Conference PaperPDF Available

Effects of bicycle helmet wearing on accident and injury rates

Authors:
  • Cycling UK

Abstract

Bicycle helmet wearing globally has increased over the past 30 years via promotion and in some cases legislation. Various reports have assessed the changes in wearing rates, accidents, injuries and cycling activity levels. A limited number of reports have analysed overall changes in accident risk per kilometres cycled, per hours cycled or in relationship to cycling levels via survey information. A significant number of findings suggest a higher accident/injury rate may result from helmet usage. Accident data from Australia, the United States, Canada, the United Kingdom and New Zealand indicate the accident rate per hour cycled or per miles cycled has increased with greater helmet usage, most likely from a greater proportion and number of upper limb injuries. Consideration is given to why the accident rate may change and if overall safety is improved or reduced by helmet usage.
1
Effects of bicycle helmet wearing on accident and injury rates
Colin Clarke1, Chris Gillham2
1 Cycling UK (CTC)– Former Councillor
9 The Crescent, Stamford Bridge, York
Email colinclarkecycling@hotmail.co.uk
2 Journalist
Author: Mandatory Bicycle Helmet Law in Western Australia (www.cycle-helmets.com)
2a Burdham Way, Balga, Western Australia
Email chris@scribeworks.com.au
Keywords:
Bicycle
Helmet
Injury
Accident rate
Cycling participation
Abstract
Bicycle helmet wearing globally has increased over the past 30 years via promotion and in
some cases legislation. Various reports have assessed the changes in wearing rates, accidents,
injuries and cycling activity levels. A limited number of reports have analysed overall changes
in accident risk per kilometres cycled, per hours cycled or in relationship to cycling levels via
survey information. A significant number of findings suggest a higher accident/injury rate may
result from helmet usage. Accident data from Australia, the United States, Canada, the United
Kingdom and New Zealand indicate the accident rate per hour cycled or per miles cycled has
increased with greater helmet usage, most likely from a greater proportion and number of
upper limb injuries. Consideration is given to why the accident rate may change and if overall
safety is improved or reduced by helmet usage.
Introduction
This study considers cyclist participation and hospital injury data in the all-age mandatory
jurisdictions of Australia and New Zealand, in Canada where various provinces have all-age or
child helmet laws, as well as the United States where child helmets are mandatory in 21 states
and with child/adult helmet laws in various town, city and county jurisdictions across the
country. These four countries have had the world's longest duration of mandatory all-age or child
bicycle helmet laws, whether nationally or in different jurisdictions.
Concerns exist about the effects of bicycle helmet wearing on the accident and injury rates
because various studies have reported negative aspects. A USA report found helmet use more
than doubled the accident rate compared with no helmet (Porter 2016). Another study
demonstrated that helmet use increased the accident/injury rate by 14% (Erke & Elvik 2007) and
a New Zealand study detailed a 20% risk increase per hours cycled in association with increased
wearing rates (Clarke 2012). At the same time, published reports analysing the influence of
bicycle helmets on cyclist injuries generally show a reduction in the proportion of head injuries
as well as an increase in arm injuries. Further, injury data is considered in the non-mandatory
jurisdiction of the United Kingdom where voluntary helmet wearing has nevertheless been
increasing.
2
Method
The patterns of cyclists’ injuries are examined in Australia, New Zealand, Canada and the United
States of America where mandatory helmet laws are applied universally or in various areas.
Evidence is sourced from both published and unpublished papers and from institutional
databases such as the US Centers for Disease Control and Prevention
(https://www.cdc.gov/injury/wisqars/nonfatal.html), the Australian Health and Welfare Institute
(https://www.aihw.gov.au), the New Zealand Household Travel Survey
(http://www.transport.govt.nz/research/travelsurvey/) and the Otago University Injury Prevention
Research Unit's Injury Query System (https://psm-dm.otago.ac.nz/niqs/).
Injury data relating to both helmeted and non-helmeted cyclists are examined, as is the overall
accident trend in the voluntary jurisdiction of the United Kingdom where helmet use has
increased. Possible reasons for changes to the accident rate with helmet usage are considered,
together with the rate of accidents for helmeted v non-helmeted cyclists.
Evidence is considered from reports relating to cycling accidents, injuries, helmet use, changes in
helmet wearing rates and helmet legislation to assess if trends show an association between
helmet use and changes in the cyclist accident rate and/or injury location on the bodies of
cyclists. A recent paper considers fatality aspects in some detail1 and reported, ‘The number of
cyclist’s deaths due to situations not involving motor vehicles appears to increase with helmet
use and needs further investigation’. A summary of these details is provided.
Cyclist injuries vary from minor to fatal, depending on factors such as speed of impact, motor
vehicle involvement, type of fall, etc. Numerous reports provide cyclist injury data. The
proportion of injuries refers to the distribution of injuries by bicyclist body part injured.
Australia
Whately2 provided data for all age cyclists from the Australian Capital Territory for 1979 to
1983, detailing the proportion of injury types from accidents stemming from falls and also
from those involving motor vehicles. The Victorian Injury Surveillance System (VISS)
provided data on children’s injuries for 19893. The Monograph 5 report4 from Queensland
provided data following helmet legislation during 1993-2008, mainly relating to road accidents
for helmeted v non-helmeted cyclists. McIntosh 20135 provides cyclist data for an 18 month
period from a major trauma centre in Sydney. Dinh et al6 provided adult (15 years and older)
accident details from an inner Sydney area.
