Cycling in Canada
by Pamela L. Ramage-Morin
The health benefits of physical activity, including cycling, are widely recognized.Note 1Note 2Note 3 In an era when nearly a third of children and youth and just under two-thirds of adults are overweight or obese,Note 4Note 5 cycling for leisure or transport is a valuable form of exercise. Cycling is also good for the environment―commuting by bicycle helps to alleviate road congestion and noise pollution and reduces emissions.Note 1Note 3
But while the benefits of cycling are acknowledged, so are the dangers. Cyclists are vulnerable in the event of a crash. Strategies to protect cyclists include infrastructure such as bicycle paths, dedicated bike lanes and traffic calming; side guards for heavy trucks; driver behaviour, with an emphasis on sharing the road; and cyclist behaviour, including increased visibility and helmet use.Note 6Note 7Note 8Note 9Note 10 The last has been debated extensively. Some resist legislated helmet use, at least for adults, on the grounds that helmets offer minimal protection and encourage risk-taking, and that such legislation impinges on personal freedom and reduces ridership; others dispute these claims.Note 9Note 11Note 12Note 13Note 14Note 15 Medical, public health and other sectors recommend that all-age helmet use be legislated and enforced across Canada.Note 16Note 17Note 18Note 19Note 20
This analysis examines bicycle and helmet use among people aged 12 or older in 1994/1995 and in 2013/2014, and cycling fatalities during the 1994-to-2012 period (see The data). Other behaviours of helmet users and non-users are compared, specifically, seatbelt use, smoking, flu shots and drinking.
Characteristics of cyclists
In 2013/2014, an estimated 12 million Canadians (41%) aged 12 or older reported that they had cycled in the previous year (Table 1). Cycling was more common at younger ages―82% among 12- to 14-year-olds versus 27% by age 50 or older―and among people in higher-education and -income households. The association with education and income may reflect residence in neighbourhoods characterized by bicycle paths, traffic-calming measures and lower perceived levels of crime, and workplaces with facilities (for example, bike racks) and policies such as flexible work hours that enable active commuting.Note 21Note 22Note 23
Males were more likely than females to have cycled in the past year (46% versus 34%), regardless of age, income or education. Males residing in population centres were more likely than those in rural areas to cycle; the opposite was true for females, who were more likely than males to report excessive traffic as a barrier to cycling.Note 24
The decrease in cycling was evident in most age groups. Cycling was more common in Quebec (48%) and Manitoba (46%) than in the rest of Canada. It was less common in the Atlantic Provinces (from 18% in Newfoundland and Labrador to 32% in New Brunswick), Saskatchewan and Ontario (both 38%), and Nunavut (23%).
Comparison with 1994/1995
In 2013/2014, an estimated 7.0 million people reported that they had cycled in the past 3 months, up from 6.5 million in 1994/1995. Despite this numerical increase, cyclists comprised a diminishing percentage of the population: 24% in 2013/2014 versus 29% in 1994/1995 (Figure 1). Even when the aging population was taken into account, the decrease persisted: if the age structure had remained unchanged during the two decades, an estimated 25% of the population would have reported cycling in 2013/2014.
Other activities may have compensated for this apparent reduction in physical activity. For example, the percentage of the population who reported running or jogging in the past 3 months almost doubled from 14% (95% CI: 14 to 15) in 1994/1995 to 27% (95% CI: 26 to 27) in 2013/2014; increases were evident in every age group.
The decline in cycling raises safety concerns, because areas where cycling is more prevalent tend to have lower rates of bicycle-related collisions and fatalities―a “safety in numbers” effect.Note 3Note 25 Teschke et al.Note 26 found an association between higher “mode share” (percentage of commuters using bicycles) and fewer traffic-related injuries.
