Understanding nonuse of mandatory e-scooter helmets

Abstract

Objectives

Head injuries resulting from e-scooter use have led to calls for helmet use to be promoted or mandatory. Helmet use is mandatory for e-scooters in Australia but observational studies have reported significant levels of nonuse, particularly by riders of shared e-scooters. The aim of this study is to understand whether nonuse in the mandatory context is a consistent behavior for an individual or is situationally-influenced, and what are the factors associated with nonuse.

Methods

An online survey was completed between 2022 and 2023 by 360 adult e-scooter riders in Canberra, Australia. Riders were asked whether they had worn a helmet on their last ride and how often they had not worn a helmet when riding in the last 30 days. The survey also asked about rider characteristics (demographics, frequency of e-scooter and bicycle use, perceived risk of e-scooter use, e-scooter ownership, and risky behaviors while riding), trip duration and perceptions of the helmet requirement (knowledge of and support for the law).

Results

Respondents were mostly male, young, highly educated, and full-time workers. Of the 29.1% of riders who reported riding without a helmet in the last 30 days, 24.4% had worn a helmet at least once during that period and 4.8% had consistently not worn a helmet. Younger age, shared e-scooter use and more frequent riding frequency (shared e-scooters only) were associated with helmet nonuse in the bivariate analyses but not in the logistic regression. Logistic regression showed that the independent predictors of helmet nonuse were the number of risky riding behaviors, lack of knowledge, and lack of support for the law.

Conclusions

Most nonuse of helmets in a mandatory context seems to be situational, rather than consistent. Many of the factors associated with nonuse of helmets for e-scooters are similar to those reported for bicycles. Nonuse of helmets appears to be one of a number of risky behaviors performed by riders, rather than being primarily an outcome that is specific to factors associated with helmets (e.g., concerns about hygiene, discomfort or availability).

Keywords:

• E-scooters
• helmet
• mandatory law
• risky behaviors

Introduction

Since the introduction of shared electrically assisted scooters (e-scooters) in the United States in 2017, their use has exponentially increased globally. This emerging mobility mode is attractive for its convenience (fast travel, traffic avoidance, low maintenance), affordability, enjoyment, and environmental sustainability (Šucha et al. 2023). Thus, e-scooters have become one of the most popular personal mobility vehicles worldwide, particularly in urban areas, replacing short car or walking trips and connecting to public transport addressing gaps in transportation networks (Serra et al. 2021).

The growth in e-scooter use has been accompanied by reports of increased e-scooter-related injuries including frequent head injuries. A review of 29 international studies of e-scooter injuries concluded that the head and/or face was the most common injury site, comprising 38.8% of emergency presentations; ranging from minor injuries (53% of head and/or facial injuries), concussions (36.1%), traumatic brain injuries (22.7%), intracerebral hemorrhages (34.4%), to skull fractures (14.3%) (Rashed et al. 2022). Another review of 34 international studies showed that over 98% of the 5,705 injured riders presenting to local Emergency departments were not wearing helmets and their most frequent mechanism of injury occurred as a result of loss of balance (74%) rather than crashes with motor vehicles (Singh et al. 2022). These falls might be linked to instability from small wheels on uneven pavements (Ma et al. 2021). In Brisbane, Australia, the head/face was the most commonly injured body part in 952 emergency presentations involving e-scooters between November 2018 and June 2021 (Vallmuur et al. 2023). In Western Australia, injured e-scooter riders who used helmets (43%) had significantly less head injuries (Raubenheimer et al. 2023), confirming the findings from an earlier Brisbane cohort (Mitchell et al. 2019).

Most countries do not require helmet wearing for e-scooters and observed wearing rates are generally below 10% (Serra et al. 2021). In Australia, helmet use has been mandatory in all states and territories since the introduction of shared e-scooters. However, despite financial penalties (151 AUD in ACT, 143 AUD in QLD) and provision of helmets on shared e-scooters (see Figure 1), nonuse of helmet remains common. Observational studies in downtown Brisbane in October 2021 found that helmets were not worn by 36.8% of shared and 7.4% of private e-scooter riders (Haworth and Schramm 2023)

Figure 1. Pictures of helmet provision on shared e-scooters in Canberra, Australia.

