Methods

The project was conducted in the community of East York, in collaboration with the East York Health Unit. East York is a health district within Metropolitan Toronto in south central Ontario with a total population of approximately 100 000, and a school aged population of approximately 10 000. For census purposes, East York had been divided into 21 census tracts. The census tracts were grouped into seven areas which were geographically and socioeconomically distinct.

Child cyclists were observed at all 111 preselected sites in the late spring and summer of 1993–97 and in 1999. Observational sites included the schoolyards of elementary and middle schools (kindergarten to grade 8), parks, major intersections, and residential streets. Sites from all seven areas were included. Observers were trained to collect reliable observational data. We have previously shown that observers can be trained to make reliable observations.⁹ The same standardized observation form was used in all years. Children were included in the observations if they were between the ages of 5 and 14 years and were riding a two wheeled bicycle. Observations of adults or young children, or of children riding any other type of vehicle were excluded. The observers remained at each site for one hour, counted the number of child cyclists, and noted whether or not the cyclist was helmeted.

For the purposes of the current study, the number of children observed cycling in each area was divided by the number of observation hours, resulting in the calculation of cyclists per hour. The rate of cyclists per hour was calculated for each of the types of observation area (schools, parks, residential streets, and major intersections), as well as an overall rate for each year. A general linear model, using Tukey's method, was used to compare the rate of cyclists per hour for each year. Tukey's method is used to test all pairwise comparisons among means, and to calculate confidence intervals for the differences between means.¹⁵ Statistical significance was considered as a p value of less than 0.05. All analyses were carried out using SAS¹⁶ software.

Results

The rate of child cyclists per hour varied over the six years in which it was recorded. Table 1 presents the number of children observed each year, and the rate of cyclists per hour for each type of site. Figure 1 depicts the rate of cyclists per hour for all sites combined. The lowest cycling rates were recorded immediately before the introduction of legislation in 1995 (4.32 cyclists per hour), and 1997 (4.57 cyclists per hour). In 1996, the year after legislation came into effect, cycling levels were significantly higher (6.84 cyclists per hour) than in 1995 (difference between means = 2.52; 95% confidence interval (CI) 0.09 to 4.95). The highest rate of cyclists was observed in 1999 (10.07 cyclists per hour), significantly different from all other years.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Changes in total cycling rate over time.

The cycling rate varied depending on the observation area, as shown in table 1. Schools and parks had the highest rate in most years, while residential streets and major intersections tended to have lower rates. The rate of children cycling to school did not change significantly throughout the study period, but in the year after legislation was introduced (1996), there was a significant increase in the number of children cycling in parks compared with 1995 (difference between means = 5.96; 95% CI 0.31 to 11.60). The rate of children seen cycling at major intersections declined slightly from 1995 to 1996 (difference between means = −0.13; 95% CI = −2.72 to 2.47).

Discussion

In this longitudinal observational survey, there was variability in the patterns of children cycling from year to year. The introduction of legislation did not appear to have an independent effect on children's cycling. We speculate that the variation in children's cycling may be related to the weather, changes in the community (for example, cultural composition, economic situation), or random variability of cycling at the observation sites.

A significant change in the location of child cyclists was observed the year after the introduction of legislation. In the year after the introduction of the law, children were more likely to cycle in parks, and somewhat less likely to be observed at major intersections. It is possible that increased public awareness may have been stimulated by the accompanying public debate. This may have led parents to encourage their children to cycle in parks, a location likely to be perceived as safer than main streets. The change in location, however, did not appear to be sustained over time.

Ecologic studies such as this one, may be limited due to attributing causal relationships at the ecologic level to the individual (the ecologic fallacy). However, failure to find a significant association between the rate of cyclists per hour and the introduction of legislation, even at the ecologic level, leads us to conclude that a cause and effect relationship between mandatory helmet legislation and a reduction in children's bicycling is unlikely.

Legislation, as previously reported,¹⁰ resulted in an increased rate of helmet use among children in East York. Contrary to the findings in Australia,¹⁴ we did not find evidence in our community that the introduction of mandatory helmet legislation had a significant negative impact on children's exposure to cycling. These findings will be important when interpreting the influence of helmet legislation on childhood cycling related injury rates.

Implications for prevention

This study provides evidence that the introduction of mandatory helmet legislation does not lead to reduced exposure to cycling. Policymakers considering the adoption of mandatory helmet legislation in their jurisdiction may find this evidence useful.