Data extracts from these reports are shown in Table 1 below, which shows the percentage of
cyclist injuries sustained as listed in each report and the sample size of each report (not all
injuries are listed). The data from Whately relates to main injuries and suggests that a higher
proportion of arm fractures result from falls rather than from accidents with motor vehicles,
14% v 8%. The VISS data for children shows 35% of injuries relate to upper limbs. Most child
injuries result from falls and a high rate of arm injuries would be expected (11% were hit by
motor vehicles, most were falls). The Monograph 5 report details in Table 14 that 28.3% of
injuries are to the arm for helmeted v 18.4% for non-helmeted. The McIntosh 2013 study had
similar results of 26% v 16%. Monograph 5 reported reduced head injuries for helmet wearers
but also mentioned “Injuries to other body regions did not differ noticeably between helmet
wearing riders and non-helmeted riders, except for shoulder and upper limb injuries” .
Helmet wearers have a higher arm injury rate than non-wearers and this suggests they
experience a higher rate of falling. With a higher fall rate comes a relatively lower proportion
of skull fractures compared with motor vehicle accidents, 9% v 21% and about half the length
of hospital stay (Whately 1985). The proportion of skull fractures could be lower for helmeted
than non-helmeted cyclists due to head protection and also a higher proportion of falls.
3
Table 1
Whately 1985
admissions
VISS
1989
Monograph 5
Table 14
McIntosh et al
2013 Table 2
Dinh et al 2015
Table 3
Bic/MV Bic
only children Helm. No
Helm. Helm No
Helm. Helm No
Helm.
Arm 35 28.3 18.3 26.4 16.0 62 41
Arm fracture 8 14
Leg 24 29.3 29.6 11.5 8.0 26 22
Leg fracture 19 10
Head 10 7.4 17.0 24.1 60 40 59
Skull fracture 21 9
Face 24 9.5 16.0 26 48
Face fracture 7 9
Concussion 20 29 32.2 50
IC, inc Conc 34.4 60
Trunk 6 14.4 12.7 6.9 8.0 20 19
Other fracture 9 5
Internal injury 8 8
Other injury 8 16
SHI/Conc 4.4 9.3
Neck 9.2 9.2
Total % 100 100 99 102.5 112.1 135.5 202.0 174 189
Medium age 36.0 21.0 39 31
Average LOS 10.5 4.5
Alcohol use 2 20
Sample size 75 323 916 9854 2037 87 50 200 54
Note: Bic – bicycle, MV – motor vehicle, Helm – helmeted, No Helm – no helmet, Conc = concussion, IC =
intracranial, SHI = Serious head injury (concussion or worse)
Olivier et al7 reported the change in number of arm injuries following helmet law enforcement in
New South Wales, increasing from 660 in 1991 to 1,334 in 2000, a 102.1% increase in the 10
year period. In Western Australia where all age helmet laws were enforced in 1992, the Health
Department also reported upper limb fractures increased from 16.4% in 1987-89 to 22.9% in
1990-92 and 28.6% in 1993-958. Meuleners et al 20039 data shows that cyclist upper extremity
injuries in Western Australia increased from 118 (16.9% of all injury locations) in 1988 to 274
(32.2%) in 1998.
4
Several reports provide data relating to changes in the accident rate with increased helmet
wearing rates. Robinson's 1996 report10 provided injury data for children from Victoria and
New South Wales. In Victoria, the equivalent injury numbers for pre-law levels of cyclist
numbers increased 15% from 1990 to 1992, a period during which all age mandatory bicycle
helmet legislation was introduced in all Australian jurisdictions. Robinson’s data in Table 2 for
children in NSW shows the equivalent number of injuries increased from 1,310 (384 head +
926 other injuries) pre law in 1991 to 2,083 (488 head + 1,595 other injuries) in 1993. The
relative injury rate proportional to cycling levels increased 59% from 1,310 to 2,083. The
relative increase for 'other' injuries was 72% and for ‘head’ was 27%.
Australia serious injury data
Table 2 below shows in 1990 there were 7,520 hospitalised cyclists compared to 5,048
pedestrians, a ratio of 1.49 to 1. In 2003-04 there were 7,929 hospitalisations for cyclists
compared to 3,716 for pedestrians, a ratio of 2.13 to 111. The ratio changes from 1.49 to 2.13
suggests cyclists are more at risk compared with pedestrians, as detailed in the table below. For
cycling level estimates refer: Evaluation of Australia's bicycle helmet laws, 2015 12 . Australian
Institute of Health and Welfare data13 show that in 2005/06 there were 4,370 cyclist hospital
cases for injury due to road vehicle traffic crashes, of which 43.6% involved the shoulder and
upper limbs. There were 2,644 pedestrian cases, of which 15.6% involved the shoulder and
upper limbs. In 2015-16, there were 12027 hospitalisations for cyclists compared to 4016
pedestrians, a ratio of 2.99 to 114.
Table 2
1990 2003/04 2005/06 2008/09 2012/13 2015/16
Cyclists 7520 7929 8814 9577 10098 12027
Pedestrians 5048 3715 3779 3686 3823 4016
C/P (actual) 1.49 2.13 2.33 2.60 2.64 2.99
Cycling level %
proportional 100 60 59 58 N/A N/A
Equivalent C/P
Best estimate 1.49 3.55 3.95 4.47 N/A N/A
Notes:
1) Cyclist data for 1990 reported 6,412 hospitalised and was further adjusted in the report for 'Unknowns', where
the mode of transport was not known, to total 7,520.
2) Victoria introduced a bicycle helmet law in mid1990 and the 1990 'Cyclist' number of 7,520 may have been
higher without the law.
3) The proportion of injuries due to mountain biking may have increased.
