Alcohol consumption among motorcyclists and bicyclists is of increasing concern given the rising prevalence of registered motorcyclists and growth in the personal and shared bicycle industry (National Association of City Transportation Officials 2020; U.S. Department of Transportation 2023; National Association of City Transportation Officials 2022; Teoh 2023). Since at least 2008, more than 38% of motorcycle riders and 20% of pedalcyclists who died in traffic crashes were under the influence of alcohol, and motorcyclists consistently show the highest prevalence of alcohol use during fatal traffic crashes compared to any other motor vehicle driver (Vargos et al. 2015; National Center for Statistics and Analysis 2021a; National Center for Statistics and Analysis 2021b). As the availability and demand for motorcycles and bicycles rises across US cities and deaths continue to rise among these vulnerable road users, national, city, and state government officials have called for greater policy and research attention to help protect their safety (Rimer 2020; Jiao and Bai 2020; National Transportation Safety Board 2022a; Herron 2023; Council Member Mackenzie Kelly, District 6 2023; McLeod 2024).

Epidemiological surveillance data for alcohol involvement in pedalcycle and motorcycle injuries, especially for crashes that did not involve a motor vehicle, are lacking. We found only two studies examining alcohol involvement and motorcycle injuries which both used National Trauma Data Bank (NTDB) data (Ahmed et al. 2020a; Brockhus et al. 2024). Brockhus et al. (2024) included all patients who were injured in a motorcycle crash and found 18% tested positive for alcohol use whereas Ahmed et al. (2020) examined only patients that were tested for alcohol, and showed that32% tested positive for blood alcohol concentration (BAC) (Ahmed et al. 2020a; Brockhus et al. 2024). To our knowledge, the only national-level sample to examine alcohol use and bicycle injuries found that among the 270,571 persons with bicycle injuries who presented to an emergency department (ED) in 2019, alcohol use was implicated in 7% of these injuries (Namiri et al. 2020).

A primary reason for the sparse evidence on the prevalence of alcohol involvement in pedalcycle and motorcycle injuries is due to methodological limitations across injury surveillance systems. An essential first step to developing effective interventions and policies to reduce the risk of alcohol-related vulnerable road user injuries and deaths is to have robust and reliable surveillance and epidemiological data to characterize the burden of these injuries and identify high-risk populations.

There at least three sources of surveillance data. The Fatality Analysis Reporting System (FARS) (National Highway Traffic Safety Administration) and National Electronic Injury Surveillance System (NEISS) (U.S. Consumer Product Safety Commission) are two public health surveillance tools that can be used for vulnerable road user fatal and non-fatal injury surveillance. FARS only collects data on motor vehicle-involved traffic crashes (National Highway Traffic Safety Administration). NEISS is a probability sample of 96 US hospitals that collect data on consumer product-related injuries and deaths resulting in an ED discharge or admission and excludes motor vehicles (e.g., motorcycles) as a product type (U.S. Consumer Product Safety Commission). These systems also use varying data coding standards for vulnerable road users. FARS uses the Model Minimum Uniform Crash Criteria (MMUCC) guidelines based on the American National Standard Institute’s (ANSI) Manual on Classification of Motor Vehicle Traffic Crashes (National Highway Traffic Safety Administration 2017; American National Standard), while NEISS uses consumer product codes that the US Consumer Product Safety Commission (CPSC) has jurisdiction over (U.S. Consumer Product Safety Commission 2019). This particular issue has risen to national attention, as highlighted by a recent report from the National Transportation Safety Board (NTSB), which cited data coding standardization across injury surveillance systems as the primary reason for the poor understanding of emerging micromobility devices (e.g., e-scooter and e-bike) associated injuries and fatalities (National Transportation Safety Board 2022b). The National Emergency Medical Services Information System (NEMSIS) public-use research dataset, which uses Emergency Medical Services (EMS) clinical and administrative data, may be a means of surveillance data for measuring alcohol involvement in vulnerable road user injuries, but no studies have used NEMSIS to describe these patterns.

The primary objective of this study was to determine the role of alcohol involvement among fatally and non-fatally injured pedalcycle and motorcycle riders and to compare these findings across the 2019 NEMSIS, NEISS, and FARS national public-use datasets. This study draws conclusions on the utility and validity as sources of information on this increasingly important public health issue.

