Cycle proficiency training was associated with some cycle-related safety behaviours in adolescence, and with cycling to school. However, it was not associated with self-reported involvement in a road traffic accident as a cyclist. A very small number of children had been admitted to hospital due to a cycle-related injury, preventing analysis of this outcome by cycle proficiency training status. Bike ownership was very common in our sample of adolescents yet less than half had received NCPS training, a proportion similar to that observed nationally [
12]. NCPS training was not available in all schools, and this remains the situation today. It is estimated that in 2013/2014 only around half of primary schools offered Bikeability training, and these schools are not evenly distributed across the country [
19].
Ownership and use of safety equipment
Children who attended an NCPS course were more likely to own a cycle helmet in adolescence, and to have used a helmet on their last cycle at age 14, than those who had not attended a course. This is consistent with a previous study of adolescents in their first year of secondary school, which found those who had had NCPS training were more likely to own and use a cycle helmet [
6]. Head injuries are the greatest health risk posed to cyclists [
5], and there is substantial evidence that helmets reduce the risk of head, brain and facial injuries, and death, in cyclists involved in a crash [
20,
21]. In the UK, there is no law compelling cyclists of any age to wear a helmet, although the Highway Code states that cyclists should wear one [
22]. The relatively high level of helmet ownership in those who attended training in our sample may reflect mandatory helmet use during training in some schools, but we have no information on this.
At 14 years those who had attended training were more likely to have worn a helmet, suggesting a beneficial impact of training on helmet use that persists for at least a couple of years. In contrast, a previous study that used the ALSPAC sample to evaluate a wide-ranging safety training programme, where cycle safety is one of many topics covered, found no association between attendance and cycle helmet ownership or use [
23]. This suggests the NCPS, with its practical nature and sole focus on cycle safety, is a more effective way of promoting some cycle safety behaviours. The increased helmet usage associated with NCPS training could be due to higher levels of knowledge of the safety benefits in trained children. However a study based in Oxford, a city with a high rate of cycling [
24], found that almost all teenagers knew wearing a helmet reduced the risk of head injury [
25]. Health behaviour models have long indicated that knowledge, while necessary, may not be sufficient to result in behaviour change as it competes with other barriers, such as social and behavioural norms [
26,
27]. In the Oxford study for example, over 70 % thought helmets ‘looked ridiculous’, and many said their friends ‘discouraged them from wearing one’ [
25]. The positive impact of the NCPS on helmet use could perhaps result from helping establish such behaviour as normal in a peer group. By 16 years, helmet use in those who had not attended training had reached similar levels to those who had. However, irrespective of training attendance, less than half of those who owned a helmet wore it on their last cycle ride. In concordance with this, a study which assessed children’s practical cycling skills found only 27 % of those who owned a helmet brought it with them for the test [
6]. In our sample, there was a positive association between distance last cycled and helmet use on that cycle; this suggests people may wrongly assume short journeys are safer. Furthermore, those whose last cycle was a longer time ago were more likely to have worn a helmet, perhaps indicating that those who cycle regularly perceive cycling as safer than those who cycle less regularly.
Use of high-visibility clothing was very low overall, although higher at 16 years than 14 years. This may reflect older adolescents being more likely to cycle for transportation, to cover longer distances, and to be on the roads in low light or dark conditions. Those who had attended cycle proficiency training were more likely to have worn reflective or fluorescent clothing, particularly at 16 years. At 14 years, those whose last cycle was a longer time ago were less likely to have worn high-visibility clothing; less frequent cyclists may be less likely to own such equipment. In the UK, there is no legal requirement for cyclists to wear high-visibility clothing but the Highway Code states cyclists should wear light-coloured or fluorescent clothing in daylight and poor light, and reflective clothing and/or accessories (e.g., ankle bands) in the dark [
22]. High-visibility clothing aims to increase cyclists’ visibility and conspicuity on the road, and hence reduce collisions, which are often due to car drivers failing to see a cyclist in time [
28]. The benefits of reflective versus fluorescent clothing depend on the time of day; fluorescent clothing is a useful daytime visibility aid but is of little use in darkness, conversely reflective clothing can greatly improve visibility at night [
29,
30]. Although many cyclists are aware of the visibility benefits of such clothing, studies have found low usage in both adults and children [
29,
31,
32]. This may be partly due to cyclists, like pedestrians, not fully appreciating their ‘invisibility’ relative to cars [
28]. Conversely, there is a danger that those who do wear such clothing over-estimate the effect it has on their visibility, particularly those wearing fluorescent clothing at night [
29,
30].
In focusing on low usage of helmets and reflective/fluorescent clothing in our sample, it is worth noting that in countries with high rates of cycling (e.g., the Netherlands, Denmark, and Germany), the use of such protective equipment is low [
33]. The safety focus in these countries is on the physical separation of bikes and cars on busy roads and intersections, together with traffic calming measures in residential areas [
33]. Additionally, there is evidence that the use of such safety equipment does not improve driver behaviour. One study found no evidence that use of high-visibility clothing changed drivers overtaking behaviour [
34], and the impact of high-visibility clothing on cyclist safety is unknown [
35]. Furthermore, drivers have been found to leave less space when overtaking cyclists wearing helmets [
36]. It is therefore important that young cyclists are not given a false sense of security from the use of such safety equipment.