4) Table 5 shows cycling to work increased from 2011 to 2016, but still well below 1991 Census ratios. See NCP
details in Discussion re declining recreational cyclist participation.
New Zealand
The chart and Table 3 below compare the change in the number of New Zealand public
hospital discharges15 and total hours cycled16 for people aged 5yrs and older. Relative injury risk
is shown for 1989-1990 and subsequent changes are detailed following their helmet law in
1994. The information shows a reduction in average hours cycled per person of between
41.24% and 59.32% and an increased accident/injury risk based on per million hours cycled of
between 34.96% and 121.31%. For 5yo+ public hospital discharges from accidents not
involving a motor vehicle, the risk of injury per million hours cycled increased 107.2%% from
23.38 in 1989-1990 to 35.85 in 1997-1998 to 48.45 in the 2003/07 period.
5
Table 3
Table 4 below details cycling levels in million hours per year for the periods shown. For the age
group 5-17 total hours cycled reduced from 23 million per year in the 1989/90 period to 7.9
million by 2003/06 period, a drop of 66% and by 2011/14 a reduction of 81%.
1989/
90
1997/
98
2003/
06
2004/
07
2005/
08
2006/
09
2007/
10
2008/1
1
2009/
12
2010/
13
2011/1
4
5-12 years 9.7 6 4.4 3.7 3.3 2.8 3.4 3.3 2.7 1.8 1.6
6
13-17 years 13.3 7.6 3.5 3.5 3.4 3.2 3.2 2.9 2.7 2.3 2.8
18 years and
over 16.2 12.3 13.8 14.8 17.7 17.8 22.3 21.2 22.2 20.3 20.4
Total ( 5+) 39.2 26 21.6 21.9 24.3 23.8 28.9 27.4 27.6 24.4 24.8
Table 4 details million hours cycle per year for age groups.
Clarke 201217 reported on changes to the injury rate and injuries per million hours of travel
following helmet legislation in 1994. The study reported that “by 2003–07, cyclists had a 20%
higher accident rate compared with pre law”. Referring to the ratio of cyclist to pedestrian
injuries per million hours of travel, pre law 1989-91 compared to post law 2004-07, it reported
that “Cyclist’s overall injuries more than doubled compared with pedestrians, 5.97 to 12.91,
indicating a major reduction in safety.”
Tin Tin et al18 provided information on the change to ‘upper extremity’ (injuries per million hours
cycled) in Figure 3. It increased from 4.4 to 13.20, indicating a 200% higher rate by 2003/07.
For New Zealand in five year periods from 1989-93 to 2009-13, the number of cyclist deaths not
involving a motor vehicle increased from 9 to 10, 12, 16 and 29.
Census data on percentage cycling to work for New Zealand and Australia both show reductions
following legislation. Details below in Table 5.
Table 5
1986 1991 1996 2001 2006 2011 2016
Australia 1.56 1.68 1.50 1.24 1.21 1.24 1.29
New Zealand 5.46 5.39 4.04 3.12 2.52 2.9 (2013 yr)
Erke and Elvik 200719 examined research from Australia and New Zealand and stated that
"There is evidence of increased accident risk per cycling-km for cyclists wearing a helmet. In
Australia and New Zealand, the increase is estimated to be around 14 per cent."
Canada
A Health Report published by Statistics Canada in 201720 shows all age cycling in the country
declined from 28.7% of the population in 1994/95 to 23.7% in 2013/14, with 12-14yo cycling
declining from 67.5% to 52.2% and 15-17yo cycling declining from 56.0% to 45.2%. British
Columbia, Ontario, New Brunswick and Nova Scotia introduced all age or under 18yo helmet
laws from 1995 to 1997, with four other provinces progressively following suit up to 2015. In
2018 Clarke21 reported, Provinces without all age helmet laws show on average a higher use of
bicycles and from 1994/95 to 2013/14 also show a better outcome with more people cycling
42.1% v 29.3%’.
British Columbia introduced an all age helmet law in 1996. Surveys were conducted in 1995 and
1999 during late July and early August. The survey counts of 3950 v 4246 showed a 7.5%
increase, similar to the population increase. In the last week of July and first week of August
1995, Vancouver had 88mm of rain compared with 18mm in 199922. The surveys show the 16-30
7
age group had reduced counts of cyclists, equating to about five fewer cycling for each extra one
wearing a helmet (i.e. 50% of 3950 – wearing rate 47% in 1995 v 35% of 4246 - wearing rate
69% in 1999, 1975 total with 928 wearing v 1486 total with 1025 wearing, 489 fewer cycling v
97 extra wearing). A Vancouver cycling study in 2015 reported young adults aged 16-25
accounted for 4% of bicycle trips and 24% of bicycle/motor vehicles accidents23 and indicates
age grouping and accident data needs to be considered separately.
Nova Scotia introduced an all age helmet law in 1997. Surveys from Halifax24 1995/96 to
1998/99 show for the age group 0-19 yrs a drop of 74% in cycling levels - see Table 6 below.
The injury data was published by LeBlance et at25.
Table 6
1995/96 1997 1998/99
Average number cyclists per day ‘A 88 34 52
% age range 0-19 yrs (number) 24.6% (22) 13.1% (4) 10.7% (5)
All injuries ‘B’ 416 222 443
Head injuries ‘C’ 15 3 7
Relative injury risk, B/A 4.72 6.53 8.52
Relative head injury risk C/A 0.17 0.09 0.13
Note - For Canada, the age specific cycling head injury rate for the age group 0-19 per 100,000 population was
approximately 51, and 5.1 for the age group 50-6426. This suggests a disproportionate reduction in head injuries
would result from a reduction in cycling levels for the 0-19 age range.
