Cook A. Significance 2004;Vol 1 issue 4:162-3
This paper poses the hypothesis that the main reason why helmet effectiveness is contentious is because it is a key issue in whether helmet use should become compulsory. Evidence supportive of helmet use is discussed, based on case-control studies and hospital admissions data. The possibility of risk compensation by helmet wearers is considered, but thought unlikely to be a significant factor due to the counter-evidence of large reductions in head injury shown by case-control studies.
In considering whether helmet use should be mandatory, the author notes that there were large reductions in cycle use in Australia when laws were first introduced, but suggests that the level of cycling recovered quickly. A Canadian paper is cited to suggest that children may be less deterred from cycling by helmet laws than adults. Legislation may be an overly paternalistic option, but there are precedents and considerations of autonomy are in any case weaker for children.
The author believes the evidence to be strong that cycle helmets offer significant protection and that uncertainties arising due to confounding and risk compensation are unlikely to alter this. Although it would be disastrous to deter cycle use, the author believes there is a strong case for legislation for children. However, cycling is fundamentally safe and serious injuries are rare, so the author says that people should cycle whether or not they choose to wear a helmet.
This paper examines very little evidence on cycle helmet effectiveness, and none that is not supportive of helmet use. The principal support for the author's argument that helmets provide useful protection comes from the work of a single team of researchers, although this is presented in a way that may suggest wider support than is the case. Thus the often-cited Seattle study (Thompson, Rivara and Thompson, 1989) is cited directly, in the guise of a Cochrane review (Thompson, Rivara and Thompson, 2002-9) (which was undertaken by the same researchers and whose own work dominated the review), and through the conclusions of a meta analysis (Towner et al, 2002) where the same researchers' papers were also dominant and no counter evidence considered. All of these references have been widely criticised for fundamental methodological errors and a lack of balance, but this is not mentioned in this paper.
The weakness of non-randomised case-control studies that do not take account of differences between cyclists other than helmet-wearing is mentioned, but the author gives no evidence for his assertion that helmet studies have been successful in adjusting for such errors, something that is much disputed. The magnitude of the effect claimed by the Seattle study is considered to be sufficient that residual confounding may be ignored, but this supremacy disappears completely if criticisms of that work are correct. No mention is made of the controversy that now surrounds case-control studies in other similar areas of medical research (BHRF, 1134).
In considering data based on hospital admissions, a Canadian study (Macpherson et al, 2002) is cited to the effect that head injury rates declined more in helmet law provinces than elsewhere. However, as a published response to that study (Robinson, 2003b) noted, most of the falls in head injury rates in the two largest helmet law provinces (accounting for 89% of cyclists studied) came either before introduction of the law or more than a year after and were most likely due to general trends unconnected with the legislation. Trends in head injury rates after the law were similar to those in non-legislation provinces.
Similarly the author makes no reference to published criticisms of cited studies in New Zealand (Scuffham, Alsop, Cryer and Langley, 2000) and England (Cook and Sheikh, 2000). In New Zealand, consideration of a less restricted time period has shown that cyclists fared no better in terms of reductions in head injury than non-cyclists and may even have fared worse (Robinson, 2005). No mention is made of the New Zealand author's subsequent paper (Taylor and Scuffham, 2002) that found that the helmet law had not been cost-effective. The English study was the author's own and has been criticised for incorrect data and interpretation. Indeed, Government data shows that head injury rates improved much more for children during a period of declining helmet use than for adults for whom helmet wearing increased by 50%. (BHRF, 1099)
No reference is made to any source that has shown no benefit from large increases in helmet use although these sources have often been based on much larger and more robust data sets.
It is mistakenly implied that even those who suggest that risk compensation takes place believe that helmets are protective in a crash. In fact, scepticism about helmet effectiveness is based on a wide range of considerations including the physical capabilities of cycle helmets and the mechanisms that lead to serious brain injury. Moreover, risk compensation is not just about people consciously taking more risks because they believe themselves to be better protected. Much risk compensation is sub-conscious and indirect.
