Authors #

Dr Glen Koorey, ViaStrada

Reviewers  #

Associate Professor Marilyn Johnson, University of New South Wales

Dr Widyarini Weningtyas, Bandung Institute of Technology

Tom Whyte, Transport for New South Wales

1. Overview of the involvement of cyclists in fatal and injury crashes #

Globally there are a wide range of people who use bicycles and other two or three-wheeled pedal devices; these include children through to older adults. However, their relative vulnerability results in a higher-than-average risk of injury or death from traffic collisions.

Several countries have developed more cycle-friendly environments to address this risk (including various measures described below), including the Netherlands, Denmark, and Sweden. Typically, this has resulted in higher proportions of people taking up cycling in these places, and often with a relatively equal gender split. However, less cycle-friendly countries are likely to have proportionately more men than women cycling, often linked with risk perception. Historically, trends show decreasing use by older adults as physical abilities decline, but the growing uptake of e-bikes has significantly changed that pattern, albeit not always with positive safety outcomes due to physical / perceptual / cognitive decline and increasing frailty (Boufous & Olivier, 2022).

In recent years in Australia there have typically been 35-40 cycling road fatalities a year and about 8000 hospitalisations a year (Department of Infrastructure, Transport, Regional Development, Communications, Sport and the Arts [DITRDCSA], 2024). Roughly 80% of these hospitalisations were males, and about 50% of hospitalisations didn’t involve another vehicle. In New Zealand, typically 10-15 cycling road fatalities are reported each year and about 1600 hospitalisations (Injury Prevention Research Unit, n.d.). Roughly 75% of those hospitalisations were males, and about 90% of hospitalisations didn’t involve another vehicle.

Cyclists tend to have a higher risk of crashes at intersections than on mid-block locations. However, this can be very location-specific with regards to the cycle facilities provided at both intersections and mid-block sections.

Typically, there is a relatively high level of under-reporting in Police data of cycling injury crashes, in comparison to hospital admission data (Koorey et al., 2023, Munster et al., 2001). That level of under-reporting is even more notable for “cycle-only” crashes not involving a motor vehicle, which are usually not captured by conventional Police crash data (typically only ~15% of Police-reported crashes do not feature a motor vehicle, including loss of control or collisions with a pedestrian, animal or object).

Although there is also a growing demand for and use of other forms of micro-mobility, such as e-scooters and e-skateboards, this overview focuses on safety risks to people who use cycles.

1.1. Cycle crash injury severity #

Cycles offer very little physical protection to the cyclist (by comparison with the car). Any crash, and in particular those involving collision with motor vehicles are likely to cause injury. The most prevalent injuries that cyclists sustain are to the legs – typically the height zone of car impacts; and the head – as the injured cyclist plunges to the ground or impacts on the vehicle. While most cyclist deaths are due to collisions with motor vehicle, the majority of injuries are not.

The risk of injury in a cycle crash is affected by the relative speed between the bicycle and other vehicle or stationary object, the unforgiving nature of the road or other objects struck, and the protection offered by clothing or a cycle helmet. The established relationship between impact speed and risk of death for pedestrians (e.g. Scott & Mackie, 2014) is thought to also be valid for cyclists. At 30 km/h there is about a 5% risk of death; at 70 km/h the chance of surviving is only about 5%. As a result, there is generally a much higher proportion of cycle crashes on rural roads being deaths and serious injuries (albeit from smaller numbers) than on urban ones.

According to the frailty bias, older riders are more likely to be severely injured or die in cycling crashes because of their age-related declines in tolerance to biomechanic forces (Oxley et al., 2004).

1.2. Overview of errors and risky behaviours that contribute to crash involvement #

There are several types of errors/behaviours that people make in crashes involving cyclists:

• As cyclists: head turn errors; balance or control errors; speeding in congested situations; rule-breaking (e.g. riding on footpaths, contra-flow cycling, failing to stop or give way, no lights at night)
• As drivers (more likely to be at-fault): failing to check/see when changing lanes, merging, or turning; limited reaction time or loss of control due to speeding; passing too closely
• As pedestrians or other path users: not aware of or noticing cycles on paths/roads; sudden movements on a path.

2. Measures to address cycling safety #

The following sub-sections describe some of the typical measures and interventions that can be applied to target safer cycling. While they cover a range of different approaches, ideally a good cycling safety strategy should be multi-disciplinary in nature and involve the complementary use of several different interventions.