While low helmet wearing rates have been widely reported in e-scooter studies, relatively little is known about the factors affecting helmet use. In one of the few examinations of the consistency of helmet nonuse (Sievert et al.2023), 39.3% of mostly US e-scooter riders reported never wearing a helmet, 28.1% wore a helmet sometimes and 32.6% reported always wearing a helmet. They reported that riders who were more concerned about collisions with vehicles and pedestrians were more likely to wear helmets, but that concern about safety of the riding environment did not influence helmet use. That study also found that riders who rode less often than monthly were less likely to wear a helmet than more frequent riders. In Brisbane, Australia, helmet nonuse was more commonly observed among riders of shared e-scooters (Haworth et al. 2021) who have been found to ride less often than private e-scooter riders (Lefrancq 2019; Laa and Leth 2020).

It is possible that the factors associated with nonuse of bicycle helmets may also be relevant for e-scooters. For bicycles, mandatory helmet legislation increases helmet use (Karkhaneh et al. 2006) but the findings of an international comparative study of helmet use identified several factors influencing the relationship between legal requirements and helmet use (Ledesma et al. 2019; Valero-Mora et al. 2020). The effectiveness of the law in increasing helmet wearing rates appeared to be greater in countries where riders were aware of the law and law was enforced. For example, self-reported helmet use was 91% in Australia where almost all riders surveyed knew that there was a law but only 45% in Argentina where knowledge and enforcement are low (Ledesma et al. 2019). The survey results also showed that the belief that helmet use is mandatory increases the likelihood of consistently wearing bicycle helmets, regardless of whether the mandatory legislation actually exists (Valero-Mora et al.2020; Siebert et al. 2021). Recent literature has examined if mandatory helmet use deters cyclists from riding (Esmaeilikia et al. 2019), suggesting that the level of rider support for the mandatory law might influence helmet wearing.

Helmet nonuse appears to be linked to involvement in other risky behaviors while riding e-scooters, as has been reported for bicycles (Esmaeilikia et al. 2019). In Berlin, Germany, before helmets became mandatory, no shared e-scooter riders were observed wearing helmets and other risky behaviors were common: Multiple riders per e-scooters (5.1%), prohibited infrastructure usage (28.3%), and travel in the opposite traffic direction (13.1%) (Siebert et al. 2021). In California, US, riders were video-recorded violating traffic rules while not wearing a helmet: 88.9% were traveling opposite to the traffic direction and 94,9% were riding on sidewalks (Todd et al. 2019). A German observational study of 4,514 bicycle and e-scooter riders showed that users who wore less protective equipment (helmet, reflective clothing) engaged in more other risky behaviors (secondary tasks like using phones or headphones and traffic rule violations) (Huemer et al. 2022). In Brisbane, observational studies also found higher helmet nonuse among shared e-scooters with multiple riders. Helmet nonuse (68%) is also listed with acute alcohol (25.1%) and other drugs intoxication (20.3%) as common contributors to e-scooter injuries by Singh et al. (2022) in their review of 34 studies, supported by the findings of Janikian et al. (2024) in their review of 81 studies. More specifically, traumatic brain injuries appear to occur commonly in alcohol intoxicated riders not wearing a helmet (Suominen et al.2022). For bicyclists, helmet nonuse has been significantly associated with alcohol intoxication (Orsi et al. 2014).

The most important barriers to helmet use listed by Norwegian bicyclists were the shortness of their trip and geographic closeness to home, explained by a minimal perceived risk of injury during short duration rides (Lajunen 2016). An earlier American survey also stressed the association of high perceived head injury risk and bicycle helmet use (Finnoff et al. 2001).

While many risky behaviors are more common for males than females, observational studies in Brisbane, Australia (Haworth and Schramm 2023) and in New Jersey, United States (Younes et al. 2023), and an online survey in the UK have found no gender differences in helmet wearing rates by gender. Helmet wearing rates also seemed to be similar for children, adolescents and adults in Brisbane (Haworth and Schramm 2023).

The aim of this study is to understand whether nonuse in the mandatory context is a consistent behavior for an individual or is situationally-influenced, and what are the factors associated with nonuse. These findings can assist in informing measures to improve helmet wearing rates. Australia provides a case study of helmet use by e-scooter riders under mandatory laws that can inform other countries regarding the outcomes of such laws.

Based on the literature presented in the Introduction, it was predicted that:

1. Shared ownership of e-scooters, low frequency of riding, low knowledge and support for the mandatory regulation, low perceived risk of e-scooter use, short trip durations and involvement in other risky behaviors while riding e-scooters would be associated with helmet nonuse.