Cycle accidents
It has been argued that accidents are too rare to use as an outcome to assess the efficacy of cycle training courses [
12]. In our sample, very few of the adolescents reported being involved in a road traffic accident as a cyclist. However, figures from Public Health England (2012/13) indicate that admissions to hospital via A&E for non-vehicular cycle accidents outnumbered cycle accidents involving another vehicle by approximately 7 to 1 for 10–13 year olds. This suggests that the ALSPAC participants may have been involved in many more accidents than those reported, if the question was interpreted as referring only to accidents involving a vehicle. In support of this, in a previous evaluation of the NCPS, over half of the accidents did not involve another vehicle (i.e., the child ‘just fell off’) [
6]. However, the total number of accidents was also small in this study and precluded any detailed analyses. Another NCPS evaluation reported much higher numbers of accidents: when 15 year olds were asked if they had ever been in a cycle accident, almost 18 % reported that they had been in one which required attendance at hospital, 45 % had had a minor accident, and 60 % had had a near miss while cycling [
12]. These higher percentages are likely due to the time frame being ‘ever’ as opposed to the previous 12 months as in our study, and the definition of accidents not being limited to road traffic accidents. Nevertheless, consistent with our study, boys were more likely to report an accident than girls, and there was no association between NCPS training and involvement in an accident. In considering whether training can have a positive impact on prevention of accidents, it is important to consider the potential for unintended negative consequences; an Australian study found that trained children, particularly boys, were more likely to be injured when cycling than untrained children, perhaps due to increased risk taking and reduced supervision [
37].
Cycling to school
Active transport to school is an important source of physical activity in young people, and cycling to school has been shown to improve the cardiovascular health of children and adolescents [
38‐
40]. Cycling also has wider benefits, including being environmentally friendly, economical, and giving young people independence [
33]. A key aim of the NCPS was to ‘promote and encourage’ cycling [
9], and the UK government have stated that they want to ‘encourage more people to cycle more safely and more often’ [
41]. We found that children who had attended training were more likely to report cycling as part of their school commute at both 14 and 16 years. Evaluations of the impact of Bikeability on cycling frequency have had mixed results. One study found that in local authorities in England where there is a longer history of delivering cycle training to primary school children, a higher proportion of children cycle to secondary school [
42], and a further study found children and their parents both reported that the children cycled more often after they received training [
4]. However, two studies that compared cycling rates in trained and untrained children did not have such positive results. One used two datasets to assess children’s cycling behaviour, and found a positive association with Bikeability training in one dataset but not the other [
19]. The other study found that two months after the course finished, the trained children did not report cycling more frequently than the untrained children, despite reporting that training had made them feel more confident about cycling on the road [
13].
Very few of the girls in our study cycled to school, and they reported that their last cycle was shorter and a longer time ago than the boys. This is consistent with previous studies [
12,
43‐
46], and is also the gender pattern observed in adults in countries with a low cycling prevalence [
47‐
49]. Barriers and incentives to cycle likely differ by gender and age. For example, boys who cycle to school may be attracted to cycling not just as a mode of transport but as a physical activity in its own right [
46]. In contrast, perceived safety issues may be of more concern to girls, as indicated by a study in Melbourne suggesting that female cyclists prefer to use routes with maximum separation from motorized traffic [
50].
In the 2014 National Travel Survey in England, only 1 % of all journeys were made by bike [
49] (as way of comparison, over a quarter of all journeys are by bike in the Netherlands [
51]). There is clearly a long way to go in establishing cycling as a viable alternative to other forms of transport in the UK for both children and adults. It is worth highlighting that ALSPAC is based in and around Bristol, a city with relatively high rates of cycling by UK standards [
24]. Bristol was designated a ‘cycling city’ by the Government in 2008, the first UK city to be given this status, and consequently awarded £22 million to invest in cycling [
52]. Since 2009 the city council have measured cycling behaviour in its annual ‘Quality of Life in Bristol’ survey, and in 2014 noted an increase for the first time: 24 % of respondents cycled at least once a week (versus 19 % in 2009), and 16 % cycled to work (versus 9 % in 2009) [
53]. However, rates of cycling vary substantially between different parts of the city: over 40 % of those in some areas cycle at least once a week, compared to less than 9 % in other areas [
53]. The contextual effect of living in an area where there is appropriate infrastructure to enable safe cycling, and where travelling by bike is viewed positively and as a social norm, could ultimately have as much influence on whether or not adolescents cycle, and their cycling behaviours, as do cycle training courses. However, it can be argued that a high rate of cycling, as in the Bristol area, renders cycle training especially important as there is some evidence, in the comparable area of car driving, that young people may acquire poor safety habits from observing their parents’ everyday behaviours [
54].
Limitations
Our study has limitations and the results should be interpreted in light of these. None of the ALSPAC measures were designed
a priori to evaluate the NCPS. The children were not asked for any details of the training they had undertaken, therefore we were unable to consider whether outcomes differed between those who were trained in a playground compared to those who trained on roads, for example. We have no measure on how the child performed during their training or whether they passed their Cycle Proficiency Test (although pass rates are generally high [
12]). The measures that relate to use of a cycle helmet and fluorescent/reflective clothing refer only to the last time cycled and may not reflect typical use. We have no information on the purpose of that last cycle ride, or the time of day. Participants were asked about accidents in the 12 months prior to data collection; as the period between the two outcome time points was over two years, we will have missed some accidents. We had no information on near misses while cycling: near misses are markedly more frequent than collisions and can have a substantial impact on an individual’s cycling experience [
55,
56]. Our hospital admissions outcome probably only captures individuals who sustained the most serious injuries; more minor cycle injuries would have been treated in A&E but a lack of detail in admission codes prevents us from being able to determine which visits to A&E were the consequence of cycle-related injuries. We cannot rule out reverse causation as an explanation for the associations observed as we have no measure of bike use or of cycle safety behaviours in earlier childhood. This was also a limitation of a recent Bikeability evaluation, and signals that more data is needed on the differences between children who do and do not receive cycle training [
19]. Finally, our sample undertook their training in approximately 2002–2004; the NCPS has since undergone some changes and been rebranded to Bikeability (although its core aims remain the same) [
8].