“Helmet Law as Cycling Deterrent” Hypothesis

There has been much recent debate in Australia regarding mandatory bicycle helmet legislation (MHL). This issue has flared up recently with recommendations from the Queensland parliamentaryreport A new direction for cycling in Queensland. One of the more contentious issues in this document is Recommendation 15 which states

The Committee recommends that the Minister for Transport and Main Roads:

  • introduce a 24 month trial which exempts cyclists aged 16 years and over from the mandatory helmet road rule when riding in parks, on footpaths and shared/cycle paths and on roads with a speed limit of 60 km/hr or less and 
  • develop an evaluation strategy for the trial which includes baseline measurements and data collection (for example through the CityCycle Scheme) so that an assessment can be made which measures the effect and proves any benefits.

Regarding this recommendation, Queensland Transport Minister Scott Emerson has stated he is “yet to be convinced of its merit.” This sentiment was echoed in a press release from the Australasian College of Road Safety (ACRS) and the Royal Australasian College of Surgeons (RACS). Speaking on behalf of the RACS, Dr Richard Lewandowski stated “Helmet laws aren’t broken – don’t fix them.”

One of the key arguments against MHL is the belief it deters people from cycling. I call it the “Helmet Law as Cycling Deterrent” Hypothesis.

I believe this hypothesis began with a paper by Dorothy Robinson[1] in which she concluded that bicycle-related head injuries declined because of less cycling and not because helmets were beneficial. To support her hypothesis, Robinson used data from helmet use surveys commissioned by the NSW Roads and Traffic Authority. These reports can be viewed in full here. For these surveys, data was collected from various areas around NSW including road intersections, recreation areas (not each year) and entrances to schools (children only).

The authors of the fourth RTA study, Smith and Milthorpe, wrote about their findings in the proceedings of a conference in 1994[2]. In the abstract, they state “The 1993 survey of over 10,000 observations found no drop in adult ridership following legislation. There were fewer children riding, but numbers varied with the area – suggesting multiple reasons for the decrease.” Later on, the authors state “The unevenness in the change in ridership – up at some sites, down in others – makes it difficult to draw conclusions about trends.”

The conclusions drawn by Robinson and Smith/Milthorpe regarding changes in cycling exposure appear to be at odds. Importantly, Smith and Milthorpe hint at heterogeneous trends from area to area, yet Robinson’s analysis is based on aggregated data by general type of area (i.e., road intersection, recreation area or school entrance). This type of analysis is acceptable if the trends are homogeneous and, if Smith and Milthorpe’s conclusions are true, then Robinson’s analysis would be susceptible to Simpson’s paradox (conclusions differ when data is aggregated or not). So, what does the actual data tell us?

First, for the “Helmet Law as Cycling Deterrent” hypothesis to be true, I believe there are at least four criteria that must be met. They are

  1. There is less cycling corresponding to the helmet law date that is beyond any existing trends
  2. The decline must be sustained
  3. The effect is consistent over different data sets and jurisdictions
  4. The decline in cycling is associated with an increase in the proportion of helmet wearing

With regards to the NSW data, there are a few important dates to consider before conducting any analysis. The helmet law was effective at different dates for adults (1 Jan 1991) and children under 16 (1 Jul 1991). The RTA surveys were collected over a monthly period in Oct 90, Apr 91, Apr 92 and Apr 93. So, there was one and two surveys before the adult and child helmet laws respectively. As I’ve noted elsewhere[3], Robinson does not consider the adult cycling counts or the 1990 data for children. Additionally, I have questioned the use of these surveys to estimate trends in cycling exposure as data is only available for 4 out of 48 months and caution should be taken for data that is repurposed[3,4]. Essentially it’s like trying to estimate a trend from a time series where over 90% of the data is missing. However, given that it’s the only NSW cycling counts available for this period, whether useful or not, we can at least look to see if there is any obvious evidence for the hypothesis.

Below is a graph of child cycling counts at road intersections for the Sydney suburbs of Hornsby, Liverpool, Woollahra and Manly. The red vertical line corresponds to the helmet law date.


The pattern in the child counts for Hornsby, Liverpool and Woollahra each appear to violate at least one of the “Helmet Law as Cycling Deterrent” hypothesis criteria. There is a decline in child counts in Hornsby but it would appear to be part of an existing trend (criterion 1), there is an increase in child cycling counts after the helmet law in Liverpool (criterion 1), and there is a post-law decline in Woollahra but it was not sustained by the next survey (criterion 2).

The apparent pattern in the Manly counts seems to meet both criteria 1 and 2. The pre-law counts in 1990 and 1991 are similar which is followed by a steady, sustained decline thereafter. This pattern is similar to that of 7 out of 24 road intersection areas for child counts including Penrith and Sutherland. Although a decline in cycling is troubling, this makes me believe the NSW child cycling counts, taken as a whole, do not meet criterion 3 (i.e., a consistent effect) and, therefore, the “Helmet Law as Cycling Deterrent” hypothesis is not supported by the NSW data.

