Close

Note: this critique can also be viewed/downloaded in PDF format (832kb)

The University of NSW/Sax Institute study The impact of compulsory cycle helmet legislation on cyclist head injuries in New South Wales, Australia by Walter, Olivier, Churches and Grzebieta attempts to prove that the ratio of cyclist head vs limb injuries justifies retention of the law, but fails to recognise within its own methodology the increased risk of cyclist accidents/injuries that negates any possible benefit from the mandatory wearing of helmets.

This failure is highlighted by the observations of lead author Jake Olivier, senior lecturer at UNSW's Prince of Wales Clinical School, published by The Conversation website on June 23, 2011.

Olivier concedes that the number of child cyclists reduced by around one third following helmet law enforcement, although he cites one study commissioned by the Road Traffic Authority but ignores the detail. For example, a 1993 study conducted by Smith MC and Milthorpe MW (An observational survey of law compliance and helmet wearing by cyclists in New South Wales, RTA 1993 (ISBN0-7305-9110-7) found:


  • school students riding to/from NSW schools: total counts 3107 in 1991 to 1648 in 1993, a drop of 47%
  • for female students the figures were 654 in 1991 down to 222 in 1993, a drop of 64%
  • for secondary female students the reduction in cycling was greater: 455 in 1991 to 106 in 1993, a drop of 77%
  • for secondary children cycling to school in Sydney the reduction was from 904 to 294, a drop of 67%.
  • the largest reduction in cycling was among secondary female students in Sydney: 214 in 1991 down to 20 in 1993, a drop of 90.6%


Between 1991 and 1993, the NSW Road Traffic Authority measured a decrease in <16 child cycling of 44% (Walker M. Law compliance among cyclists in New South Wales, April 1992. A third survey.: Road and Traffic Authority Network Efficiency Strategy Branch, July 1992 / Smith N, Milthorpe F. An observational survey of law compliance and helmet wearing by bicyclists in New South Wales - 1993.: Roads and Traffic Authority, 1993.)

Whether it's a third, 44% or 90.6% reduction in child cycling, it could reasonably be expected that such declines would see a roughly commensurate reduction in accidents/injuries among children after helmet law introduction, regardless of whether or not they wore helmets.

Walter et al note there are data and timing flaws in the Voukelatos/Rissel study, The effects of bicycle helmet legislation on cycling-related injury: The ratio of head to arm injuries over time (p50) which prompted their analysis, but critics concede that its age-group-specific injury counts are fairly accurate.

As presented in the Voukelatos/Rissel study, those injury trends for children were:


Table 2. Cases of head and arm injuries for hospitalised cycling-related injuries by age group

Ages 0-14
YearHeadArm
ICD9ICD10ICD9ICD10
1988-89421334
1989-90423409
1990-91356338
1991-92291397
1992-93310446
1993-94315476
1994-95311521
1995-96330617
1996-97373595
1997-98386640
1998-99288554484587
1999-00339620567712
2000-01574612
2001-02466615
2002-03544675
2003-04479678
2004-05480753
2005-06496641
2006-07445657
2007-08403526


Head injuries came down: comparing the first three years of data (88-91 av 400) to the following three years (91-93 av 305), a reduction of 24%; or comparing the first four years of data (88-92 av 373) to the following four years (92-96 av 317), a reduction of 15%.

Arm injuries went up: comparing the first three years of data (88-91 av 360) to the following three years (91-93 av 440), an increase of 22%; or comparing the first four years of data (88-92 av 370) to the following four years (92-96 av 515), an increase of 39%.

So Olivier concedes that child cycling fell by about 33% when the helmet law was introduced (with children helmet wearing rates increasing from about 20% to more than 60% within two months of helmet law introduction), yet is satisfied that head injuries only declined by 24% on a three year comparison and just 15% on a four year comparison.

More disturbing is that even though child cycling fell by about 33%, according to Olivier, arm injuries increased by 22% on a three year comparison and by 39% on a four year comparison.

The NSW University/Sax Institute study asserts that 1988-89 should be discounted due to an absence of admission data.

Excluding 1988-89, head injuries came down: comparing the first two years of data (89-91 av 390) to the following two years (91-93 av 301), a reduction of 23%; or comparing the first three years of data (89-92 av 357) to the following three years (92-95 av 312), a reduction of 13%.