2.1. Crash prevention measures targeting cyclists  #

There are several physical infrastructure treatments that can be introduced to reduce the likelihood of harm to people cycling. The potential measures include:

• Provision of cycle lanes and paths
  • The choice of most appropriate cycle facilities is typically dependent on a combination of both traffic volumes and traffic speeds. On-road or shared facilities are usually better suited to lower volume, lower speed streets while separated or protected facilities work better along busier and faster roads. Koorey & Parsons (2016) found that the installation of several painted cycle lanes in Christchurch, NZ, resulted in an average 23% reduction in cycle crash rates. The New Zealand Transport Agency Crash Estimation Compendium ([NZTA], 2025) found slightly lesser effects, with narrow (<1.4m) cycle lanes producing a 10% crash reduction factor and wider cycle lanes producing a 20% reduction.
  • The safety effects of separated cycle paths and shared paths are less clear. The NZTA Crash Estimation Compendium (2025) suggests that the increased risks at intersections and side accesses may cancel out the benefits that occur along mid-block sections. The TOI Road Safety Handbook (Transportøkonomisk institutt, 2025) found that roads with cycle paths have, on average, fewer cycle crashes on mid-block sections than mixed traffic, but the effects are very uncertain. Notably, at intersections two-way separated cycle paths can be less safe than one-way cycle paths, presumably due to lack of driver expectation.
• Intersection treatments (colour surface, advanced stop line)
  • The NZTA Crash Estimation Compendium (2025) estimated a 35% crash reduction factor from the provision of advanced stop boxes of adequate depth. Koorey & Mangundu (2010) investigated the effect of coloured surfacing on motor vehicle encroachment into advanced stop boxes (ASBs) or advanced stop lines (ASLs). Drivers were much less likely to encroach on coloured cycle spaces in comparison to uncoloured ones, particularly ASLs.
• Road speed management and traffic calming
  • Boufous et al. (2012) show that in Victoria, Australia the risk of serious injury is 28% higher in sparsely populated areas with higher traffic speeds than in densely populated areas. Similarly, Cripton et al. (2015) found that in Toronto and Vancouver, Canada, the risk of serious injuries in cycle crashes increases by about 27% for every 10km/h increase in average speed.
  • Raised cycle crossings are another form of providing speed management along roads. Schepers et al. (2011) found that raised cycle crossings with a right-of-way for cyclists reduce the number of cycle crashes by ~50%.

2.2. Injury prevention/mitigation measures #

As noted earlier, the impact speed of a collision with a cyclist can greatly affect the relative risk of injury or death. However, some other protective measures can also affect the likely severity as a result of a cycle crash. These include:

• Bicycle helmets
  • Looking across several studies, Høye (2018a) found that helmets often reduce head injuries by around 50% but are less effective against facial or neck injuries, and will have no effect on serious trauma to other parts of the body. It has been speculated that some riders wearing helmets may feel more protected and end up riding less carefully, thus negating the protective effect. However, Høye (2017) found that cyclists who wear helmets have, on average, less serious accidents than cyclists who do not wear helmets, and are less likely to engage in risky behaviours while riding.
  • The TOI Road Safety Handbook (Transportøkonomisk institutt, 2025) found that mandatory helmet laws led to a roughly 20% reduction in injuries to those cycling, typically due to an increase in the proportion of people wearing helmets. Mandatory laws often lead to an initial decline in the amount of cycling (typically by about 20%), reducing the associated health benefits; however, this decrease is often relatively short-lived (Høye, 2018b).
• Under-run guards for heavy vehicles
  • These measures can help prevent cyclists from being crushed underneath heavy vehicles (that are often turning at the time); such barriers can be located either at the front or sides of vehicles, with side guards often placed between wheels.
  • Side under-run barriers are most effective for reducing serious or fatal injuries to cyclists near trucks, especially when overtaking. Cookson & Knight (2010) found that mandatory under-run barriers in the UK had led to a roughly 60% reduction in DSIs to cyclists being overtaken. The effect is less clear for turning trucks versus through-cyclists.

2.3. Exposure control measures targeting cyclists #

The greatest source of danger for people cycling is from motor traffic; therefore, any initiatives to reduce exposure to traffic are likely to reduce the overall risk. Some of this can be using infrastructure measures to physically protect riders from adjacent traffic (discussed above), but other measures include:

• Traffic volume reduction:
  • In London, several suburbs have been converted in “mini-Holland” neighbourhoods by means of Dutch-style cycling infrastructure such as traffic calming, segregated bike lanes, safety measures at junctions, and blocking residential streets to car traffic. These aimed to encourage modal shift from car to bike for short journeys.
  • The TOI Road Safety Handbook (Transportøkonomisk institutt, 2025) found that cyclists have a lower crash risk on local low-speed streets or “bicycle boulevards” (typically about 50-60% less) than on parallel main roads. Many of these roads feature in-lane “sharrow” markings, and these were found to produce about a 65% reduction in cycle crashes, largely due to cyclists keeping a greater distance from parked cars, and motor vehicles on average keeping a greater distance from cyclists when overtaking.