A helmet law for cyclists aged under 18 was introduced in Alberta, in 2002. Child cycling
decreased by 56% from 2000 to 2006 but injuries per cyclist increased following helmet law
enforcement in the province of Alberta27. From Karkhaneh 2011 report, Table 7 below provides
details of the number of cyclists per hour from observational surveys in Alberta (pre-law in 2000,
post-law in 2006) and compares these to the numbers treated in emergency rooms (ER) for non-
head injuries. For children and teenagers there was a marked reduction in cycling levels and a
substantial increase in injuries relative to cycling activity. Safety slightly improved relative to
cycling levels for adults without a helmet requirement.
Table 7
Cyclists per hour Non-head injuries (ER) Change in injuries relative to
change in cycling
Pre Post CRatio Pre Post IRatio
Children 3.56 1.58 0.44 1676.3 1762.0 1.05 2.37
Teenagers 1.92 1.41 0.73 870.3 1101.0 1.27 1.72
Adults 18+ 6.29 7.58 1.21 1846.7 2062.5 1.12 0.93
Clarke 200928 reported on helmet laws in Canada comparing injury data for areas with helmet
legislation to those without. Overall, the injury outcome adjusted for estimations on cycling
levels found provinces without helmet legislation had a better outcome. In 2015 Teschke et al
8
reported: Helmet legislation was not associated with reduced hospitalisation rates for brain,
head, scalp, skull or face injuries, indicating that factors other than helmet laws have more
influence on injury rates 29. The study focus was on more serious cycling injuries requiring a
hospital stay.
For Canada from 2010 -13, road traffic accidents (RTA, involving motor vehicles) recorded 249
cyclist deaths, whereas Statistics Canada recorded 345 (all deaths) a difference of 96, It appears
more cyclists are dying from falls than occurred in the early 1990s when the data shows a
difference of 7, 363 RTA v 371 SC.
United States of America
Mandatory bicycle helmet laws for youth have been introduced in 21 of the US states and child
helmet wearing is high in all jurisdictions - 79% for the under 12 age group30. Estimates on
cycling levels in the USA vary substantially31. Survey details from 1998, indicated more than 17
billion hrs of cycling annually32, population approximately 270 million, an average of about 63
hrs per capita. The age group 0-15 having the highest levels, mean riding time of 300 hr/yr for
those cycling (about 150 hrs per capita).
Cycling participation among 7-17-year-old children in the US declined 23.1% from a 1995-2003
average of 18,593,000 to a 2004-2012 average of 14,296,889, despite a population increase close
to 10% during that time33. Over the same time period, 7-17yo all-body injuries declined 23.7%
from an average 291,970 to 222,869, while concussion injuries among 7-17-year-old cyclists
declined 2.1% from a 1995-2003 average of 6,555 to a 2004-2012 average of 6,420, refer Figure
1 below. The relative injury rate increased from 1.27% (291607/22948000) in 1995 to 1.87%
(201926/10800000) in 2012, a relative increase of 47%.
Fig 1 Participation – black, Injuries – blue, Concussions- red
The data below in Table 8 compare head, upper body and lower body injury hospital admissions
in the US from 1994-2004 to 2005-2015. It is noteworthy that US Census Bureau data show
22,948,000 cyclists aged 7-17yo in 1995 and 13,196,000 in 2009, a 42.5% reduction. Cycling
9
among 18yo+ was 33,360,000 in 1995 and 24,942,000 in 2009, a 25.2% reduction. A 2012
summer survey estimated that 29% of US adults always wear a bike helmet and 56% never wear
one. Among children aged 5 to 7 years, 42% always wore a helmet and 31% never wore one
while riding34.
Table 8
It should be noted that Outdoor Industry Association35 data show US 6-17yo cycling declined
from 15,550,000 in 2007 to 12,461,000 in 2015, down 19.9%. US 18yo+ cycling increased from
26,576,000 in 2007 to 30,612,000 in 2015, up 15.2%. The decline in upper body injuries detailed
above among 7-17yo is due to an ongoing significant decline in child and teenage bike riding
since the 1990s. On the other hand, 18yo+ cycling has been increasing since the 1990s with
surveys showing a gradual increase in adult helmet wearing. Concussion, head and upper body
injuries have been increasing but lower body injuries have been decreasing when comparing
1994-2004 with 2005-2015.
Research published in the Journal of the American Medical Association36 found injuries among
cyclists aged 18+ in the US increased 28%, hospital admissions increased 120% and head
injuries increased 60% from 1998 to 2013.
The Porter 201637 report in the US detailed that cyclists wearing helmets had more than twice
the odds of suffering an injury than cyclists not wearing helmets. The report’s Figure 1 shows
the percentage injured among wearers and non wearers in the previous two years. The
percentage injured for wearers was 2.81 times higher than for non-wearers.
Carpenter and Stehr reported a reduction in cycling of 4% - 5% occurred due to legislation from
1991 to 2005 for the age group 5-15 years. Chatterji and Markowitz estimated a possible 9%
reduction due to helmet laws. The National Sporting Goods Administration suggests a reduction
in cycling for the 6-17 age group from 1998 to 2016 of about 40%.
Clarke 2018 examined the cycling fatality risk data and reported an increased risk from death not
involving motor vehicles38. For the USA, cyclist deaths not involving a motor vehicle in five
year periods from 1986-90 to 2011-15 increased from 105 to 114, 150, 209, 258 and 296. Of the
1,024 bicyclist deaths in 2017, 679 died in motor vehicle crashes and 345 in other incidents39.