Arm injuries went up: comparing the first two years of data (89-91 av 374) to the following two years (91-93 av 422), an increase of 13%; or comparing the first three years of data (89-92 av 381) to the following three years (92-95 av 481), an increase of 26%.

Helmets do not prevent arm injury - indeed, they may increase arm injury - and being indicative of non-head injury the results suggest there was an increased number of accidents and thus injuries despite a smaller pool of child cyclists.

A further consequence of reduced cycling is a significant decrease in public health as people abandon regular recreational exercise during an era of high obesity. The ongoing discouragement of public cycling participation is apparent in the failure of Bike Share schemes in Melbourne and Brisbane, despite the success of such schemes in all other countries without mandatory all-age helmet legislation.

Olivier contends that "any major drop in cycling rates would have resulted in a drop in head and arm injury rates. So the comparisons we made were 'exposure free'". This implies that no decrease in arm (or leg) injuries would prove that cyclist numbers didn't fall, ignoring the possibility that the ratio of arm (or leg) injuries per cyclist might have increased.

Smith-Milthorpe's 1993 study suggested a 67% drop in secondary student cycling, and the closest data range from the Voukelatos/Rissel study is the 15-24 age group:

Head injuries came down among 15-24yo: comparing the first three years of data (88-91 av 160) to the following three years (91-93 av 128), a reduction of 20%; or comparing the first four years of data (88-92 av 153) to the following four years (92-96 av 123), a reduction of 20%.

Arm injuries went up among 15-24yo: comparing the first three years of data (88-91 av 106) to the following three years (91-93 av 111), an increase of 5%; or comparing the first four years of data (88-92 av 103) to the following four years (92-96 av 130), an increase of 26%.

Walter et al speculate whether the downturn in child cycling was temporary or permanent. Trends in Australian children traveling to school 1971-2003: burning petrol or carbohydrates? (final page Appendix 4) shows the decline in walking and cycling to school by children across Australia aged 5-9 and 10-14 from 1971 to 2003 (other = cycling). This and other surveys suggest the downturn was permanent.


australian school transport


But what of adults? Olivier states that "there seemed to be an increase in adult riders" after the helmet law was introduced, and in his study cites Walker, Cameron, Smith and Milthorpe (see below).


nsw child cyclist numbers


nsw cyclist numbers


Report CR69, Day-to-Day Travel in Australia 1985-86, a statistical data analysis commissioned by the Department of Transport and Communication and the Federal Office of Road Safety, estimated 147,200 people above the age of nine used bicycles daily in Sydney (see below), five years before helmet law enforcement.


australian cyclist numbers


Cycling in Sydney: Bicycle Ownership and Use suggests that in 2000, there were an average 94,571 bicycle trip per day in Sydney - down 35% on the 1985/86 numbers. On Census day 2006, an estimated 22,161 people used a bicycle for their journey to work in NSW (see below) and cycling as a proportion of work trips in Sydney was at the same level as it was in 1986 (see below).


nsw commuter trips


australian bike share to work

Above is an extract from Cycling Down Under: A Comparative Analysis of Bicycling Trends and Policies in Sydney and Melbourne by John Pucher, Jan Garrard and Stephen Greaves (Journal of Transport Geography, Vol. 18, 2010). Based on population (1986 - 0.7% x 3,472,000 population = 24,304 / 2006 - 0.7% x 4,282,000 population = 29,974) this suggests about 5,670 more Sydney cyclists riding to work in 2006 compared to 1986.


australian commuter numbers to work

Above is an extract from Travel to work in Australian capital cities, 1976-2006: an analysis of census data by Mees, Sorupia and Stone (GAMUT and University of Melbourne, December 2007) showing Sydney adult cycling to work fell from 9,262 in 1986 to 8,934 in 1991 to 8,193 in 1996 to 9,223 in 2001.