2.4. Behavioural interventions targeting cyclists #

Both cyclists and motorists can be influenced in their interactions with each other by means of various perceptual measures or promotional initiatives applied. These include:

• Lights/reflectors and high-visibility (“hi-vis”) gear
  • The TOI Road Safety Handbook (2025) found that overall, it is most likely that bicycle lights reduce involvement in collisions and that the effect is greater in darkness than in daylight and when the cyclist is not also wearing a yellow or reflective vest. In developing policy around the issues, Cycling Action Network (2013) found generally inconclusive evidence regarding the effectiveness of hi-vis clothing. From a study of New Zealand cycle fatalities, Koorey (2014) noted that the proportion of drivers not noticing a cyclist prior to a fatal crash was not significantly different whether they were wearing hi-vis clothing or not.
  • Interestingly, Wood et al. (2009) noted that moving pedal reflectors and retro-reflective strips attached to knees/ankles were often more effective than fixed lights/reflectors.
• “Safety in Numbers” effects
  • Safety-in-numbers is the tendency for cycle crashes to increase at a lesser rate than in proportion to cycle volumes, usually due to other road users becoming more aware of them. For example, in the meta-analysis by Elvik & Goel (2019), a doubling of the number of cyclists resulted in only an average 32% increase in the number of bicycle-motor vehicle collisions.
  • However, for crashes between cyclists and heavy vehicles, no notable safety-in-numbers effect was found by Kaplan & Prato (2015), i.e. the number of such crashes increased roughly proportionally to cycle traffic.
• Cycle skills training
  • Teyhan et al. (2016) investigated the impact of cycle proficiency training on cycle-related behaviours and crashes in adolescence. While cycle training led to improved cycling behaviours, they found no strong effect of its impact on crashes. Other studies on young cyclist training have found similar effects or even worse safety outcomes, possibly due to inadvertent encouragement of risk-taking or of cycling with inadequate supervision.
  • Keppner et al. (2023) undertook a randomised trial of cycle training for older (≥65-year) adults. Compared with a control group, participants in the training trial were found to make fewer errors in a subsequent cycling course.
• Cycling safety and “share the road” campaigns
  • Promotional campaigns highlighting the dangers of poor cycling and encouraging courteous road behaviour can either be disseminated via various media formats (TV, radio, print) or via in-person campaigns. The relative effectiveness is often mixed though.

3. References #

Boufous, S., de Rome, L., Senserrick, T., & Ivers, R. (2012). Risk factors for severe injury in cyclists involved in traffic crashes in Victoria, Australia. Accident Analysis & Prevention, 49, 404-409. https://doi.org/10.1016/j.aap.2012.03.011

Boufous, S., & Olivier, J. (2022). Age, crash type and the changing patterns of cycling fatalities in Australia between 1991 and 2022. Injury Prevention, 30(2), 167-173. https://injuryprevention.bmj.com/content/30/2/167

Cycling Action Network. (2013). High-visibility clothing – Policy statement. https://web.archive.org/web/20220201211751/https:/can.org.nz/canpolicy/high-visibility-clothing-0

Cookson, R., & Knight, I. (2010). Sideguards on heavy goods vehicles: Assessing the effects on pedal cyclists injured by trucks overtaking or turning left ( Report No. PPR514). Transport Research Laboratory. https://www.trl.co.uk/uploads/trl/documents/PPR514_secure.pdf

Cripton, P. A., Shen, H., Brubacher, J. R., Chipman, M., Friedman, S. M., Harris, M. A., Winters, M., Reynolds, C. C. O., & Teschke, K. (2015). Severity of urban cycling injuries and the relationship to personal, trip, route and crash characteristics: Analyses using four severity metrics. BMJ Open, 5(1), e006654. https://bmjopen.bmj.com/content/5/1/e006654

Department of Infrastructure, Transport, Regional Development, Communications, Sport and the Arts. (2024). Road crashes involving cyclists. Road Safety Data Hub. Australian Government. https://datahub.roadsafety.gov.au/safe-systems/safe-vehicles/road-crashes-involving-cyclists

Elvik, R., & Goel, R. (2019). Safety-in-numbers: An updated meta-analysis of estimates. Accident Analysis & Prevention, 129, 136-147. https://www.sciencedirect.com/science/article/pii/S0001457519303641