10
United Kingdom
The helmet wearing rate in Great Britain has increased from 2002/03 to 2014/15. In 2002/03
the wearing rate was about 27%, by 2008 about 35%40 and data from police reported accidents
for 2013 shows that 49% of cyclists in known cases were wearing helmets 41. During this period
the accident rate per billion miles cycled increased by approximately 17%. Dodds et al 2018
reported on NHS England data from March 2012 to September 2017 for cyclists older than 15
years, with 61.5% wearing helmets and also having an upper limb injury rate of 61.5% v
47.5% for non-wearers, suggesting helmet wearers having a higher fall off rate (the study did
not have data on how the accidents occurred). Injuries to chest, spine and upper and lower
limbs averaged 0.89 for non-wearers v 1.43 for wearers (66% higher). Alcohol use was
reported as 2.1% for wearers v 15.6% for non-wearers (data from the USA and other countries
show other differences in addition to alcohol use may also occur, ref 45). The mortality data
reports 1.8% for wearers (76 cases) v 5.6% for non-wearers (143 cases) and 102 cases
unknown helmet status. It appears probable that wearers had a higher fall off rate resulting in
less severe injuries. Their injury severity score was 12 v 16 for non-wearers.
Great Britain fatality data (DfT Stats19) not involving motor vehicles from 2012 to 2016 shows
22 helmeted v 15 non helmeted, with 38 unknowns42. Cyclist fatalities involving motor vehicles
were 123 helmeted v 110 non helmeted, with 309 unknowns. Table 9 data below indicates an
increase in the fatal and serious injury rate of between 16% and 19% from 2002/03 to
2012/2015. From 2003 to 2016 cycle traffic increased by 25% and the number of serious injuries
rose by 48 per cent43.
Years 2002/03 2004/05 2006/07 2008/09 2010/11 2012/13 2014/15
Average
rate 938 937 991 964 1048 1122 1095
Table 9 GB data – fatal and serious accident rate per billion miles cycled44
Complimentary evidence
Recent studies have also cast doubt on the reliability of published case control studies and meta-
analyses regarding the effectiveness of bicycle helmets. Zeegers 201545 states that “Three cases
could be found in the literature with sufficient data to assess both risk ratios and odds ratios: the
Netherlands, Victoria (Australia) and Seattle (U.S.A). In all three cases, the problem of
overestimation of the effectiveness of the helmet by using odds ratios did occur. The effect ranges
from small (+ 8 %) to extremely large (> + 400 %). Contrary to the original claim of these
studies, in two out of three cases the risk of getting a head injury proved not to be lower for
helmeted cyclists. Moreover, in all three cases the risk of getting a non-head injury proved to be
higher for cyclists with a helmet.”
The Seattle study, Thompson et al46 claimed, “The use of helmets can reduce the risk of head
injury by 85%”. The 400% overestimation from the Seattle study shows the degree of error that
can occur from using comparisons and odds ratio calculations.
Increased risk
Several possible reasons can be considered that may explain why there is an apparent increased
accident/injury risk associated with helmet use. In nearly all cases they are difficult to evaluate
and may require further research. Cyclists might believe they are safer with a helmet and take
11
increased risks, for example in mountain biking events or travelling at higher speeds. Risk
Compensation and Bicycle Helmets published by Phillips, Fyhri and Sagberg47 in 2011 reported
“Our results show increased cycling speed and decreased risk perception in a helmet-on
compared to a helmet-off condition among cyclists used to wearing helmets, a finding that is in
line with the theory of risk compensation. However, for those cyclists not used to helmets there
were no differences in either risk or behaviour between the helmet-off and helmet-on
conditions.” These findings support the theory of risk compensation among cyclists in a
mandatory helmet environment where they are used to wearing helmets.
Bicycle helmets – A case of risk compensation?48 was published in 2012 with a survey of 1,504
bicycle owners in Norway finding two sub-populations: one speed happy group who cycle fast
and have a lot of equipment including helmets, and one traditional group without much
equipment who cycle slowly. In 2016, Gamble and Walker49 found “In a controlled study in
which a helmet, compared with a baseball cap, was used as the head mount for an eye tracker,
participants scored significantly higher on laboratory measures of both risk taking and sensation
seeking. This happened despite there being no risk for the helmet to ameliorate and despite it
being introduced purely as an eye tracker. The results suggest that unconscious activation of
safety-related concepts primes globally increased risk propensity.
Other explanations for increased accident and injury occurrence for helmeted cyclist may be
increased head diameter, impaired vision, impaired hearing, sideways wind shear and forces50,51,
reduced riding stability and loss of "safety in numbers" due to reduced cycling participation
following helmet law enforcement. In all-age mandatory bicycle helmet jurisdictions, there is
speculation that increased vehicular traffic density might also be a factor as a result of
discouraged cyclists instead driving. With regards to "single-cycle non-collision accidents", the
contributory factor most frequently attributed by UK police in such accidents was "loss of
control" (67% of fatal single-cycle accidents and 44% of serious).52
Research has reported cyclists incurring up to 10g forces due to hitting pot holes and lower g
forces from road humps, manholes covers and situations where the road/path is not smooth and
even. A recent article details g forces from slow cycling speeds incurring up to 6g acceleration
forces53. A lightweight helmet at 0.25 kg incurring a 4g acceleration would involve a force of
about 10N or 2.2 lbs force imperial. In general, helmet use results in extra forces per typical hour
cycled and, on some occasions, they may add to problems in maintaining balance. Therefore, the
increased "other" rate, mainly falls, is what could be expected from helmet use.
Robinson 199654 refers to the Wasserman data that detailed the incidence of cyclists hitting
their head/helmet during an 18-month period was “significantly higher for helmet wearers
(8/40 vs 13/476 - i.e. 20% vs 2.7%, p 0.00001)." A bare head width of approximately 150mm
may avoid contact compared to a helmeted head at approximately 200mm wide (Clarke
200755). Assuming the 20% and 2.7% figures are typical, on a yearly average for helmeted and
non-helmeted the risk of hitting their helmet or head would be 13.2% and 1.8% respectively.
The increased risk of impact for helmeted is about seven times higher. A degree of protection
could be expected plus a degree of risk from the extra impacts.
Discussion
There is not universal agreement that bicycle helmets or any other road safety equipment are
related to risk compensation, but it is a plausible explanation for the increased ratio of injuries
per cyclist in mandatory helmet jurisdictions. We further speculate that parents may allow
children to cycle at a younger age with the added protection of a helmet, or children may think
they are safer and engage in higher risk cycling. It appears that a combination of factors could
result in the higher accident rate, increased risk-taking and factors affecting balance. In the event
12
of an accident, helmet use typically results in higher risk of impact to the helmet than occurs for
a bare head.
It has been 29 years since Australia became the first country to mandate all-age bicycle helmets
and the evidence is accumulating that such laws have a long-term detrimental impact on cycling
participation and the cyclist crash/injury ratio. For example, the National Cycling Participation
(NCP) surveys conducted by the Australian government show the total population proportion
cycling at least once per week decreased from 18.2% in 2011 to 13.8% in 2019, with annual
cycling down from 40.2% to 35.0%56 Both of these surveys included children cycling in their
garden, unlike traditional cycling surveys.
Despite the participation decline, cyclist injuries continued to increase. For example, the NCP
survey data show the population proportion who cycle weekly in Victoria declined from 19.9%
in 2011 to 13.7% in 2019, yet recently published research57 shows the number of cyclist major
trauma cases hospitalised in the state increased 8% per annum between 2007 and 2015.
We speculate that in the generational timespan of 29 years there is a demographic bulge of baby
boomers who have maintained Australian participation levels but are now retiring due to age and
not being replaced by younger people who have been discouraged by helmet laws as they grew
up. The authors of the NCP 2011-2017 survey reports also state: "… it is likely that the gradual
ageing of the Australian population has contributed to the participation trend, and this
demographic shift is likely to exacerbate the challenge of increasing cycling participation in
future as the population continues to age. The strong correlation between age and cycling
participation means that over time we would expect cycling participation to decline without
significant policy intervention or natural cultural shifts."
In Australia from 1990 to 2014-15, hospital admissions from road accidents decreased for car
occupants by 8.4% and for pedestrians by 45.1%, but cyclist admissions increased 7.1%58. For
Alberta Canada, the change in injuries relative to change in cycling for children is calculated at
237%. From the USA, Porter reports the percentage injured for wearers was 2.81 times higher
than for non-wearers.
For New Zealand 5yo+ages, the data in Table 3 above shows the accident rate per million hours
increased from 31.35 to 69.38. This related to all accidents involving motor vehicles and for
"other" reasons, probably mainly falls and riding into objects or stationary vehicles. For those
involving only motor vehicles it did not change very much, roughly eight per million hours to
seven per million hours. Accidents due to "other" increased substantially from approximately 23
to 63 per million hours. The 23 accidents per million hours is approximately one per 43,000
hours of cycling on average. If cycling 10 hours per week, this would equate to once in 82 years
on average. The disadvantages to helmets and increased risk taking only needs to be minute to
increase the accident rate and exceed the expected benefits in a proportion of accidents.
Conclusions
Published studies and research within this paper from jurisdictions with all-age, child and/or
adult mandatory helmet jurisdictions consistently suggest reductions in cycling participation
when laws are enforced but either no corresponding decrease in hospitalised total injuries or an
increase in all-body injuries. Further, there is strong evidence that helmeted cyclists suffer a
higher rate of upper body limb injuries than non-wearers, suggesting a higher rate of falls than
non-wearers. This may be due to factors such as risk compensation, imbalance caused by the
helmet or peripheral vision occasionally obscured by some helmet designs, particularly among
cyclists who lower their heads to improve their aerodynamics.
13
The cause is uncertain but participation and injury data from the countries examined show a
negative safety outcome in terms of increased helmet wearing rates and the accident rate per
hour and per kilometre cycled. In mandatory helmet jurisdictions, the proportion of head injuries
is reduced, although this is partly because the number of upper limb injuries increases.
Confounding factors that are not analysed in this study which may influence participation and
injury trends in different countries include changes in cyclist preference for riding on or off road,
demographic shifts in cycling participation, varying public compliance with helmet regulations,
improved bicycle and motor vehicle safety features, cycle path infrastructure improvements and
stricter road traffic laws in different jurisdictions.
Nevertheless, this study presents evidence that helmet use tends to increase the accident/injury
rate per cyclist, potentially outweighing any head protection benefits. It reinforces the findings of
numerous published studies that mandatory helmet laws reduce cycling participation, which is
detrimental to public health and is likely to also increase vehicular traffic if discouraged bike
riders alternatively drive a car.
The possible reasons for increased risk of injury per cyclist, particularly upper extremities,
appear to be due to increased falls. It appears that helmet use increases the accident rate by more
than 40%. This should be the subject of further research to determine why overall accident and
injury rates outweigh head injury benefits provided by helmets.
Contact Colin Clarke for further information (colinclarkecycling@hotmail.co.uk )
1 Clarke CF Evaluating cycling fatality risk with a focus on cycle helmet use http://worldtransportjournal.com/wp-
content/uploads/2018/12/24.4opt.pdf
2 Whately S, Bicycle Crashes in the Austrian Capital Territories, CR 35, FORS, 1985,p13
3 Victorian Injury Surveillance System, Hazard Vol 6, Dec 1990
4 Bicycle Helmet Research, CARRS-Q Monograph Series - Monograph 5, Centre for Accident Research & Road Safety
Queensland, 2010
5 McIntosh AS, Curtis K, Rankin T et al. Associations between helmet use and brain injuries amongst injured pedal-and
motor-cyclists: A case series analysis of trauma centre presentations. J Australas Coll Road Saf 2013;24:11–20.
6 Dinh MM, Kastelein C, Hopkins R, Royle TJ, Bein KJ, Chalkley DR, Ivers R. Mechanisms, injuries and helmet use
in cyclists presenting to an inner city emergency department. Emerg Med Australas. 2015 Aug;27(4):323-7. doi:
10.1111/1742-6723.12407. Epub 2015 May 4
7 Olivier J, Walter SR, Grzebieta RH, Long term bicycle related head injury trends for New South Wales, Australia
following mandatory helmet legislation, Accid Anal Prev 2013 Jan;50:1128-34. doi: 10.1016/j.aap.2012.09.003. Epub
2012 Sep 28
8 http://www.cycle-helmets.com/wa-bicycle-injuries.html
9 Meuleners L, Gaviin A, Cercarelli L, Hendrie D, Bicycle Crashes and Injuries in Western Australia, 1987-2000
http://www.cycle-helmets.com/RR131.pdf
10 Robinson DL; Head injuries and bicycle helmet laws; Accid Anal Prev, 28, 4: p 463-475, 1996 http://www.cycle-
helmets.com/robinson-head-injuries.pdf
11 Australian Institute of Health and Welfare: Jesia G Berry, & James E Harrison 2007. Serious injury due
to land transport accidents, Australia, 2003–04. AIHW cat. no. INJCAT 107. Canberra: AIHW &
ATSB.
12 Clarke CF, Evaluation of Australia's bicycle helmet laws, The Sports Science Summit, O2 venue
London UK http://www.cycle-helmets.com/au-assessment-2015.pdf Presented 14 January 2015
13 Australian Institute of Health and Welfare, Serious injury due to land transport accidents,
Australia, 2005–06
14 Australian Institute of Health and Welfare, Trends in hospitalised injury, Australia 1999-00 to
2012-13 https://www.aihw.gov.au/getmedia/fc940eea-91aa-47d6-ac20-94243a83a5e9/15555.pdf
https://www.aihw.gov.au/getmedia/fbff1344-6b3f-4b2a-8649-fb879dcd69ab/aihw-injcat-
203.pdf.aspx?inline=true
15 Otago University Injury Prevention Research Unit https://psm-dm.otago.ac.nz/niqs/
16 New Zealand Household Travel Survey 2015-2017 December 2017
http://www.transport.govt.nz/assets/Uploads/Research/Documents/Household-Travel-Survey-intro-
Dec2017.pdf
17 Clarke CF, Evaluation of New Zealand’s bicycle helmet law , Med J NZ http://www.cycle-
helmets.com/nz-clarke-2012.pdf ,
18 Tin Tin S. Injuries to pedal cyclists on New Zealand roads, 1988-2007. BMC Public Health
2010;10:655. http://www.biomedcentral.com/1471-2458/10/655
19 Erke A, Elvik R. Making Vision Zero real: Preventing Pedestrian Accidents And Making
Them Less Severe, Oslo June 2007. page 28. https://www.toi.no/getfile.php/Publikasjoner/T%d8I
%20rapporter/2007/889-2007/889-2007-nett.pdf
20 Ramage-Morin, Statistic Canada, Health Report Cycling in Canada
http://www.statcan.gc.ca/pub/82-003-x/2017004/article/14788-eng.htm
21 Clarke CF, Evaluating cycling fatality risk with a focus on cycle helmet use
http://worldtransportjournal.com/wp-content/uploads/2018/12/24.4opt.pdf
22 Environment Canada, Climate Services.
23 CITY OF VANCOUVER CYCLING SAFETY STUDY FINAL REPORT January 22, 2015
24 Chipman M, Hats off (or not?) to helmet legislation, CanMedAssocJ, 5 March ,166,(5) 2002
25 LeBlanc JC, Beattie TL, Culligan C. Effect of legislation on the use of bicycle
helmets. CMAJ 2002;166(5):592-5.
26 Head Injuries in Canada, A Decade of Change, Canadian Institute of Health Information, August
2006
27 BHRF, Alberta's helmet law – children's cycling halved, injuries increased per cyclist
http://www.cyclehelmets.org/1250.html
28 Clarke, CF, Evaluating bicycle helmet use and legislation in Canada, 2009, http://www.cycle-
helmets.com/canada-helmet-assessment.doc
29 Teschke K, Koehoorn M, Shen H, et al. Bicycling injury hospitalisation rates in Canadian jurisdictions: analyses
examining associations with helmet legislation and mode share. BMJ Open 2015;5:e008052. doi:10.1136/bmjopen-
2015- 008052
30 Rodger GB, State bicycle helmet law effectiveness, Injury Prevention 2002;8:42–46
http://injuryprevention.bmj.com/content/injuryprev/8/1/42.full.pdf
31 Clarke CF Evaluating cycling fatality risk with a focus on cycle helmet use
http://worldtransportjournal.com/wp-content/uploads/2018/12/24.4opt.pdf
32 Rodgers GB, Bicycle and Bicyle Helmet use Patterns in the United States in 1998, J of Safety
Research, Vol 31, No 3, pp. 149-158, 2000.
