Objective: To determine changes in helmet use in cyclists following the introduction of a bicycle helmet law for children under age 18.
Methods: Cyclists were observed by two independent observers from July to August 2004 (post-legislation) in Edmonton, Alberta. The data were compared with a similar survey completed at the same locations and days in July to August 2000 (pre-legislation). Data were collected for 271 cyclists in 2004 and 699 cyclists in 2000.
Results: The overall prevalence of helmet use increased from 43% (95% CI 39 to 47%) in 2000 to 53% (95% CI 47 to 59%) in 2004. Helmet use increased in those under 18, but did not change in those 18 and older. In the cluster adjusted multivariate Poisson regression model, the prevalence of helmet use significantly increased for those under age 18 (adjusted prevalence ratio (APR) 3.69, 95% CI 2.65 to 5.14), but not for those 18 years and older (APR 1.17, 95% CI 0.95 to 1.43).
Conclusion: Extension of legislation to all age groups should be considered.
- repeat cross sectional survey
- bicycle helmet
Statistics from Altmetric.com
According to 2004 Alberta Transportation data, there were 613 cyclists injured or killed on Alberta roadways and 40% of these were under the age of 20.1 Approximately 20% of cyclist emergency department visits are for head injuries;2 though the proportion can rise to over 75% for those fatally injured.3
Evidence indicates that bike helmets prevent head, face, and brain injuries.4,5 Two systematic reviews of the scientific evidence found that helmets reduce fatal injuries by 73%,4 and are effective even in those crashes involving motor vehicles, reducing the risk of head injury under these circumstances by 69%.5
Many studies have demonstrated a post-legislation increase in the proportion of helmeted cyclists.6,7,8,9,10,11,12,13,14,15,16 Among studies using direct observation of helmet use both pre- and post-legislation, only one involved age groups not affected by the law.17 That study included a helmet giveaway program and educational component, which may have confounded the legislative effect.
We were presented with an opportunity in Alberta to examine helmet wearing rates a full two years after helmet legislation was implemented in 2002 in age groups both covered (<18 years old) and not covered (18 years and older) by the law. This evaluation is important in order to inform other jurisdictions considering whether to implement legislation and to which age groups it should apply.
The first phase of this repeat, cross sectional survey was conducted in 2000 to estimate the prevalence of bicycle helmet use in the cities of Edmonton and Calgary, along with surrounding communities if they were within 50 km from either city center and their population exceeded 9500 (Airdrie, Cochrane, Okotoks, Fort Saskatchewan, Leduc, Sherwood Park, Spruce Grove, and St Albert).18 These locations were divided into six strata: schools, parks, commuter routes, designated cycling paths, universities/colleges, and residential areas. One observer at each site collected information on riding companionship, helmet use of riding companions (if any), helmet use, gender, and approximate age.
During the summer of 2004, two trained observers and a project coordinator visited the same sites, in Edmonton only, on the same weekday and at the same time periods used in 2000.18 Only those sites where at least 10 riders were observed in 2000 were visited by the team to maximize data collection efficiency. If data collection times overlapped, we chose the site where most observations were recorded.
Five sites had to be rescheduled due to inclement weather. We were unable to reschedule data collection for one site. No school observations were made in 2004 as data collection did not commence until July. Thus, in 2004 we collected data at 22 of the 23 eligible 2000 observation sites in Edmonton.
In 2004, the project coordinator directed observations on cyclists and pedestrians who passed by on the observers’ side of the street, but traveling in either direction. Data were collected on age (<6, 6–12, 13–17, 18–54, and 55+), sex, helmet use, travel mode, clothing visibility, observer assessed speed, and bicycle reflective devices. This report presents cyclist data only and does not include the additional information on clothing visibility, observer assessed speed, or reflective device use. Once finished recording observations, the observers would verbally cue the project coordinator who would then direct observations on the next subject (pedestrian or cyclist).
As two estimates of helmet prevalence were available in 2004, we randomly selected which observer’s data to use for each site. Interobserver agreement was assessed using Kappa (κ) for age category (<18, 18+), sex, and helmet use.
Change in helmet prevalence between 2000 and 2004 was examined by age, sex, location, and neighborhood average annual household income based on 2001 Statistics Canada census data divided into three strata: <$50,000, $50,000–$59,999, and $60,000+.19
Poisson regression was used to directly model the prevalence ratio,20 with the robust (sandwich or Huber-White) estimator to account for clustering by site.21 All analyses were conducted in Stata version 8.0.22 Main effects and interactions between year of observation (2004 = 1; 2000 = 0) and all other variables (age: <18 = 1, 18+ = 0; sex: male = 1, female = 0; location: campus, residential, cycle path, park; neighborhood average annual household income: <50,000, 50,000–59,999, 60,000+) were included. Interaction terms were simultaneously tested and those that were significant (p<0.05) were retained in the model.
Ethical approval was obtained from the University of Alberta Health Research Ethics Board. We notified the Edmonton Police Service about the study and provided a letter explaining the project details to concerned citizens in 2004.
Data were collected for 271 cyclists in 2004 and 699 cyclists in 2000. Helmet use was not recorded for two cyclists in 2000 and two cyclists in 2004 and so these observations were excluded from the analysis.
We noted almost perfect agreement between the two observers on cyclist helmet use (κ = 0.95; 95% CI 0.86 to 1.0) and sex (κ = 0.95; 95% CI 0.84 to 1.0). We also found substantial agreement for age categories (κ = 0.78; 95% CI 0.66 to 0.89).
Helmet use: unadjusted
Helmet use increased from 43% in 2000 to 53% in 2004 (table 1). Helmet prevalence increased threefold in those younger than 18; however, prevalence changed little in those 18 and older (from 49% to 48%). Helmet use was estimated to increase in residential areas and on commuter routes in both males and females, regardless of average annual income.