Høye, A. (2017). Road safety for cyclists (Report No. 1597/2017). Institute of Transport Economics (TOI). https://www.toi.no/publications/road-safety-for-cyclists-article34654-29.html 

Høye, A. (2018a). Bicycle helmets – To wear or not to wear? A meta-analyses of the effects of bicycle helmets on injuries. Accident Analysis & Prevention, 117, 85-97. https://doi.org/10.1016/j.aap.2018.03.026

Høye, A. (2018b). Recommend or mandate? A systematic review and meta-analysis of the effects of mandatory bicycle helmet legislation. Accident Analysis & Prevention, 120, 239-249. https://doi.org/10.1016/j.aap.2018.08.001

Kaplan, S., & Prato, C. G. (2015). A spatial analysis of land use and network effects on frequency and severity of cyclist-motorist crashes in the Copenhagen region. Traffic Injury Prevention, 16, 724-731.

Keppner, B., Krumpoch, S., Kob, R., Rappl, A., Sieber, C. C., Freiberger, E., & Siebentritt, H. M. (2023). Safer cycling in older age (SiFAr): effects of a multi-component cycle training. a randomized controlled trial. BMC Geriatrics, 7(23), 131. https://pubmed.ncbi.nlm.nih.gov/36882759/

Koorey, G. (2014, March 23-26). Investigating common patterns in New Zealand cycling fatalities [Conference Paper]. IPENZ Transportation Group Conference (IPENZTG 2014), Wellington, New Zealand. http://hdl.handle.net/10092/9718

Koorey, G., Dioni, G., & Sim, P. (2023, March). How do we measure harm in transport? (TG2023) [Conference Paper]. NZ Transportation Conference, Tauranga NZ. https://viastrada.nz/pub/2023/measure-trpt-harm

Koorey, G., & Mangundu, A. (2010, January 10-14). Effects on motor vehicle behavior of color and width of bicycle facilities at signalized intersections [Meeting Presentation]. 89th Transportation Research Board Annual Meeting, Washington, DC, USA. http://hdl.handle.net/10092/3891

Koorey, G., & Parsons, J. (2016, January). The effect of cycle lanes on cycling numbers and safety [Meeting Presentation]. 95th Transportation Research Board (TRB) Annual Meeting, Washington DC, USA. ViaStrada. https://viastrada.nz/pub/effect-cycle-lanes-cycling-numbers-and-safety

Munster, A., Koorey, G., & Walton, D. (2001). Role of road features in cycle-only crashes in New Zealand (Report No. 211). Transfund New Zealand. https://nzta.govt.nz/resources/research/reports/211

NZ Transport Agency Waka Kotahi. (2025). Crash estimation compendium: New Zealand crash risk factors guidelines. https://nzta.govt.nz/resources/crash-estimation-compendium

Injury Prevention Research Unit. (n.d.). NZ Injury Query System. University of Otago. https://psm-dm.otago.ac.nz/niqs/index.php

Oxley, J., Corben, B., Fildes, B., O’Hare, M., & Rothengatter, T. (2004). Older vulnerable road users – Measures to reduce crash and injury risk (Report No. 218). Monash University Accident Research Centre. https://www.monash.edu/muarc/archive/our-publications/reports/muarc218

Schepers, P., Kroeze, P., Sweers, W., & Wüst, J. (2011). Road factors and bicycle-motor vehicle crashes at unsignalised priority intersections. Accident Analysis & Prevention, 43(3), 853-861. https://pubmed.ncbi.nlm.nih.gov/21376876/

Scott, J., & Mackie, H. (2014). Speed and injury risk curves: Analysis of evidence and consideration for updated curves. https://www.bikeauckland.org.nz/wp-content/uploads/2018/09/Mackie-Research-Report_Speed-vs-injury-risk.pdf

Teyhan, A., Cornish, R., Boyd, A., Joshi, M. S. & Macleod, J. (2016). The impact of cycle proficiency training on cycle-related behaviours and accidents in adolescence: findings from ALSPAC, a UK longitudinal cohort. BMC Public Health, 16, 469. https://pmc.ncbi.nlm.nih.gov/articles/PMC4899925/

Transportøkonomisk institutt. (2025). Road Safety Handbook. https://www.tshandbok.no/

Wood, J. M., Tyrrell, R. A., Marszalek, R., Lacherez, P., & Carberry, T. (2009). Drivers’ and cyclists’ experiences of sharing the road: Incidents, attitudes and perceptions of visibility. Accident Analysis & Prevention, 41(4), 772-776. https://doi.org/10.1016/j.aap.2009.03.014