The 2008/09 Sydney Household Travel Survey (p26) shows there were 106,000 bike trips on average weekdays in 2008/09, compared to 101,000 bike trips in 2001/02 (see below), and data from Cycling in New South Wales: What the data tells us shows Sydney had Australia's lowest proportion of bicycle use for journeys to work at 0.7% in 2006 (see below).


nsw travel mode


australia average bike share


Above is an extract from Cycling in New South Wales: What the data tells us prepared for the Premier's Council for Active Living, December 2008

Note: Bicycle helmets are not mandatory in Darwin for adults cycling on public paths. Extracted from the same report, the Northern Territory cyclist injury rate based on 2006 Census data is displayed below:


northern territory bicycle injuries


Age-group-specific and all-age data in the Voukelatos/Rissel study is not consistent and the all age totals in particular are considered unreliable. Nevertheless, the Rissel data shows:


Table 2. Cases of head and arm injuries for hospitalised cycling-related injuries by age group

All ages
YearHeadArm
ICD9ICD10ICD9ICD10
1988-89702499
1989-90770666
1990-91640638
1991-92509627
1992-93579765
1993-94513692
1994-95505756
1995-96532904
1996-97581913
1997-98618979
1998-9951111708121212
1999-0058113239661421
2000-0112931341
2001-0213211462
2002-0313551540
2003-0415191731
2004-0515141863
2005-0616241956
2006-0716191955
2007-0814431754


Head injuries came down: comparing the first three years of data (88-91 av 704) to the following three years (91-93 av 534), a reduction of 24%; or comparing the first four years of data (88-92 av 655) to the following four years (92-96 av 532), a reduction of 19%.

Arm injuries went up: comparing the first three years of data (88-91 av 601) to the following three years (91-93 av 695), an increase of 16%; or comparing the first four years of data (88-92 av 608) to the following four years (92-96 av 779), an increase of 28%.

As critics believe the all-age data in the Voukelatos/Rissel study is most likely to be inaccurate, the all-age table can be recalculated based on the age-group-specific data:


Table 2. Cases of head and arm injuries for hospitalised cycling-related injuries by age group

All ages
YearHeadArm
ICD9ICD10ICD9ICD10
1988-89662499
1989-90747632
1990-91629541
1991-92527609
1992-93548735
1993-94551731
1994-95537804
1995-96562955
1996-97615959
1997-986521014
1998-9952111588151212
1999-0059113179701418
2000-0112621319
2001-0212281389
2002-0312651472
2003-0411401479
2004-0512021582
2005-0613141638
2006-0712941630
2007-0811601399


Head injuries came down: comparing the first three years of data (88-91 av 679) to the following three years (91-93 av 542), a reduction of 20%; or comparing the first four years of data (88-92 av 641) to the following four years (92-96 av 550), a reduction of 14%.

Arm injuries went up: comparing the first three years of data (88-91 av 557) to the following three years (91-93 av 692), an increase of 24%; or comparing the first four years of data (88-92 av 570) to the following four years (92-96 av 806), an increase of 41%.

The NSW University/Sax Institute study asserts that 1988-89 should be discounted due to an absence of admission data.

Excluding 1988-89, head injuries came down: comparing the two pre-law years of data (89-91 av 688) to the following two years (91-93 av 538), a reduction of 22%; or comparing the first three years of data (89-92 av 634) to the following three years (92-95 av 545), a reduction of 14%.

Arm injuries went up: comparing the first two years of data (89-91 av 587) to the following two years (91-93 av 672), an increase of 14%; or comparing the first three years of data (89-92 av 594) to the following three years (92-95 av 757, an increase of 27%.

Annual results in all ICD9 data suggest increasing head and particularly arm injuries beyond 1996 to 2000, a worsening trend maintained within the ICD10 data to 2008 as cyclist numbers gradually recovered from their 1991/92 slump.

Whichever way the head/arm injury data is diced from the Voukelatos/Rissel study, it indicates a lower proportion of head injuries but an overwhelming increase in total accidents/injuries per cyclist on the road.

Walter et al repeated the arm and leg analyses on pedestrians "so if we found a big drop in pedestrians, that would be an indication of general road safety improvements. But we did not see a reduction in pedestrian head injury at all relative to limb injuries."


NSW pedestrian fatalities and Injuries


YearKilledInjured
19603674022
19702914346
19802524161
19901773944
19971142985
20001102979
2001882861
2002942607
2003942490
2004852301
2005962220

2006722126
2007682119

Source : RTA document Road Traffic Crashes in New South Wales 2007


Walter et al ignore the fact that since bicycle helmet laws were introduced, pedestrian injuries have almost halved, indicative of general road safety improvements - yet cyclist injuries have increased.