Another point that seems relevant here comes from a comment by Colin Clarke. When I asked him why Robinson did not include adult cycling counts or the 1990 data for children, his response was essentially bad weather in Oct 1990. It is well known that weather is a major deterrent to cycling and the presumption then is the cycling counts in that month are lower than they are without poor weather. It has been proposed the increase in NSW adult cycling after their helmet law is just an artifact of this. That argument has some face validity as I doubt helmet legislation would lead to more cycling.

What seems to be missing from this argument, however, is that weather is also a deterrent for children. So, the 1990 child cycling counts in Manly are lower than expected. If helmet legislation has no effect on adult cycling, we can use the average of the 1991-93 counts as a guide for the 1990 count. Specifically, the 1990 adult count in Manly of 252 is about 85% of the 1991-93 average of 296.3. We could then use that to adjust the Manly child count for 1990. A graph of both Manly adult and child counts is given below with and without the 1990 adjustment for children.


From the left panel, it is clear the adult cycling counts did not change in a negative manner with the helmet law. The 1993 count of 284 adult cyclists is a 12.7% increase from the 1990 count. The Manly child count data with the 1990 weather adjustment fits a line extremely well (R2=0.996), although caution should be exercised from making inferences from only four points. Although this is just a guess, the Manly child counts don’t satisfy criterion 1 as the trend existed before the helmet law.

In addition to the cycling count data, the “Helmet Law as Cycling Deterrent” hypothesis is also not supported by the hospitalisation record[5,6]. If this hypothesis were true, then there would be a decline in non-head bicycle related hospitalisations, yet these have changed little beyond the decline in casualties for all road users (i.e., criteria 1 and 3 are not supported).

On the other end of the spectrum, Chris Rissel, in a piece for The Conversation, reiterates many points against helmet legislation found in his work and on many anti-helmet websites. This includes the “Helmet Law as Cycling Deterrent” Hypothesis where he states “any reductions in head injuries attributed to the legislation actually due to a marked reduction in the number of people cycling.” With the citation being Dorothy Robinson’s 1996 article discussed above. As demonstrated, there is little solid evidence for “Helmet Law as Cycling Deterrent” Hypothesis or any of his other arguments as discussed here, here, here, here, here and here.

  1. Robinson, D.L. (1996). Head injuries and bicycle helmet laws. Accident Analysis and Prevention, 28, 463-475.
  2. Smith, N.C., Milthorpe, F.W. (1994). Bicycle helmet wearing in New South Wales after legislative mandate. Proceedings of Pedestrian and Bicyclist Safety and Travel Workshop.
  3. Olivier, J., Grzebieta, R., Wang, J.J.J. & Walter, S. (2013). Statistical Errors in Anti-Helmet Arguments. Australasian College of Road Safety Conference.
  4. Walter, S.R., Olivier, J., Churches, T. & Grzebieta, R. (2013). The impact of compulsory helmet legislation on cyclist head injuries in New South Wales, Australia: A response.
  5. Williams, M-A. (1994). Evaluation of the NSW Introduction of Compulsory Bicycle Helmet Legislation. NSW Roads and Traffic Authority, Rosebery, NSW.
  6. Walter, S.R., Olivier, J., Churches, T., & Grzebeita, R. (2011). The impact of compulsory cycle helmet legislation on cyclist head injuries in New South Wales, Australia. Accident Analysis and Prevention, 43, 2064–2071.

Bicycle Helmets and Diffuse Axonal Injury

In an examination of a meta-analysis demonstrating bicycle helmets are effective in mitigating head injury[1], Bill Curnow suggested helmets exacerbate rotational injuries, the more serious being diffuse axonal injury (DAI) [2]. The implication of this hypothesis is clear — bicycle helmets have the potential to increase injury risk. However, no supportive cycling injury data is presented and many have taken this hypothesis as fact[3-8]. Note that these groups openly advocate against helmet use and/or helmet laws and Curnow runs an anti-helmet website (crag.asn.au). The notable exception is a website maintained by the University of Michigan School of Public Health which, given the lack of evidence discussed below, is disappointing.

There are now several studies, with varying degrees of experimental control and real-world comparability, that have investigated the proposed bicycle helmet/DAI link. None of these studies has found evidence supportive of Curnow’s hypothesis.

  • A series of computer simulations found helmets did not increase the likelihood of neck injury for children or adults[9,10],
  • a series of oblique impact dummy tests found helmets did not increase angular acceleration[11],
  • there were no DAI cases found among 110 cyclists in a trauma registry[12] and
  • only 12 potential DAI cases were identified among 6745 cyclists involved in a collision with a motor vehicle (7 were unhelmeted)[13].

In summary, there is no evidence supportive of Curnow’s DAI hypothesis for bicycle helmets. In fact, given available injury data, diffuse axonal injury is uncommon for cyclists presenting to a hospital or trauma center. Given the lack of supportive evidence, Curnow’s DAI hypothesis should not be used to influence cycling safety policy.

