|Home | About | Journals | Submit | Contact Us | Français|
The association of demographic, socioeconomic, and attitudinal factors with children's cycle‐helmet use was investigated using self‐reported data from 9775 children aged 8–12 in England, where there are currently no legal requirements that cyclists (of any age) wear helmets. The results suggest that demographic factors, socioeconomic status, and attitudes influence whether children wear bicycle helmets but that attitudinal factors are most important.
Wearing a helmet when cycling can reduce mortality related to collisions1 and lowers the risk of head, brain, and face injuries.2,3,4,5,6 The protective effect of helmets is greatest in children, who have most cycling‐related head injuries.5,7 Legislation has proved highly effective in increasing helmet use,8 and interventions have been identified that significantly improve helmet‐wearing.9,10 In the absence of either legislation or intervention, helmet use has been found to be associated with parental encouragement,11 peer influences,11,12 children's attitudes,13,14,15,16 and levels of neighborhood deprivation.10,17,18
In this study, I assessed the combined effects of demographic, socioeconomic, and attitudinal factors on cycle‐helmet wearing in children aged 8–12 in the absence of legislation or local ordinances on helmet wearing. I used nationally representative cross‐sectional data to compare the effects of demographic, socioeconomic, and attitudinal factors, and to assess the strength of their relationship with children's cycle‐helmet use.
Data for this study came from consecutive waves of the Health Survey for England (HSE) from 1997 to 2004. The HSE is a nationally representative survey of individuals aged 2 and over living in households in England, the aim of which is to provide information about health and to assess progress towards health targets. HSE data are publicly accessible from the Data Archive of the UK Economic and Social Data Service (www.esds.ac.uk). From 1997 to 2004, 13985 children aged 8–12 participated in the HSE and were asked questions relating to cycle‐helmet use. I used data on 9775 of these children in this study. Figure 11 gives a flow chart showing how respondents were selected for inclusion.
I classified explanatory variables into three categories: demographic, socioeconomic, and attitudinal. Demographic variables were: age (in years); gender; degree of urbanization (urban, suburban, rural; categorized by interviewers when visiting households). Socioeconomic variables were occupational social class of the head of household in six categories (professional; managerial and technical; skilled non‐manual; skilled manual; semi‐skilled manual; unskilled) and standardized equivalized household income split by deciles (categories: £34211 and above; £24595–34210; £19439–24594; £15690–19438; £12466–15689; £10196–12465; £7685–10195; £5532–7684; £4160–5530; £0–4159).
As part of the HSE, children were asked to respond to six statements about cycle helmets, and the attitudinal variables are the results of these (all answered yes/no):
Respondents who did not own, or who never rode, a bicycle were excluded from analysis (fig 11).). Those who cycled were asked whether they wore a helmet when they rode; possible responses were “always”, “sometimes”, and “never”. A subsequent question asked respondents whether they sometimes forgot to put their helmet on. I constructed a binary outcome coded 1 for those who reported that they always wore a helmet and never forgot to put it on, and 0 for those who sometimes or never wore a helmet, or sometimes forgot their helmet.
Logistic regression was used to estimate the odds ratios of always wearing a helmet associated with the explanatory variables. Because the aim was to compare the effects of groups of variables (socioeconomic and attitudinal) rather than assess the explanatory value of individual independent variables, I entered variables into the model in blocks to enable comparison of their effects and explanatory value. I controlled for clustering within households and within each round of the survey and conducted all analyses using Stata SE V9.2.
Table 11 shows the demographic and socioeconomic characteristics of the children involved in this analysis. Table 22 shows percentages of children who reported that they always wore a helmet in relation to demographic, socioeconomic, and attitudinal variables, plus the associated unadjusted odds ratios (ORs). Overall, 17.1% (95% CI 15.7% to 18.6%) reported that they always wore a helmet when they rode a bike. In unadjusted models, all six attitudinal variables were statistically significantly associated with helmet wearing.
Table 33 shows logistic regression results. Model 1 contains demographic and socioeconomic variables and shows a statistically significant reduction in the OR of wearing a helmet associated with age, and a statistically significant increase associated with being female (compared with male) or living in a rural area (compared with an urban one). The OR of wearing a helmet decreased as the occupational social class of the household reference person declined. Compared with a child in a household with a professional head of household, the estimated OR for a child who lives in a household where the head of household was unskilled was 0.38 (95% CI 0.25 to 0.57). Each 10% reduction in equivalized household income was associated with a 9% reduction in the chances of reporting always wearing a helmet (OR 0.91, 95% CI 0.90 to 0.93).
Model 2 includes demographic, socioeconomic, and attitudinal variables. In this model, the effect of age was no longer statistically significant. In this adjusted model, statistically significant differences in the ORs of reporting wearing a helmet were associated with all attitudinal variables except “Helmets look good”. The largest effect was associated with the statement “Helmets make me feel safer when I ride a bike”, to which an answer “Yes” compared with “No” was associated with an OR of 7.37 (95% CI 5.42 to 10.03).
The difference in χ2 (and pseudo‐R2) between model 1 and model 2 was markedly greater than that between model 1 and the null model, suggesting that attitudinal variables provide more explanatory power than demographic and socioeconomic variables alone, and that this additional explanatory power is additional to that of the demographic and socioeconomic variables. This is the case when controlling for the demographic and socioeconomic variables, as in table 33,, and when the attitudinal variables are entered separately (results not shown).
These findings suggest that, in the absence of relevant legislation, there are demographic, socioeconomic, and attitudinal factors associated with whether children aged 8–12 report wearing helmets when cycling. However, it appears that the largest proportion of the differences in helmet use is explained by attitudinal variables. These findings are in keeping with previous findings that helmet use is less common in children from deprived neighborhoods,10 and with observations in smaller populations that children's attitudes are important determinants of helmet wearing.14,16 I have built on these previous studies by comparing the effects of demographic, socioeconomic, and attitudinal variables in a single model, using data from repeated waves of a large‐scale nationally representative survey.
Some methodological issues should be borne in mind. I used self‐reports of helmet wearing, and both the independent and dependent variables are based on what children (or their guardians, for the socioeconomic variables) reported about themselves. In general, children and adolescents have been found to be good informants on their behaviors, including problem behaviors.19 Studies of helmet wearing rely on either self‐reported or observed behavior,9 and there is little published information about which of these measures has greater reliability and validity. Nevertheless, I cannot exclude the possibility that children who over‐report risk‐averse attitudes may also over‐report risk‐averse behaviors.
Because the findings are from cross‐sectional data, caution must be exercised in drawing causal inferences. Longitudinal data would enable us to assess whether changes in attitudes over time (which are implied by the differences in attitudes at different ages) bring about changes in behavior.
Legislation is a proven way of increasing rates of cycle‐helmet use. As long as legislation remains politically unpalatable, as is currently the case in England, attempts to understand motivations for helmet wearing remain important. These findings suggest that promotion efforts should include consideration of how to bring about change in children's attitudes to cycle helmets and helmet‐wearing.
It was found that reported cycle‐helmet use by children aged 8–12 was more associated with their attitudes to cycle helmets than with socioeconomic factors. Age did not have a significant effect on helmet use when attitudes were controlled for. In the absence of appropriate legislation, interventions to promote cycle‐helmet use by children should take into account their attitudes to helmets.
I thank the following people for their helpful comments on earlier versions of this paper: Sara Gibbs; William Henley; George T H Ellison.
The author is a SpT in Public Health funded by the NHS South West Public Health Training Scheme.
Competing interests: None.