The 1991 introduction of speed cameras in NSW is likely to be a significant factor in the pedestrian death/injury rate after 1990, yet the Walter et al study contends that, for cyclists, "speed modifying interventions would not be expected to have a marked differential effect on head and arm injury rates". Why not?

Within the NSW University/Sax Institute study of June 2011, it is stated that "counts of road accidents for pedestrians, motorists and cyclists showed a marked decline approximately between 1989 and 1992 (Fig. 1)". Figure 1 shows that annual casualty rates for pedestrians, cyclists and motor vehicle occupants had all been falling from 1986/87 to 1991/92, after which they all plateaued.

This is not dissimilar to the finding by Voukelatos/Rissel, who calculated injuries among both on-road and off-road cyclists had been falling prior to helmet law enforcement. On and off road cyclist injuries actually increased after 1991/92.

In considering whether helmets are currently a barrier to cycling, Olivier writes:

Helmets aren’t a major barrier. There’s a widely cited survey by the Cycling Promotion Fund and the National Heart Foundation that suggests it is one of many coming in as the tenth most selected barrier. However, there were some problems with their methodology in terms of finding the primary barriers to cycling.

The researchers asked, "What do you find are the barriers to cycling?" and gave the respondents a list of choices, allowing them to tick as many as they wanted. The problem with that is you don’t get an idea of what the main barrier of cycling is for these people.

The results showed around 16% said the helmet law was a barrier to cycling and it was ranked the tenth most common barrier. So when you consider that this might not be the main barrier, the actual figure is likely to be much lower than 16%.

No it's not. The CPF survey asked 1000 respondents if they own or have access to a bike and 596 said they do.

Among 158 who have ridden a bike for transport in the past month, 26 ticked that they don't ride more often for transport because they don't like wearing a helmet. What of the other 438 of the 596 who own or have access to a bike but haven't ridden for transport in the past month? Is their reason invalid because they rode two or 12 months ago?

Within reasons for not owning or having access to a bike, the 404 respondents were not given a multiple choice option referring to helmets, although 13.1% ticked "Other".

Among 515 of the 596 who don't ride a bike for transport, 81 said it was at least partly because they don't like wearing a helmet. Within reasons for not being interested in riding a bike for transport, 515 were not given a multiple choice option referring to helmets, although 11.9% ticked "Other".

All 1000 were asked which conditions would encourage them to ride a bike more often than they currently do (or at all) and among a multiple choice of nine answers, none of the choices offered by the CPF referred to helmets.

The 16.5% is perfectly valid, if not an underestimate, and that rounds to one in six people (who own a bike and have cycled in the past month) who cite helmets as a barrier to cycling. What might be the opinion of everybody else on the planet?


Conclusion

The data accepted as accurate by pro helmet law researchers shows that child cyclist road numbers fell by a greater percentage than the reduction in child head injuries, at the same time actually increasing the number of non-head injuries. All survey data within the public domain suggests a similar result for adult and all-age cyclists.

This is the same result recorded in other Australian state jurisdictions. In Western Australia, Main Roads WA surveys showed a reduction of at least 30% in cyclist numbers but no decrease in hospital admissions, with admissions increasing by about 30% once cyclist numbers had recovered to pre-law levels by 2000.

Cyclist road number/injury data since helmet law enforcement suggests an approximate 10% reduction in head injury risk but an approximate 30% increase in accident/injury risk due to factors such as risk compensation, safety in numbers, helmet size, rotational brain injury and increased car traffic density when discouraged cyclists drive instead.

Among all age groups, mandatory helmet wearing reduces the risk of head injury but increases the accident risk by a greater per cent, resulting in a larger number of head injuries and a substantially larger number of non-head injuries from the same size pool of cyclists on the road.

Based on the analysis by Walter et al, the question of whether mandatory helmet laws reduce head injury, increase accident risk or damage public health should be properly debated in NSW, and the law should be repealed.


Note: this critique can also be viewed/downloaded in PDF format (832kb)

Mandatory Bicycle Helmet Law in Western Australia

Top of page

Close