Whilst acknowledging that risk compensation has been shown to occur in many scenarios, the author claims that there is little evidence of this with regard to cycle helmet wearing. This is not so. In one study by the Transport Research Laboratory (Halliday, White, Finch and Ward, 1996), children were found to be more likely to take risks if helmeted. Another study (Mok et al, 2004) - which included a prominent researcher previously dismissive of risk compensation - concluded that there was clear evidence of a risk compensation effect among helmeted cyclists, at least in those situations that were the subject of the study. As another example, UK Department for Transport data (Gregory, Inwood and Sexton, 2003) shows clearly that cyclists are much more likely to wear a helmet when riding on busy main roads than on quieter roads.
The author has misinterpreted evidence from Spaite et al, 1991 in suggesting higher speeds among riders without helmets. Spaite analysed collisions involving cyclists with motor vehicles (only). It is likely that when reporting that unhelmeted cyclists were more often in higher impact collisions than helmeted riders (a situation that applied even when head injury was not a factor), that this was due to motor vehicle speeds, not higher speeds of cyclists. Spaite concluded that helmet use might be a marker for other behavioural differences and that some of the benefit attributed to helmets might simply be a reflection of the severity of impact sustained by helmet users compared with that sustained by non-users. There is other evidence of behavioural differences between helmet users and non-users. For example, one study (Farris et al, 1997) showed a strong association between helmet use and safer riding practices, helmet wearers being 2.6 times more likely to obey stop signs and 7.1 times more likely to give proper hand signals.
Noting large reductions in cycle use in Australia when helmet laws were first introduced, it is suggested that the level of cycling recovered within a few years. A single reference (Finch, Heiman and Neiger, 1993), based on a single town (Melbourne), is cited to support this. However, this is not a complete reflection of that study which also reported that teenager cycling remained suppressed by 43% in the last year surveyed compared with pre-law. An apparent recovery in overall cycling numbers was distorted by a cycling rally passing through one of the sites surveyed. Elsewhere, there is evidence to suggest that it took at least ten years for cycling levels to recover in absolute numbers in some towns whilst in many others cycling is still not back to pre-law levels. Moreover, where cycle use has recovered, the profile of cycling has changed, with little recovery for utility journeys and amongst children.
The latter fact makes it all the more strange that the author should claim that "the deterrent effect may be less among children". In fact, it is widely acknowledged, even among people sympathetic to helmet laws, that children - and in particular teenagers - are the most likely to be deterred from cycling by helmet laws. In both Australia and Canada, the falls in cycling among teenagers were approximately double those for the population as a whole, and at one Sydney school 90% of teenage girls abandoned cycling (BHRF, 1096).
A single source (Macpherson et al, 2002) is cited in support of the author's prognosis, based on a study of children in one community in the City of Toronto, Ontario. It is not clear that the children in the community studied were typical of all cyclists in the Province of Ontario, where the helmet law was in any case never enforced. Whilst this law appears not to have led to a reduction in cycling in this particular community, nor did it lead to an increase in helmet wearing (Macpherson, Parkin and To, 2001).
It is suggested that there are clear precedents for overruling personal choice, such as with regard to car seatbelts. A difference, however, is that seatbelts had been shown to be beneficial at least to their users based on evidence from multiple sources with no contradictory evidence. (There remains controversy about the repercussions of the law for the safety of more vulnerable road users – Adams, 1982; Adams, 2004). In the case of cycle helmets, there is a sharp conflict of evidence between small-sample non-randomised case-control studies and time series and other data, often based on much larger populations. Furthermore, cycling has substantial health benefits that are eroded by discouraging the activity and that do not apply to motorised transport.
