program #191310x). The senior author had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Contributorship

Ophelie Lavoie-Gagne: Contributions to the conception and design of the work; data acquisition; data analysis and interpretation; drafting and critical review of work.

Matthew Siow: Contributions to the conception and design of the work; data acquisition; data analysis and interpretation; drafting and critical review of work.

Will E. Harkin: Data acquisition; critical review of work. Alec R. Flores: Data acquisition; critical review of work.

Paul J. Girard: Drafting and critical review of work; final approval of version to be published; agreement to be accountable for all aspects work related to accuracy and integrity.

Alexandra K. Schwartz: Drafting and critical review of work; final ap- proval of version to be published; agreement to be accountable for all aspects work related to accuracy and integrity.

William T. Kent: Contributions to the conception and design of the work; Data interpretation; drafting and critical review of work; final ap- proval of version to be published; agreement to be accountable for all aspects work related to accuracy and integrity.

Author contribution

Ophelie Lavoie-Gagne: literature search, study design, data collec- tion, data analysis, data interpretation, writing, critical revision.

Matthew Siow: literature search, study design, data collection, data analysis, data interpretation, writing, critical revision.

William E. Harkin: data collection, data analysis, data interpretation, critical revision.

Alec R. Flores: data collection, data analysis, data interpretation, crit- ical revision.

Paul J. Girard: data interpretation, writing, critical revision. Alexandra K. Schwartz: data interpretation, writing, critical revision. William T. Kent: literature search, study design, data interpretation,

writing, and critical revision.

Declaration of Competing Interest

There were no outside sources of funding used in this study and the authors have no conflicts of interest to disclose.

Acknowledgments

The senior author had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2021.02.019.