From 1994 through 2012, 1,408 cyclists died in crashes, an average of 74 each year (Figure 2). MillarNote 27 reported 1,665 cyclist deaths between 1980 and 1994, an average of 111 annually. At that time, 57% of cyclist deaths involved people younger than age 20,Note 27 compared with 30% of the more recent deaths. Most of those killed were male―around 80% in the earlier periodNote 27 and 84% in 1994 through 2012.
The fatality rate for cyclists was 2.6 deaths per million population in 2012. This was lower than rates for pedestrians and motor vehicle occupants―9.0 and 43.0 deaths per million population, respectively.Note 28 Annual cycling fatality rates fluctuated during the 1994-to-2012 period, from a low of 1.6 deaths per million population (2003) to a high of 2.9 deaths per million population (1994), but remained below rates in the 1980s, which ranged from 3.0 to 4.5 per million population.
While useful for comparisons, rates based on the entire population are not indicative of the risk specific to cyclists. Based on the number of person-trips by mode of travel, Beck et al.Note 29 reported fatality rates for cyclists, pedestrians and motor vehicle occupants at 21.0, 13.2 and 9.2 per 100 million person-trips, respectively.
Cycling deaths can be prevented, according to the Ontario Office of the Chief CoronerNote 6 after a detailed review of deaths from 2006 through 2010. In addition to recommendations for improved road infrastructure, side guards for heavy trucks, legislation and enforcement of road safety laws, mandatory helmet use was proposed, a position supported by medical professionals and other interest groups.Note 17Note 18Note 19Note 20 It has been argued that cyclists of all ages require head protection, and that compliance is higher among younger riders when they observe adult cyclists wearing helmets.Note 30
In 2013/2014, of the estimated 12 million cyclists aged 12 or older, 5 million (42%) reported “always” wearing a helmet (users); the remaining 7 million did so “most of the time,” “rarely” or “never” (non-users) (Table 2).
Although females were less likely than males to cycle, they were more likely to wear a helmet―46% compared with 39%. Higher helmet use (47%) in the early teen years coincided with the ages when cycling prevalence was highest. These are also the years when parents and caregivers may find it easier to enforce helmet use, which is legislated in several provinces, territories and cities.Note 26 However, in the later teen (15 to 17) and early adult (18 to 24) years, helmet use dropped to 28% and 25%, respectively. At older ages, the percentage rose again―49% of cyclists aged 50 or older “always” wore a helmet.
Twenty years ago, helmet use was much less common. In 1994/1995, 19% of those who had cycled in the past 3 months “always” wore a helmet, compared with 45% in 2013/2014 (Figure 1). The increase among 12- to 14-year-olds was almost threefold―from 18% to 50%―which may reflect, at least in part, introduction of bicycle helmet legislation.Note 31
People who commuted by bicycle were more likely than leisure-time-only cyclists be helmet users―48% (95% CI: 45 to 51) versus 44% (95% CI: 43 to 46). Even when age and sex were taken into account, the odds of commuters wearing a helmet were higher than the odds for non-commuters (1.3; 95% CI: 1.2 to 1.5). Cyclists who lived in population centres were more likely than those in rural areas to wear helmets, which may reflect concerns about traffic safety and aggressive drivers (Table 2).Note 24
Other precautionary behaviours
It has been suggested that people who use protective equipment such as bicycle helmets tend to take more risks, a concept known as “risk compensation.”Note 12Note 32 Alternatively, helmet use may indicate a person who is more cautious, and therefore, less likely to take risks and experience injuries.Note 33 Data for 2013/2014 support the latter: cyclists who “always” wore a helmet were more likely than other cyclists to always use a seatbelt when driving or as a passenger (90% versus 76%) and to have had a flu shot in the past year (33% versus 20%) (Table 3). Helmet users were less likely than non-users to smoke (10% versus 21%) or to engage in heavy episodic drinking (17% versus 27%). Associations between helmet use and these other behaviours remained even when age, sex and education were taken into account. The tendency for helmet users to adopt other safety practices was also reported by Bolen et al.,Note 33 who found that they were more likely than non-users to have a household smoke detector and a fire escape plan, and to use seatbelts.