Although many people do believe that cycle helmets offer significant protection against head injuries, that does not mean that they are necessarily right. Many people also have an exagerrated view of the likelihood of head injury, in part due to the ways helmets are promoted. Similarly, although the cited meta-analysis (Towner et al, 2002) concluded that there is much evidence that helmets are effective, there is also much evidence to the contrary, none of which was considered by either the meta-analysis or the present author.
The data and debate on helmet effectiveness presented in this paper is shallow, poorly informed and selective. It is constrained by what appears to be the author's limited perception of the arguments against helmet efficacy and compulsion, and also by his belief that non-randomised case-control studies are the 'best data' to inform the debate, irrespective of the quality of those studies, the considerable contrast between their outcomes and real-world experience over the 30 years during which cycle helmets have been available, or the wider controversy that currently surrounds studies of that type.
Adams J, 1982. The efficacy of seat belt legislation: A comparative study of road accident fatality statistics from 18 countries. SAE Transactions 1982:2824-2838.
Adams J, 2004. Seat belt laws - a clumsy solution. Oxford conference on clumsy solutions .
Cook A, Sheikh A, 2000. Trends in serious head injuries among cyclists in England: analysis of routinely collected data. BMJ 2000;321:1055.
Farris C, Spaite DW, Criss EA, Valenzuela TD, Meislin HW, 1997. Observational evaluation of compliance with traffic regulations among helmeted and nonhelmeted bicyclists. Ann Emerg Med 1997 May;29(5):625-9.
Finch C, Heiman L, Neiger D, 1993. Bicycle Use and Helmet Wearing Rates in Melbourne, 1987 to 1992: the influence of the helmet wearing law. Monash University Accident Research Centre Report 45.
Gregory K, Inwood C, Sexton B, 2003. Cycle helmet wearing in 2002. Transport Research Laboratory Report 578.
Halliday ME, White C, Finch H, Ward K, 1996. Attitudes to cycle helmets - a qualitative study. Transport Research Laboratory Report 154.
Macpherson AK, To TM, Macarthur C, Chipman ML, Wright JG, Parkin PC, 2002. Impact of Mandatory Helmet Legislation on Bicycle-Related Head Injuries in Children: A Population-Based Study. Pediatrics 2002; 110(5):e60.
Macpherson AK, Parkin PC, To TM, 2001. Mandatory helmet legislation and children's exposure to cycling. Injury Prevention Inj Prev 2001;7:228-230.
Mok D, Gore G, Hagel B, Mok E, Magdalinos H, Pless IB, 2004. Risk compensation in children's activities: A pilot study. Paediatr Child Health 2004;9(5):327-330.
Robinson DL, 2003. Confusing trends with the effect of helmet laws. Pediatrics P3R 7 Jul 2003.
Robinson DL, 2005. Head injuries and helmet laws in Australia and New Zealand. .
Scuffham P, Alsop J, Cryer C, Langley JD, 2000. Head injuries to bicyclists and the New Zealand bicycle helmet law. Accident Analysis and Prevention 2000 Jul;32(4):565-73.
Spaite DW, Murphy M, Criss EA, Valenzuela TD, Meislin HW, 1991. A prospective analysis of injury severity among helmeted and non helmeted bicyclists involved in collisions with motor vehicles. Journal of Trauma 1991 Nov;31(11):1510-6..
Taylor M, Scuffham P, 2002. New Zealand bicycle helmet law - do the costs outweigh the benefits?. Injury Prevention 2002;8:317-320.
Thompson RS, Rivara FP, Thompson DC, 1989. A case control study of the effectiveness of bicycle safety helmets. New England Journal of Medicine 1989 v320 n21 p1361-7.
Thompson DC, Rivara FP, Thompson RS., 2002. Helmets for preventing head and facial injuries in bicyclists (Cochrane Review). Cochrane Database Syst Rev issue 4, 2002.
Towner E, Dowswell T, Burkes M, Dickinson H, Towner J, Hayes M, 2002. Bicycle helmets - a review of their effectiveness: a critical review of the literature. Department for Transport Road Safety Research Report 30.