A more detailed analysis (and other cycling-related analyses) can be found in our peer-reviewed paper[14].

  1. Attewell, R.G., Glase, K. & McFadden, M. (2001). Bicycle helmet efficacy: a meta-analysis. Accident Analysis and Prevention, 33, 345–352.
  2. Curnow, W.J. (2003). The efficacy of bicycle helmets against brain injury. Accident Analysis and Prevention, 35, 287-292.
  3. BHRF. (2003). Cycle helmets and rotational injuries. Bicycle Helmet Research Foundation. Available at: http://www.cyclehelmets.org/1039.html. (accessed 19.07.13)
  4. Bicycle Australia. (2010). Bicycle Helmets. Available at: http://bicycleaustralia.org/helmets.php. (accessed 19.07.13)
  5. Bicycle NSW. (2013). Bicycle Helmets. Available at: http://bicyclensw.org.au/advocacy/positions/legal/helmets/. (accessed 19.07.13)
  6. Gillham, C. (2011). Mandatory bicycle helmet law in Western Australia. Available at: http://www.cycle-helmets.com/. (accessed 19.07.13)
  7. Stewart, M. (2012). The Myth of the Bicycle Helmet. Available at: http://www.risksense.org/2012/06/14/the-myth-of-the-bicycle-helmet/. (accessed 19.07.13)
  8. Rissel, C. (2012). The impact of compulsory cycle helmet legislation on cyclist head injuries in New South Wales, Australia: A rejoinder. Accident Analysis and Prevention, 45, 107-109.
  9. McNally, D.S. & Rosenberg, N.M. (2013). MADYMO simulation of children in cycle accidents: A novel approach in risk assessment. Accident Analysis and Prevention, 59, 469–478.
  10. McNally, D.S. & Whitehead, S. (2013). A computational simulation study of helmet wearing on head injury risk in adult cyclists. Accident Analysis and Prevention, 60, 15–23.
  11. McIntosh, A.S., Lai, A. & Schilter, E. (2013). Bicycle Helmets: Head Impact Dynamics in Helmeted and Unhelmeted Oblique Impact Tests. Traffic Injury Prevention, 14, 501-508.
  12. Dinh, M.M., Curtis, K. & Ivers, R. (2013). The effectiveness of helmets in reducing head injuries and hospital treatment costs: a multicentre study. MJA, 198, 416-417.
  13. Walter, S.R., Olivier, J., Churches, T. & Grzebieta, R. (2013). The impact of compulsory helmet legislation on cyclist head injuries in New South Wales, Australia: A response. 
  14. Olivier, J., Grzebieta, R., Wang, J.J.J. & Walter, S. (2013). Statistical Errors in Anti-Helmet Arguments. Australasian College of Road Safety Conference.

Safety in Numbers Hypothesis for Cycling

Safety in Numbers is a well known hypothesis in cycling safety. Essentially, the argument is the number of injuries per cyclist decreases as the amount of cycling increases [1,2]. Or put another way, the risk of cycling injury increases when cycling amounts decrease.

The mathematical expression for safety in numbers can be written as


where I and C represent number of injuries and amount of cycling respectively. The exponent 0.4 has been suggested by Robinson[2].

Is there evidence supporting this phenomena using NSW hospitalization and cycling participation surveys? The data used can be found here [3,4].

Here are plots of the expected (red dashed line) number of head and arm injuries (left panel) and head injuries only (right panel) if the Safety in Numbers hypothesis is true. The black line represents the observed number of injuries by year.

SiN Effect

There is a clear divergence between what was observed and what was expected. Therefore, the evidence does not support Robinson’s safety in numbers hypothesis. In fact, the estimated exponent is 0.94 (95% CI: 0.59-1.30) and suggests increases in cycling is associated with a roughly equal increase in injury.

A more detailed analysis (and other cycling-related analyses) can be found in our peer-reviewed paper[5].

  1. Jacobsen, P.L. (2003). Safety in numbers: more walkers and bicyclists, safer walking and bicycling. Injury Prevention, 9, 205-209.
  2. Robinson, D.L. (2005). Safety in numbers in Australia: more walkers and bicyclists, safer walking and bicycling. Health Promotion Journal of Australia, 16, 47-51.
  3. Olivier, J., Walter, S.R., & Grzebieta, R.H. (2013). Long-term bicycle related head injury trends for New South Wales, Australia following mandatory helmet legislation. Accident Analysis and Prevention, 50, 1128–1134.
  4. Australian Bureau of Statistics, 2001. Participation in Exercise, Recreation and Sport 2001. ABS, Canberra.
  5. Olivier, J., Grzebieta, R., Wang, J.J.J. & Walter, S. (2013). Statistical Errors in Anti-Helmet Arguments. Australasian College of Road Safety Conference.