Table 1 shows the descriptives characteristics for adult (≥21 years) pedalcyclist fatal and non-fatal injuries across the 2019 NEMSIS, NEISS, and FARS datasets by alcohol involvement. Among US adults, we identified 26,295 pedalcyclist injuries that required an EMS response in the 2019 NEMSIS dataset, and alcohol involvement was noted in 10.2% of these injuries (n=2,690). Using the NEISS dataset, there were an estimated 253,186 (95% CI 191,580 to 314,802) pedalcycle injuries with 7.3% (95% CI 5.8 to 9.0) involving alcohol. FARS data revealed a total of 770 pedalcyclists fatalities from traffic crashes, and 27% of these fatalities indicated alcohol involvement (BAC of .01 g/dL or higher). For pedalcyclist injuries as reported by both NEMSIS and NEISS datasets, across each age category, adults ages 45–64 had the highest proportion of alcohol-related injuries (13% and 11%, respectively) compared to all other age groups. For fatally injured pedalcyclists in traffic crashes, the 25–34 age group had the highest prevalence of alcohol involvement (36%), closely followed by the 45–65 age group (33%). Across all three datasets, men had a higher prevalence of alcohol involvement compared to women for pedalcyclist injuries resulting in an EMS response (11.8% vs 5.3%), when treated at an ED (8.7%, 95% CI 7.0 to 11.0 vs 3.1%, 95% CI 1.9 to 5.0), or for fatalities following a crash (27% vs 23%).

Table 2 show the descriptives characteristics for adult (≥21 years) motorcycle rider fatal and non-fatal injuries across the 2019 NEMSIS, NEISS, and FARS datasets by alcohol involvement. In total, there were 50,122 motorcycle rider injuries resulting in an EMS response in the 2019 NEMSIS dataset, and 8.5% of these injuries were reported to involve alcohol (n=4,643). This was slightly larger than the 6.1% (95% CI 4.6 to 8.2) of moped/power-assisted cycle/minibike/two-wheeled, powered, off-road vehicle injuries to persons treated in the ED among the estimated 46,398 (95% CI 37,046 to 55,749) total injuries in the NEISS dataset. FARS data revealed that among the 4,534 motorcycle rider fatalities from traffic crashes, 42% involved alcohol (BAC of .01 g/dL or higher), and 34% were alcohol impaired (BAC of .08 g/dL or higher). Early middle-aged adults (35–44 age groups) had the highest proportion of alcohol involvement in motorcycle rider injuries resulting in EMS response (9.8%) and motorcycle rider fatalities in traffic crashes (50%) compared to all other age groups as reported by the NEMSIS and FARS datasets. Early and late-middle aged adults (35–44 and 45–64 age groups) had the highest proportions for alcohol involvement in moped/power-assisted cycle/minibike/two-wheeled, powered, off-road vehicle injuries for those admitted to an ED (9.5% and 10.3%, respectively). The proportions were all higher for males compared to females for alcohol involvement in motorcycle rider injuries with EMS response (9.1% vs 6.0%), moped/power-assisted cycle/minibike/two-wheeled, powered, off-road vehicle injuries with ED visits (6.9%, 95% CI 5.1 to 9.3 vs 1.8%, 95% CI 0.34 to 9.1), and motorcycle rider fatalities from automobile crashes (42% vs 34%).

This is the first study to use the NEMSIS national public-use dataset for describing the burden of motorcycle and pedalcycle injuries from transport crashes involving both motorized and non-motorized vehicles, and to compare findings with two other national public-use datasets by alcohol consumption status. Overall, about 10% and 8.5% of EMS responses for injured pedalcyclists and motorcycle riders, respectively, involved alcohol. These were larger estimates compared to the 7.3% of pedalcyclist and 6.1% of moped/power-assisted cycle/minibike/two-wheeled, powered, off-road vehicle injuries involving alcohol among adult patients treated at an ED. However, the proportions for pedalcycle and motorcycle riders who were fatally injured in motor vehicle traffic crashes that involved alcohol were much greater (27% and 42%, respectively). Pedalcycle and motorcycle fatal and non-fatal injuries among middle aged adults and men were more likely to involve alcohol compared to older adults and early aged adults and women.

Differences in estimates across these datasets may be explained by sources of bias and error in the sample populations and measures for pedalcycles, motorcycles, and alcohol across these national public-use datasets (Figure 1). Selection bias is a limitation of these findings as FARS only collects BAC test results for motor vehicle crashes where there was a fatal injury, compared to NEISS and NEMSIS which primarily collect data on non-fatal injuries from all types of crashes (motor vehicle involved and not) and data are reported by clinicians. ED and EMS clinicians also do not routinely screen all patients for alcohol-intoxication (Chezem 2004). The small estimates we found for alcohol involvement in motorcycle and pedalcycle injuries (between 6.1% - 10%) suggest that the limited alcohol screening results in a small subgroup being reported for alcohol involvement, who are likely the most severely injured patients as alcohol use by bicyclists and motorists is associated with more severe injury and death (Sethi et al. 2016; Ahmed et al. 2020b).