References

1. Shared Micromobility in the US: 2018, 2019. [accessed 2019, July 6].
2. Herrman M. A Comprehensive Guide to Electric Scooter Regulation Practices. De- partment of Landscape Architecture and Regional & Community Planning. Master of Regional and Community Planning. Kansas State University; 2019.
3. Badeau A, Carman C, Newman M, Steenblik J, Carlson M, Madsen T. Emergency de- partment visits for electric scooter-related injuries after introduction of an urban rental program. Am J Emerg Med. 2019. https://doi.org/10.1016/j.ajem.2019.05.003.
4. Siman-Tov M, Radomislensky I, Israel Trauma G, Peleg K. The casualties from electric bike and motorized scooter road accidents. Traffic Inj Prev. 2017;18(3):318-23. https://doi.org/10.1080/15389588.2016.1246723.
5. Trivedi B, Kesterke MJ, Bhattacharjee R, Weber W, Mynar K, Reddy LV. Craniofacial injuries seen with the introduction of bicycle-share electric scooters in an urban set- ting. J Oral Maxillofac Surg. 2019;77(11):2292-7. https://doi.org/10.1016/j.joms. 2019.07.014.
6. Trivedi TK, Liu C, Antonio ALM, Wheaton N, Kreger V, Yap A, et al. Injuries associated with standing electric scooter use. JAMA Netw Open. 2019;2(1):e187381. https:// doi.org/10.1001/jamanetworkopen.2018.7381.
7. Aizpuru M, Farley KX, Rojas JC, Crawford RS, Moore Jr TJ, Wagner ER. motorized scooter injuries in the era of scooter-shares: a review of the national electronic sur- veillance system. Am J Emerg Med. 2019;37(6):1133-8. https://doi.org/10.1016/j. ajem.2019.03.049.
8. Shang S, Zheng Y, Shen M, Yang X, Xu J. Numerical investigation on head and brain injuries caused by windshield impact on riders using electric self-balancing scooters. Appl Bionics Biomech. 2018;2018:5738090. https://doi.org/10.1155/2018/5738090.
9. Alwani M, Jones AJ, Sandelski M, Bandali E, Lancaster B, Sim MW, et al. Facing facts: facial injuries from stand-up electric scooters. Cureus. 2020;12(1):e6663. https:// doi.org/10.7759/cureus.6663.
10. Ishmael CR, Hsiue PP, Zoller SD, Wang P, Hori KR, Gatto JD, et al. An early look at op- erative orthopaedic injuries associated with electric scooter accidents: bringing high-energy Trauma to a wider audience. J Bone Joint Surg Am. 2020;102(5):e18. https://doi.org/10.2106/JBJS.19.00390.
11. Puzio TJ, Murphy PB, Gazzetta J, Dineen HA, Savage SA, Streib EW, et al. The electric scooter: a surging new Mode of transportation that comes with risk to riders. Traffic Inj Prev. 2020;21(2):175-8. https://doi.org/10.1080/15389588.2019.1709176.
12. Blomberg SNF, Rosenkrantz OCM, Lippert F, Collatz Christensen H. Injury from elec- tric scooters in Copenhagen: a retrospective cohort study. BMJ Open. 2019;9(12): e033988. https://doi.org/10.1136/bmjopen-2019-033988.
13. Brownson AB, Fagan PV, Dickson S, Civil ID. Electric scooter injuries at Auckland City Hospital. N Z Med J. 2019;132(1505):62-72https://www.ncbi.nlm.nih.gov/ pubmed/31697664.
14. Beck S, Barker L, Chan A, Stanbridge S. Emergency department impact following the introduction of an electric scooter sharing service. Emerg Med Australas. 2019. https://doi.org/10.1111/1742-6723.13419.
15. Liew YK, Wee CPJ, Pek JH. New peril on our roads: a retrospective study of electric scooter-related injuries. Singapore Med J. 2020;61(2):92-5. https://doi.org/10. 11622/smedj.2019083.
16. Mayhew LJ, Bergin C. Impact of e-scooter injuries on emergency department imag- ing. J Med Imaging Radiat Oncol. 2019;63(4):461-6. https://doi.org/10.1111/1754- 9485.12889.
17. Namiri NK, Lui H, Tangney T, Allen IE, Cohen AJ, Breyer BN. Electric Scooter Injuries and Hospital Admissions in the United States, 2014-2018. JAMA Surg. 2020. https:// doi.org/10.1001/jamasurg.2019.5423.
18. 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. 2019. https://doi.org/10.1136/injuryprev-2019-043418.
19. Mitchell G, Tsao H, Randell T, Marks J, Mackay P. Impact of electric scooters to a ter- tiary emergency department: 8-week review after implementation of a scooter share scheme. Emerg Med Australas. 2019;31(6):930-4. https://doi.org/10.1111/ 1742-6723.13356.
20. Schlaff CD, Sack KD, Elliott RJ, Rosner MK. Early experience with electric scooter in- juries requiring neurosurgical evaluation in District of Columbia: a case series. World Neurosurg. 2019;132:202-7. https://doi.org/10.1016/j.wneu.2019.08.237.
21. Campbell A, Wong N, Monk P, Munro J, Bahho Z. The cost of electric-scooter related orthopaedic surgery. N Z Med J. 2019;132(1501):57-63 https://www.ncbi.nlm.nih. gov/pubmed/31465328.
22. Schapiro AH, Lall NU, Anton CG, Trout AT. Hoverboards: spectrum of injury and as- sociation with an uncommon fracture. Pediatr Radiol. 2017;47(4):437-41. https:// doi.org/10.1007/s00247-016-3766-9.