2. Demographic variables including age, gender, education, and employment would have little influence on helmet nonuse.

Data collection for this study occurred in Canberra, the capital city of Australia. Use of private e-scooters is legal in Canberra. Shared e-scooter schemes commenced in September 2020, with two providers operating (Beam and Neuron). Approximately 4,500 e-scooter trips occurred per day in Canberra between October 2020 and March 2021 (Curijo 2021). E-scooters are permitted to ride on footpaths with a maximum speed of 15 km/h and 25 km/h on bicycle paths or roads. Helmet use is mandatory for both e-scooters and bicycles.

Methods

The Queensland University of Technology Committee on Ethics in Research on Humans gave its approval for the study (approval number 5247).

Participants

Participation was restricted to adults (18 years old or older), living in the Australian Capital Territory or border communities, who had ever ridden an e-scooter in Canberra. While the overall survey comprised riders and non-riders: Only the e-scooter rider data was analyzed in this paper. A total of 374 e-scooter riders responded to the survey but one response was excluded due to missing information.

Participants were asked to indicate if they had worn a helmet during their last trip (the last time you used a e-scooter, did you wear a helmet?) and how often they had ridden without a helmet over the last 30 days (over the last 30 days, how often did you ride an e-scooter without a helmet?). Self-reported helmet use over the last 30 days was chosen instead of use on the last trip as the key dependent variable because of the theoretical interest in whether helmet use is consistent over time. In addition, the small number of riders (55) who reported not wearing a helmet on the last trip was too small to support multivariate analyses. Reported helmet use on the last trip and reported helmet use over the last 30 days were consistent except for 17 riders who reported always wearing a helmet over the last 30 days but not during their last trip. For 14 of these riders, their low riding frequency suggested that they had probably not ridden in the last 30 days, thus they were excluded from the analysis. The remaining three riders who reported using e-scooters frequently, were reclassified as riding without a helmet at least once over the last 30 days. The final sample included 360 riders, with 253 (70.1%) never riding without a helmet, 44 (12.2%) at least once, 29 (8%) sometimes, 12 (3.3%) often, and 22 (6.1%) almost always in the last 30 days. The responses were grouped into two categories: Consistent helmet users (70.1%) and occasional helmet users (29.9%).

Data collection

A 15-min online survey was administered between April and June 2022 and between February and April 2023 using KeySurvey software. Overall, almost 80% of responses were submitted in autumn (March-May, average daily maximum temperature 20 °C), with about 10% in early winter (average daily maximum temperature 12 °C) and about 10% in late summer (average daily maximum temperature 27 °C). Participants were recruited via paid Facebook advertisements, online newspaper announcements, and emails from shared e-scooter companies. Participants could enter a prize draw for an AUD 100 electronic gift card.

Based on previous finding, the survey items addressed:

• Rider characteristics: Demographics (gender, level of education, current occupation, e-scooter ownership (“The last time you used an e-scooter, was it your own or a rented one?”), 12 other risky behaviors while riding e-scooters (riding after consuming alcohol or illegal drugs, speeding in various environments, riding in inaccessible areas, carrying more than one person, red-light running, phone use, use of headphones, lack of lights and reflective gear at night), and frequency of riding e-scooters and bicycles. Given previous surveys and observational studies in Australia have shown helmet wearing rates of more than 95% among private bicycle riders (e.g., Debnath et al. 2016; Haworth et al. 2021), this item was included to identify whether the e-scooter riders were also bicycle riders and thus likely to be accustomed to wearing a helmet.

• Trip characteristics: Duration of the last e-scooter trip.

• Attitude: Perceived risk of using e-scooters.

• Knowledge of the law (“In Canberra, wearing a helmet while riding an e-scooter is?” Mandatory for any user, mandatory for children less than 12, not mandatory but highly recommended) and level of support for the law (“Do you/would you oppose or support mandatory helmet use?” With responses on a Likert scale from 1 = Strongly oppose to 5 = Strongly support).

Data analysis

All statistical analyses were performed using IBM SPSS software (Version 29). Descriptive statistics were calculated for each of the variables as a function of helmet use (consistent or occasional) and as a function of whether the last trip was made using a private or shared e-scooter, as large differences of helmet nonuse were observed (36.8% of shared and 7.4% of private e-scooter riders) in previous Australian studies (Haworth et al. 2021).

Logistic regression was used to identify which of the factors were independently associated with riding without a helmet at least once in the last 30 days or not. The types of risky riding behaviors were highly correlated and the sample size was not sufficiently large to include all the risky behaviors in the model. Therefore, a risky riding scale was created by adding the number of risky behaviors that participants reported having engaged in at least once during the last 30 days. Riding in the dark without reflectors was omitted from the scale because it is a characteristic of the e-scooter, rather than a rider choice. Riding at faster than walking pace was also excluded from the risky riding scale because this behavior was reported by almost all riders. The scale had high internal consistency (Cronbach’s alpha = .90). The risky riding score was treated as a continuous variable. The categorical variables entered in the logistic regression were age group, gender, e-scooter ownership (private or shared), frequency of e-scooter riding, frequency of bicycle riding, trip duration, perceived risk of e-scooter riding, knowledge of the helmet law and support for the helmet law. To ensure sufficient cell sizes, “strongly oppose” and “oppose” were combined in support for the helmet law, and “very unsafe” and “unsafe” were combined in the perceived risk variable.

Results

Descriptive statistics

Among the 360 participants, 75% were male, half were aged between 35 and 54 years old (50.3%), highly educated (62.7% with a bachelor’s degree or postgraduate), and predominantly full-time workers (80.3%). Their last ride was on a shared e-scooter for 67.9% of riders and on a private e-scooter for 31.9% of riders. Shared e-scooter riders were more likely to have ridden without a helmet on the last trip (17.9% vs 9.6%) and in the last 30 days than private e-scooter riders (32.7% vs 23.5%). The sociodemographic characteristics of private and shared e-scooter riders based on their helmet use over the last 30 days are summarized in Table A1 (see online supplement). Age was the only demographic variable significantly associated with helmet use. Non-helmet wearers were significantly younger than helmet wearers for both private (χ²(4) = 12.998, p=.011) and shared (χ² (4) = 24.983, p<.001) e-scooters. Helmet use was not associated with gender (private: χ²(1) = 0.039, p=.844; shared: χ²(1) = 0.319, p=.572). There was no significant association between helmet use and level of education (private: χ²(3) = 0.169, p=.982; shared: χ²(3) = 4.007, p=.261) or helmet use and employment status (private: χ²(1) = 3.613, p=.057; shared: χ²(1) = 0.024, p=.876).

Table 1. Helmet nonuse in the last 30 days as a function of knowledge and support for the law.

	Knowledge of law
	Mandatory	Not mandatory or only mandatory for children	Total
Level of support for law
	Helmet use/ helmet nonuse	Helmet use/ helmet nonuse
Opposed or strongly opposed	27/36 (75%)	14/17 (83%)	41/53 (78%)
Neutral, support or strongly support	50/276 (18%)	16/26 (62%)	66/302 (22%)
Total	77/312 (25%)	30/43 (70%)

Private e-scooter users rode more frequently, but no association was found between their frequency of riding and their helmet use (χ²(5) = 1.542, p=.908). Shared e-scooter users tended to ride less often, mostly a few times a month or less (72.7%), and helmet use was associated with less frequent riding among this group (χ²(5) = 12.478, p=.029). Since no significant difference (χ² (1) = 3.154, p=.076) was found in the type of ownership between consistent and occasional helmet nonusers in our study, the results that follow are presented without separating private and shared e-scooter riders. Frequency of bicycle riding in the last year (Table A2 in online supplement) was not significantly associated with wearing a helmet when riding an e-scooter (χ²(3) = 4.042, p= .257). No significant association was found between duration of the last trip and helmet wearing in the last 30 days (χ²(3) = 6.392, p= .094). When we conducted the analysis for all riders in the survey, there was a trend toward lower helmet use during their last trip when it was shorter (less than 10 min) than when it was longer (over 10 min), but this association felt just short of statistical significance (χ² (3) = 7.602, p= .055). We were concerned that this analysis may have been confounded by potentially higher wearing rates and longer trips for private compared to shared e-scooter riders, so we repeated the analysis separately for the two groups. Our suspicions were confirmed, with the apparent relationship disappearing (private: χ² (3) = 2.373, p= .499; shared: χ² (3) = 4.715, p= .194). During the last 30 days there was no significant association between perceived risk of using e-scooters (χ² (4) = 7.114, p= .130) and helmet wearing (Table A2 in online supplement).

Table 2. Logistic regression results for self-reported helmet nonuse at least once when riding an e-scooter in the last 30 days.

Variables	β	Odds ratio	95% CI
Constant	0.862	2.369
Age group
18–34^a
35–44	−0.699	0.497	0.172–1.434
45 and over	−0.471	0.624	0.266–1.468
Gender
Male^a
Female	0.315	1.370	0.564–3.327
Ownership status
Private^a
Shared	0.443	1.557	0.612–3.965
Trip duration
Less than 5 min^a
5–10 min	−1.019	0.361	0.103–1.269
10–20 min	−1.034	0.355	0.092–1.380
Greater than 20 min	−1.134	0.322	0.076–1.354
Frequency of e-scooter riding
Several times a day^a
Several times a week	−0.637	0.529	0.152–1.836
One to a few times a month	−0.804	0.448	0.167–1.202
Less often or not anymore	−0.734	0.480	0.153–1.501
Frequency of bicycle riding
Never^a
One to a few days a year	0.058	1.060	0.377–2.978
One to a few days a month	0.685	1.983	0.747–5.263
At least once a week	−0.241	0.786	0.254–2.430
Risky riding score	0.640	1.896	1.528–2.352
Perceived risk
Very unsafe or unsafe^a
Neutral	0.077	1.080	0.316–3.690
Safe	−0.420	0.657	0.203–2.128
Very safe	−0.244	0.784	0.215–2.856
Knowledge of the law
Mandatory for all^a
Mandatory for children under 12 or not mandatory but highly recommended	1.299	3.664	1.279–10.492
Support for the law
Strongly oppose or oppose^a
Neutral	−1.816	0.163	0.046–0.572
Support	−2.636	0.072	0.020–0.252
Strongly support	−2.976	0.051	0.015–0.168
Full model vs null model: χ^2(21) = 167.529, p<.001; Nagelkerke R^2 = .591.

^a Reference level.

Helmet wearing was significantly associated with knowledge (χ²(1) = 27.798, p<.001) and support for the law (χ²(3) = 81.878, p<.001). Among non-helmet wearers, 28.0% incorrectly responded that helmets were not mandatory for adults (either that there was no law or that it only applied to children), compared to 5.2% of helmet wearers. Overall, 60.2% of helmet wearers stated that they “strongly supported” the law, compared with 26.2% of non-wearers. Table 1 shows that helmet nonuse was low (18%) among riders who knew that there was a law and were not opposed or strongly opposed to the law, who were the largest sub-group in the study (276 riders). Helmet nonuse was high among the smaller groups of riders who were opposed or strongly opposed to the law (whether they knew that the law was mandatory or not, 75% and 83% nonuse, respectively) and among those who did not know about the law and were not opposed or strongly opposed to it (62% nonuse).

Among the risky behaviors listed, riding faster than walking speed on the footpath was the most common, being reported by 68.7% of helmet-wearers and 97.2% of non-wearers. Nonuse of helmets was significantly associated with each of the other risky behaviors (all p<.001). Table A3 (see online supplement) shows that all the other risky behaviors were significantly correlated at p<.01 except riding at greater than walking pace on the footpath and riding after using illegal drugs (p<.042). The highest correlations were between not wearing a reflective jacket in the dark and not using front and rear lights (r = .683), phone usage for talking and texting (r = .679), and riding on the footpath at over walking speed and over 15 km/h (r = .714). The riders who occasionally wore a helmet (M = 4.95, SD = 3.32) reported significantly more risky behaviors than riders who consistently wore a helmet (M = 1.58, SD = 1.60), t(342) = 12.180, p <.001. Only 5% of riders who reported engaging in less than two types of risky behavior had ridden without a helmet at some time in the last 30 days, which increased to 33.9% of riders who had engaged in two to four types of risky behaviors, and 78.6% of riders with five or more types of risky behaviors.