Helmet use: adjusted
Significant interactions were found for year of observation by age and year of observation by location. From table 2, after adjusting for sex, location, and average annual income, the prevalence of helmet use increased 3.7-fold from 2000 to 2004 for those younger than 18 years (95% CI 2.65 to 5.14). For those riders 18 years and older, the effect was less pronounced and not significant (adjusted prevalence ratio: 1.17; 95% CI 0.95 to 1.43). The prevalence of helmet use was also estimated to increase for residential neighborhoods, but not at other locations.
Our results demonstrate a significant increase in helmet prevalence following legislation in those cyclists under 18 years old and are consistent with the findings of other investigators.6,7,8,9,10,11,12,13,14,15,16 Bicycle helmet evaluations in other Canadian Provinces with all-ages bicycle helmet legislation have shown a consistent increase in helmet use across age groups.9,11 However, adjusting for trends in those aged 18 and older strengthens the findings. Considering that children riding with helmeted adults are almost 10 times more likely to be wearing a helmet than children riding with non-helmeted child companions,23 policy makers should consider extending current children-only helmet legislation in Alberta and other locations. Moreover, regions contemplating bicycle helmet legislation should use these results to argue for universal bicycle helmet legislation.
In the 2004 survey, the project coordinator directed observations on subjects (that is, selected which subjects to observe); however, this direction followed a standard protocol that precluded consideration of helmet wearing, but was simply based on the next subject to pass by the research team. This approach was necessary to ensure that each observer recorded information on the same cyclist in order to calculate agreement statistics. It seems unlikely that selection bias occurred under these circumstances.
There was likely some non-differential misclassification of neighborhood average annual income as we did not find that this variable modified the pre- to post-legislation assessment of bicycle helmet prevalence. This result conflicts with the findings of Parkin et al,15 who noted a greater post-legislation increase in the prevalence of helmet use for low and middle socioeconomic status (SES) areas compared with high SES areas. The discrepancy is likely explained by the more sensitive definition of SES units used by Parkin et al and the lack of heterogeneity of neighborhood average annual income levels in our investigation.
We adjusted for most variables that have been shown to influence helmet use; however, because we did not stop cyclists and collect other information regarding personal, trip, or cycling characteristics, these influences could not be excluded. However, because we compared the change in prevalence for those younger than 18 years affected by the helmet law with similar trends for adults, it is unlikely that our results could be explained by a concomitant increase in general safety. Public helmet awareness campaigns and targeted school health activities were implemented in the spring of 2004 (Kathy Holgate, KIDSAFE Connection, Stollery Children’s Hospital, personal communication), which may have increased helmet use independent of the legislative effect. Compared with legislation, though, education on its own has proven to be a less effective intervention.7 Enforcement activities also increased in 2004 as reflected in Edmonton Police Service helmet infraction data showing 16 bicycle helmet tickets in 2003 compared with 48 in 2004 (Nancy Leake, Edmonton Police Service, personal communication). We would argue that this still represents minimal enforcement for a large urban center. Therefore, the increase in helmet prevalence for children and adolescents is likely due to the legislation—not education or enforcement.
Because we captured more information (for example, clothing color) on each cyclist, did not capture information on all cyclists, but only those passing by the research team, and recorded information on pedestrians in 2004, we cannot comment on the number of cyclists seen in 2000 and 2004.
Implications for prevention
The introduction of helmet legislation restricted to youth in Alberta was associated with an increase in helmet use for this age group, but had little effect on adult riders, suggesting that adult legislation should be considered.
Many studies have demonstrated a post-legislation increase in the proportion of helmeted cyclists, but few have included age groups not affected by the law.
We noted substantial interrater agreement for helmet use, age, and sex.
Adjusted helmet prevalence estimates increased substantially after legislation for those under 18 years old (that is, those affected by helmet legislation), but changed little for adults (that is, those not affected by the legislation).
Adjusting the child helmet use trends for adult trends makes it unlikely that other factors besides the legislation caused the increase in helmet use seen in those under the age of 18 years.
The results suggest that adult legislation should be considered.
Jacques Rizkallah was funded by an Alberta Heritage Foundation for Medical Research Summer Studentship (AHFMR, Edmonton, AB). Andrea Lamy was funded by the University of Alberta Summer Temporary Employment Program and the Alberta Centre for Injury Control & Research. We would like to thank KIDSAFE Connection, the injury prevention program of the Stollery Children’s Hospital, for assistance with the conduct of this investigation and Terri Vaive of the Alberta Centre for Injury Control & Research for help with data collection. We also thank Dr Ambikaipakan (Sentil) Senthilselvan for helpful comments on an earlier draft of this manuscript. Dr Brent Hagel holds the recently appointed position of Professorship in Child Health and Wellness funded by the Alberta Children’s Hospital Foundation, through the support of an anonymous donor and Canadian National Railway Company. Dr Rowe is supported by a Canada Research Chair from the Canadian Institutes of Health Research (CIHR, Ottawa).
Competing interests none.
Field work conducted while Dr Hagel was with the Alberta Centre for Injury Control & Research, Department of Public Health Sciences, Faculty of Medicine and Dentistry, University of Alberta.
These results were presented, in part, at the Canadian Association of Emergency Physicians Annual Meeting in Edmonton, AB, 29 May–1 June 2005 the Canadian Injury Prevention and Safety Promotion Conference in Halifax, NS, 6–8 November 2005, and the 8th World Conference on Injury Prevention and Safety Promotion, Durban, South Africa, 2–5 April 2006.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.