33 Gillham C, Rissel C, Children’s cycling participation, injuries, fatalities and helmet legislation in
the United States, 2015 http://www.eco-logica.co.uk/pdf/wtpp21.1.pdf
34 Bicycle helmet use among persons 5 years and older in the United States, 2012, A. Jewett, L.F.
Beck, C. Taylor, G. Baldwin https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5189688/
35 2017 Outdoor Recreation Participation, Outdoor Foundation https://outdoorindustry.org/wp-
content/uploads/2017/05/2017-Outdoor-Recreation-Participation-Report_FINAL.pdf
36 Bicycle Trauma Injuries and Hospital Admissions in the United States, 1998-2013, T. Sanford,
C.E. McCulloch, R.A. Callcut, et al, Journal of the American Medical Association, 2015
https://jamanetwork.com/journals/jama/fullarticle/2432153
37 Porter AK, Salvo D, Kohl HW, Correlates of Helmet Use Among Recreation and Transportation
Bicyclists, AJPM 2016. https://uthealth.influuent.utsystem.edu/en/publications/correlates-of-helmet-use-among-
recreation-and-transportation-bicy
38 Clarke CF Evaluating cycling fatality risk with a focus on cycle helmet use
http://worldtransportjournal.com/wp-content/uploads/2018/12/24.4opt.pdf
39 Bicyclist deaths, Injury Facts, https://injuryfacts.nsc.org/home-and-community/safety-
topics/bicycle-deaths/
40 Sharrat C, Walters LK, Anjum O, Cycle helmet wearing 2008, Project PPR 420, TRL UK
41 GB police data provided via DfT,
42 DfT information provided to C Clarke.
43 DfT Pedal Cycling Road Safety Factsheet March 2018
44 DfT Road Casualties GB 2015, RAS20001 p103.
45 Zeegers T, Overestimation of the effectiveness of the bicycle helmet by the use of odds ratios’
http://www.fietsberaad.nl/?
lang=nl&repository=overestimation+of+the+effectiveness+of+the+bicycle+helmet+by+the+use+of
+odds+ratios
46 R.S. Thompson, P.R. Rivara and D.C. Thompson, “A case-control study of the effectiveness
of bicycle safety helmets”, The New England Journal of Medicine 320 (1989), pp. 1361-
1367.
47 Phillips RO1, Fyhri A, Sagberg F. Risk compensation and bicycle helmets.
https://www.ncbi.nlm.nih.gov/pubmed/21418079
48 Fyhri A Bjørnskau T Backer-Grøndahl A, Bicycle helmets – A case of risk compensation?,
https://www.sciencedirect.com/science/article/abs/pii/S1369847812000587
49 Gamble t, Walker I, Bicycle Helmet Can Increase Risk Taking and Sensation Seeking in Adults,
2016 http://journals.sagepub.com/doi/abs/10.1177/0956797615620784
50 Fintelman, DM, Sterling, M, Hemida, H & Li, FX 2014, 'The effect of crosswinds on cyclists: An
experimental study' Procedia Engineering, vol 72, pp. 720-725. DOI: 10.1016/j.proeng.2014.06.122
51 Brownlie L, Ostafichuk P, Tews E, Muller H, Briggs E , Franks K, The wind-averaged
aerodynamic drag of competitive time trial cycling helmets, 8th Conference of the International
Sports Engineering Association (ISEA), 1877-7058 c 2010 Published by Elsevier Ltd.
https://www.sciencedirect.com/science/article/pii/S1877705810002638
52 TRL (2009) ‘TRL Report PPR 445: Collisions involving cyclists on Britain’s roads: establishing the causes’.
www.trl.co.uk/online_store/reports_publications/trl_reports/cat_road_user_safety/report_collisions_inv
olving_pedal_cyclists_on_britain_s_roads_establishing_the_causes_.htm
53 Zang, K.; Shen, J.; Huang, H.; Wan, M.; Shi, J. Assessing and Mapping of Road Surface
Roughness based on GPS and Accelerometer Sensors on Bicycle-Mounted
Smartphones. Sensors 2018, 18, 914.
54 Robinson DL; Head injuries and bicycle helmet laws; Accid Anal Prev, 28, 4: p 463-475, 1996 http://www.cycle-
helmets.com/robinson-head-injuries.pdf
55 Clarke CF, The Case against bicycle helmets and legislation, VeloCity Munich, 2007.
http://www.ta.org.br/site/Banco/7manuais/colin_clarke_cycle_helmet.pdf
56 National Cycling Participation Survey 2017, Austroads and Australian Bicycle Council
https://www.onlinepublications.austroads.com.au/items/AP-C91-17 http://www.cycle-
helmets.com/cycling-1985-2019.html
57 Road safety: serious injuries remain a major unsolved problem, B. Beck, P.A. Cameron, M.C.
Fitzgerald, R.T. Judson, W. Teague, R.A. Lyons, B.J. Gabbe, Medical Journal of Australia, 2017
https://www.mja.com.au/journal/2017/207/6/road-safety-serious-injuries-remain-major-unsolved-
problem
58 Hospitalised Injury, Bureau of Infrastructure, Transport and Regional Economics, Australian
Government Department of Infrastructure, Regional Development and Cities -
https://bitre.gov.au/publications/ongoing/hospitalised-injury.aspx