The measures for assessing acute alcohol involvement also likely contribute to the selection bias in these findings and further induce measurement error. FARS collects BAC results from police administered breath-tests or toxicology reports from the Medical Examiner’s office, whereas NEISS data includes the medical record report of alcohol consumption prior to or during the incident, which often include a BAC test result for the patient. However, NEMSIS uses the EMS clinician’s evaluation of alcohol involvement at the scene, which is a subjective measure. The alcohol involvement measure in NEMSIS is likely a more sensitive test, as shown in the larger proportion estimates were found compared to the estimates using NEISS data. However, the objective measures for acute alcohol use reported in FARS and NEISS are more specific, which likely results in a higher proportion of true cases. Denial of coverage to patients, time constraints, safety, patient condition, and training are all potential barriers to routine screening for alcohol intoxication or use of breathalyzers to capture objective data in injury incidents that result in EMS response or treatment at an ED (Chezem 2004). To determine the true prevalence of alcohol involvement in vulnerable road user crashes, future work is needed to identify valid and feasible screening tools for acute alcohol involvement.

The various coding criteria for “motorcycles” and “pedalcycles” also result in measurement errors in these findings and may explain differences in estimates across datasets. For example, ANSI defines a motorcycle as a “motorcycle, moped, motor scooter, minibike, or pocket bike” compared to NEMSIS using ICD10 codes (V20-V29) codes, which include “mopeds, motor scooters, motorcycles, motorized bicycles, and speed-limited motor-drive cycles” transportation modes. While NEISS does not contain a specific motorcycle product code, “mopeds/power assisted cycles”, “minibikes”, and “two-wheeled powered off-road cycles” are collected and are listed under the ANSI or ICD10 umbrella terms for motorcycles. None of the 2019 datasets had specific codes for e-bikes and e-scooters so we did not examine these modes within our analyses. However, the 2023 NEISS coding manual now includes product codes for both e-bikes and e-scooters, as recommended by the NTSB (National Transportation Safety Board 2022b). Likewise, the 2021 and 2023 ICD10 codes were updated to include separate coding categories for injuries associated with e-bikes and e-scooters. Future research should determine the validity of ICD10 coding standards for micromobility modes to ensure accurate injury surveillance data is being collected.

Similar analyses of 2019 NEISS data using the same bicycle product codes (codes 5033 and 5040) also found that alcohol involvement in pedalcyclist injuries comprised 7.0% of ED visits among adult patients aged 18 and older (Namiri et al. 2020). Middle-aged adults and men were also similarly more likely to sustain an injury that involved alcohol compared to other age groups and women. In fact, our study found that regardless of the data source, men had between 3% to 8% points higher proportions of alcohol involvement among pedalcyclist and motorcyclist injuries compared to women, which may be explained by the higher prevalence of alcohol consumption among men (White 2020). However, our estimates for alcohol use and motorcyclist injuries were much lower than found by one study using the NTDB to examine adult patients (18 years and older) injured in motorcycle crashes, which found a positive BAC test in about one-third of patients who were tested for alcohol use (Ahmed et al. 2020b). This estimate fell between the estimates from FARS and NEISS/NEMSIS, which may again be an indicator of selection bias, as patients included in the NTDB are only those admitted to a trauma center with serious injuries.

These analyses are subject to a number of important limitations. First, there are likely a large number of additional motorcycle and pedalcycle crashes involving alcohol where the pedalcyclist or motorcycle rider does not seek EMS or hospital care, perhaps due to concerns surrounding overconsumption. This likely has resulted in underestimations within the NEISS and NEMSIS findingsSecond, we onlyexamined 2019 data due to the substantial impact on transportation patterns during the COVID-19 pandemic and that alcohol involvement data were only included in the NEISS dataset as of 2019, yet this will be important to examine in future research(U.S. Consumer Product Safety Commission 2019; Padmanabhan et al. 2021; Wada et al. 2023). We also did not include e-bikes or e-scooters because the included datasets did not yet have variables/codes for these vulnerable road users. As micromobility shared systems (e.g., docked and dockless bike shared systems) are being promoted by US cities and private companies (National Association of City Transportation Officials 2022; Council Member Mackenzie Kelly, District 6 2023; Transportation Alternatives), typically in the same or adjacent areas where alcohol-serving venues and nightlife districts are being developed as part of urban economic development initiatives (Caulfield 2017; New York City Mayor’s Office of Media and Entertainment and City of New York 2019; Alcorn and Jiao 2023; Dean and Zuniga-Garcia 2023), studies are needed to understand how these initiatives will impact alcohol-related pedalcyclist and motorcyclist injuries. While existing research has shown that more alcohol establishments are associated with alcohol-related driving crashes (Saeed et al. 2020) and pedestrian injuries (DiMaggio et al. 2016; Nesoff et al. 2018), the associations with injuries to bicyclists, motorcyclists, and other emerging modes of transportation have yet to be understood.