Logistic regression

Logistic regression was used to identify the independent contribution of factors toward having ridden an e-scooter without a helmet at least once in the last 30 days (Table 2). The overall model was statistically significant when compared to the null model, (χ²(21) = 167.529, p<.001), explaining 59.1% of the variation in helmet use (Nagelkerke R²) and correctly predicted 85.7% of cases. The likelihood of having ridden without a helmet increased with the risky riding score (OR = 1.936, p<.001). Riders who thought that helmets were not mandatory for them (either because they thought there was no law or the law only applied to children under 12), were more likely to have ridden without a helmet (OR = 3.664, p = .016). Riders who supported the law were less likely to have not used a helmet. Compared to riders who strongly opposed or opposed the law, the odds of not having worn a helmet were almost 84% lower for riders who were neutral toward the law (OR = .163, p = .005), and riders who supported the law (OR = .072, p<.001) or strongly supported the law (OR = .051, p<.001) were even less likely to have ridden without a helmet in the last 30 days.

Discussion

This study aimed to better understand whether helmet nonuse in the mandatory context is a consistent behavior for an individual or is situationally-influenced, and what are the factors associated with nonuse. The factors investigated comprised rider demographics (age, gender, level of education, occupation), frequency of e-scooter and bicycle riding, private or shared e-scooter used on last trip, risky behaviors while riding, trip duration, perceived risk, and knowledge and support for the helmet law. The study findings suggest that most nonuse of helmets when mandatory is situational, rather than consistent. Of the 29.1% of riders who reported not wearing a helmet in the last 30 days, 24.4% had not worn a helmet at some time during that period and only 4.8% had consistently not worn a helmet. However, a US study (Sievert et al.2023) found that consistent nonuse was more common than occasional nonuse (39.3% versus 28.1%), suggesting that nonuse may be more consistent where helmet use is not mandatory. There were only 22 consistent non-wearers in this study which prevented a detailed analysis of how the characteristics of consistent and occasional non-wearers differ. However, there were no marked differences in the gender, age, frequency of e-scooter use or whether the last ride was on a private or shared e-scooter between consistent and occasional non-wearers.

The factors shown to be independently associated with helmet nonuse were the number of risky riding behaviors, lack of knowledge, and lack of support for the law. The association between helmet nonuse and other risky behaviors when riding an e-scooter mirrors the findings of earlier observational studies in Germany (Huemer et al.2022), the US (Todd et al.2019) and in Brisbane (Haworth et al.2021) and reports of higher levels of alcohol and drug intoxication among non-helmeted injured e-scooter riders (Singh et al.2022; Suominen et al.2022; Janikian et al.2024). One explanation for this association could be the predominantly young male population who ride e-scooters as their higher tendency for risky behaviors has been shown (Wilson and Daly 1985). Another explanation could be linked to societal attitudes toward e-scooters as they are often perceived as a “harmless toy” with low perceived risk (Useche et al.2022) and therefore little need for protective behaviors. Low levels of enforcement of rules relating to e-scooter riding or a sense of anonymity provided by the absence of vehicle identification plates (Useche et al.2022) might also facilitate engaging in multiple illegal behaviors

Almost 90% of riders knew that the law applied to them and more than 70% supported or strongly supported the law. Overall, 25% of riders who knew the law applied to them reported having ridden without a helmet in the last 30 days. Lack of knowledge and lack of support for the law were both associated with nonuse of helmets. The crosstabulation of these factors in Table 1 suggests that knowledge of the law has little effect on helmet wearing for riders who oppose or strongly oppose the law, but that knowledge of the law is associated with less nonuse of helmets by riders who are neutral or supportive of the law. Opposing the mandatory law is often related to not understanding the purpose of the regulation either by not grasping the risk of head injuries linked to e-scooters or the protective role of helmets in reducing head injuries (Fleiter et al.2014). It can also be attributed to not sharing a prosocial view of helmet wearing to ensure own and others’ safety but rather to viewing helmet wearing as a personal choice and the law as a freedom restriction. For these helmet non-wearers, targeting prosocial considerations for helmet wearing, a better understanding of the risk of potentially severe head injuries associated with e-scooters, and of the protective role of helmets in educational campaigns could inform better-targeted prevention of helmet wearing.

Based on previous research, it was predicted that nonuse of helmets would be associated with riding shared e-scooters, low frequency of riding, shorter trips and a lower perceived risk of riding. The reported rate of helmet nonuse for shared e-scooters appeared to be greater than for private e-scooters (32.7% vs 23.5%) but this factor was no longer significant when adjusted for the other factors in the logistic regression. It is possible that this factor became no longer significant after adjusting for rider age because older riders were more likely to wear helmets and more likely to ride private e-scooters (as found in the bivariate analysis).

In the bivariate analysis, shared e-scooter riders who rode less frequently were less likely to have not worn a helmet, the opposite to the findings reported previously and there was no significant relationship for private e-scooters. The surprising finding may have reflected that older shared e-scooter riders rode less frequently and that these older riders were more likely to wear helmets (as found in the bivariate analysis). Once other factors were adjusted for in the logistic regression, the association between frequency of riding and helmet use disappeared.

The finding in the literature that shorter trips were associated with lower helmet wearing rates for bicycles (Lajunen 2016) was not found for e-scooters in the current study, either in the bivariate analyses or the logistic regression. Other studies have not yet examined this relationship for e-scooters.

It was expected that nonuse of helmets would be associated with lower perceptions of the risk of e-scooter use based on earlier surveys of e-scooter (Sievert et al.2023) and bicycle riders (Lajunen 2016). However, this was not the case in the current study, despite the wide range of reported levels of perceived risk.

Most nonuse of helmets in a mandatory context seems to be situational, rather than consistent. Many of the factors associated with nonuse of helmets for e-scooters are similar to those reported for bicycles. Nonuse of helmets appears to be one of a number of risky behaviors performed by riders, rather than being primarily an outcome that is specific to factors associated with helmets (e.g., concerns about hygiene, discomfort or availability). Knowledge of the helmet law appears to have little effect on helmet wearing for e-scooter riders who oppose or strongly oppose the law, but lack of knowledge of the law is associated with less nonuse of helmets by riders who are neutral or supportive of mandating helmet use. Thus, increasing support for the helmet law through making riders more aware of the risk of head injury should increase helmet use.

Strengths and limitations

This study is one of the first studies to examine how knowledge and attitudes may influence helmet nonuse by e-scooter riders. Most previous studies observed riders which could measure some characteristics associated with helmet wearing nonuse such as gender, approximate age and riding behaviors but were limited in their ability to identify the factors underlying helmet wearing or to understand whether nonuse was a consistent or occasional behavior. Observations are often restricted to a small number of locations which may not be representative of the entire riding environment. While using a rider survey has advantages over observations, the limitations of a survey approach should be acknowledged. It is possible that a survey approach resulted in a sample with different characteristics than an observed sample. There is no published observational data from Canberra to compare with the survey results. However, survey and observational studies in Brisbane, Australia found similar patterns (Haworth et al.2021, 2024). However, the survey results showed a slightly higher proportion of men (77% versus 63%), private e-scooters (50% versus 42%) and higher helmet wearing rates for shared e-scooters on the last trip (88% versus 59%) but not for private e-scooters (Haworth and Schramm 2023). The latter finding suggests that this may reflect difference in the samples, rather than over-estimation of self-reported helmet wearing because of inaccurate memory or desirability bias.

Some variables, shown to be associated with helmet nonuse, were missing in our survey and could be investigated such as the riding location (helmets were observed to be less worn on footpaths than on roads in Brisbane; Haworth and Schramm 2023), perceived inconvenience of helmets including discomfort with hot temperatures, impact on vision, concern regarding appearance, the price of private helmets, hygiene for shared helmets (Serra et al.2021), and frequency of helmet use when riding other transport modes. As most survey responses were submitted during the autumn season in Canberra which has a comfortable temperature, changes in helmet discomfort due to seasonal effects was not considered as a relevant factor in our analysis. To better understand the motivations of helmet nonusers and better inform variables to be targeted in educational campaigns, variables related to perceived enforcement, perceived risk of head injuries, perceived protective role of helmets, and prosocial consideration of helmet wearing could be investigated in further studies. Items from the bicycle helmet use questionnaire deriving from the health belief model (Quine et al.1998), that could be adapted to e-scooters include: Perceived vulnerability to head injuries (in case of an accident, would the e-scooter be fast enough to hurt my head seriously?), perceived severity of head injury, and perceived helmet protection). Items related to perceived enforcement could include: Perceived fear of punishments (being fined), social sanction (embarrassment of being caught, peer pressure), punishment avoidance (previous experience or vicarious exposure to other being punished) (Hasan et al.2023).

Acknowledgements

We would also like to thank our participants for their time and interest.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Access to the data can be requested from the corresponding author and will be subject to agreement by the funder and conditions of the ethical approval for this project.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

This research was funded by the ACT Government Road Safety Fund Grants Program. The preparation of this paper was supported by a Vacation Research Experience Scheme scholarship from the Queensland University of Technology (QUT) Faculty of Health to the first author.