23. Sewalt CA, Venema E, Wiegers EJA, Lecky FE, Schuit SCE, Den Hartog D, et al. Trauma models to identify major trauma and mortality in the prehospital setting. BJS. 2019. https://doi.org/10.1002/bjs.11304.
24. Tamim H, Al Hazzouri AZ, Mahfoud Z, Atoui M, El-Chemaly S. The injury severity score or the New Injury Severity Score for predicting mortality, intensive care unit admission and length of hospital stay: experience from a university hospital in a de- veloping country. Injury. 2008;39(1):115-20.
25. Baker S, O’Neil B, Haddon WJ, Long W. The injury severity score: a method for de- scribing patients with multiple injuries and evaluating emergency care. J Trauma. 1974;14(3):187-96.
26. Barsoum WK, Murray TG, Klika AK, Green K, Miniaci SL, Wells BJ, et al. Predicting pa- tient discharge disposition after total joint arthroplasty in the United States. J Arthroplast. 2010;25(6):885-92. https://doi.org/10.1016/j.arth.2009.06.022.
27. Wuerz TH, Kent DM, Malchau H, Rubash HE. A nomogram to predict major compli- cations after hip and knee arthroplasty. J Arthroplast. 2014;29(7):1457-62. https:// doi.org/10.1016/j.arth.2013.09.007.
28. Gronbeck C, Cote MP, Halawi MJ. Predicting inpatient status after primary total knee arthroplasty in medicare-aged patients. J Arthroplasty. 2019. https://doi.org/10. 1016/j.arth.2019.03.009.
29. Ogura K, Yasunaga H, Horiguchi H, Fushimi K, Tanaka S, Kawano H. Nomogram predicting severe adverse events after musculoskeletal tumor surgery: analysis of a national administrative database. Ann Surg Oncol. 2014;21(11):3564-71. https://doi.org/10.1245/s10434-014-3760-0.
30. Jalali A, Alvarez-Iglesias A, Roshan D, Newell J. Visualising statistical models using dynamic nomograms. PLoS One. 2019;14(11):e0225253. https://doi.org/10.1371/ journal.pone.0225253.
31. Dunn WR, Kuhn JE, Sanders R, An Q, Baumgarten KM, Bishop JY, et al. Symptoms of pain do not correlate with rotator cuff tear severity: a cross-sectional study of 393 patients with a symptomatic atraumatic full-thickness rotator cuff tear. J Bone Joint Surg Am. 2014;96(10):793-800. https://doi.org/10.2106/JBJS.L.01304.
32. Calcagno V. Model selection and multimodel inference made easy. J Stat Softw. 2013;34(12).
33. First/Last Mile Solutions. In: Resources RaT, ed.; 2018.
34. Krebs J. The importance of public-health ethics. Bull World Health Organ. 2008;86 (8):577-656.
35. The World Health Report 2000. Health systems: Improving performance Geneva ; 2000.
36. Public health: ethical issues. London; 2007.
37. Jochelson K. Nanny or steward? The role of government in public health. London: Kings’ Fund; 2005.
38. Rivara FP. Shareable 2-wheeled vehicles-a new Public health problem? JAMA Netw Open. 2019;2(1):e187407. https://doi.org/10.1001/jamanetworkopen.2018.7407.
39. Thompson DC, Rivara FP, Thompson RS. Effectiveness of bicycle safety helmets in preventing head injuries. A case-control study. JAMA. 1996;276(24):1968-73 https://www.ncbi.nlm.nih.gov/pubmed/8971066https://jamanetwork.com/ journals/jama/article-abstract/412317.
40. Allem JP, Majmundar A. Are electric scooters promoted on social media with safety in mind? A case study on Bird’s Instagram. Prev Med Rep. 2019;13:62-3. https://doi. org/10.1016/j.pmedr.2018.11.013.
41. Dormanesh A, Majmundar A, Allem JP. Follow-up investigation on the promotional practices of electric scooter companies: content analysis of posts on Instagram and twitter. JMIR Public Health Surveill. 2020;6(1):e16833. https://doi.org/10.2196/ 16833.
42. Xu J, Shang S, Qi H, Yu G, Wang Y, Chen P. Simulative investigation on head injuries of electric self-balancing scooter riders subject to ground impact. Accid Anal Prev. 2016;89:128-41. https://doi.org/10.1016/j.aap.2016.01.013.
43. Xu J, Shang S, Yu G, Qi H, Wang Y, Xu S. Are electric self-balancing scooters safe in vehicle crash accidents? Accid Anal Prev. 2016;87:102-16. https://doi.org/10. 1016/j.aap.2015.10.022.
44. Bai L, Liu P, Guo Y, Yu H. comparative analysis of risky behaviors of electric bicycles at signalized intersections. Traffic Inj Prev. 2015;16(4):424-8. https://doi.org/10. 1080/15389588.2014.952724.
45. Qian Y, Sun Q, Fei G, Li X, Stallones L, Xiang H, et al. Riding behavior and electric bike traffic crashes: a Chinese case-control study. Traffic Inj Prev. 2020;21(1):24-8. https://doi.org/10.1080/15389588.2019.1696963.
46. Sikka N, Vila C, Stratton M, Ghassemi M, Pourmand A. Sharing the sidewalk: a case of E-scooter related pedestrian injury. Am J Emerg Med. 2019. https://doi.org/10.1016/ j.ajem.2019.06.017.