In 2013/2014 an estimated 41% of Canadians reported that they had cycled in the previous 12 months. Although the number of cyclists increased between 1994/1995 and 2013/2014, the percentage of the population who cycled decreased. Cycling fatalities averaged 74 each year during the study period, and fatality rates fluctuated between 1.6 and 2.9 deaths per million population. Helmet use among past 3-month cyclists more than doubled from 19% to 45% during the two decades, with significant increases in every age group. Helmet users were more likely than non-users to engage in what are generally considered precautionary behaviours.
Start of text box
This study is based on cross-sectional data from the 1994/1995 National Population Health Survey (NPHS) and the 2013/2014 Canadian Community Health Survey (CCHS). The NPHS covers household residents in all provinces, except people living on Indian reserves, on Canadian forces bases, and in some remote areas. CCHS coverage includes the territories, with exclusions similar to the NPHS. The respective household- and person-level response rates are 88.7% and 96.1% (NPHS), and 75.9% and 87.3% (CCHS). Study samples comprised 17,626 (NPHS) and 128,310 (CCHS) respondents. Death data are from the Vital Statistics–Death Database. Details are available at www.statcan.gc.ca.
- Cyclist (past 3- and 12-month) included bicycle use for leisure or transport. Commuters had cycled to school or work in the past 3 months versus those who cycled only during leisure time.
- Helmet users “always” (versus ‘’most of the time,” “rarely” or ‘’never”) wore a helmet when cycling.
- Education (less than postsecondary graduation or postsecondary graduation) and income (quintiles) were estimated at the household level.
- Population centres are areas with a population of at least 1,000 and apopulation density of 400 or more per square kilometre. All areas outside population centres are defined as rural.
- Seat belt use as a driver or passenger was classified as “always” versus “most of the time,” “rarely” or “never.” This CCHS 2013/2014 optional content was selected by Ontario, Manitoba, Alberta, Yukon and Nunavut.
- Current smokers (daily or occasional) were grouped together, distinct from former smokers or those who had never smoked.
- Heavy episodic drinking for respondents aged 18 or older was consuming 4 (female)/5 (male) or more drinks on one occasion.
- Received a seasonal flu shot was classified as “within past year” versus “one year ago or longer” or “never.”
- Cyclist deaths in 2000 through 2012 were identified by codes V10 to V19 in ICD-10.Note 34 Before 2000, ICD-9 codes were E810 to E819 (with .6 to identify the injured person as a pedal cyclist), E820 to E825 (with .6) and E826.1.Note 35 Deaths of non-residents of Canada were excluded.
Weighted frequencies and cross-tabulations were used to examine the prevalence of cycling and bicycle helmet use and factors associated with both. For consistency, comparisons between the NPHS and CCHS were limited to respondents aged 12 or older living in the ten provinces. The 2013/2014 prevalence of 3-month bicycle use was age-standardized to an average of the 1994 and 1995 populations to assess whether the change over time was related to the aging population. Regressions that modelled behaviours on helmet use controlled for sex, age, age2 (to account for the non-linear relationship between age and helmet use) and education. Cyclist fatality rates were age-standardized to the 2012 population. Standard errors and coefficients of variation were estimated using the bootstrap technique to account for survey design effects.
The cross-sectional data preclude an assessment of behaviour changes over time, and therefore, have limited utility for questions about how legislation affects helmet use and ridership. Comparisons between 1994/1995 and 2013/2014 were limited by the NPHS questions, which pertained to past 3-month cycling in the 10 provinces. Respondents were categorized based on residence (population centre or rural), which was not necessarily the area where they cycled. Estimation of a cycling-specific fatality rate is not possible without additional data (for example, number of cyclists, distance cycled, person-trips).
End of text box
- Date modified: