|Evidence that helmets are effective||Evidence that casts doubt on helmet efficacy|
Case control studies
Most of the evidence in favour of helmet effectiveness has come from 'case control studies' where a group of cyclists with head injuries is compared with one or more groups without.
Case-control studies are less reliable than randomized controlled studies or cohort studies; there are several examples where case-control studies produced incorrect and totally misleading results. Some studies attempt to match ‘cases’ (subjects with head injury) to non-head injured 'controls' with the same age and background. However, as was the case in the first case-control study of helmet wearing (Thompson, Rivara and Thompson, 1989), it is not always possible to find suitable controls. In other studies, researchers attempt to statistically ‘adjust’ for differences such as age or motor vehicle involvement (known as confounders) between cases and controls. Again this can be very difficult to do. For example, helmet wearers tend to be more safety conscious and there is evidence that they have lower impact collisions (Spaite et al, 1991). Case-control studies have not tried to measure safety consciousness or impact speed, so it would be very difficult to fully adjust for them
Randomized controlled trials have not been used in helmet research for practical reasons. Since injuries to cyclists are rare, the researchers would have to find a very large number of participants and persuade some, chosen at random, to wear helmets, others not. The researchers would then have to follow up these people and find out what injuries, if any, they might have had. This would be extremely time-consuming and difficult, and almost all of the participants would have no relevant injury at all.
The Medical Research Library of Brooklyn provides these explanations of medical research methods:
The likelihood of a crash when cycling is not the same for all cyclists or in all situations, and not all crashes are equally likely to result in a head injury. For example, the risk of a crash is very much influenced by social factors (Grayling et al, 2002), cycling skill, where the person is riding and even personal temperament. Serious head injuries are much less likely in crashes with no motor vehicle involved and when a person rides more cautiously. When people choose whether or not to wear a cycle helmet, that too may be influenced by their attitude to and perception of risk and other personal factors. It is probably not possible to compensate for all these variables when comparing groups of cyclists.
Just as gloves prevent minor wounds to the hands, cycle helmets are likely to prevent similar wounds to the head, but these are not serious injuries or in any way life-threatening. In the largest case-control study, 73% of head injuries did not involve concussion or other brain injury (Thompson, Rivara and Thompson, 1996). Based on this wide definition, including all wounds to the head, helmet wearers will have lower rates of head injury. However, the case-control study evidence for other types of head injury is not as consistent as is often suggested. Although one frequently cited but much-criticised study claims that helmets reduce head injuries by 85%, all other studies predict more modest benefits, such as 39% (McDermott, Lane, Brazenore and Debney, 1993) or none at all (for some types of cyclists, e.g. cyclists over 15 riding on-road) (Jacobson, Blizzard and Dwyer, 1998).
Some studies have found that only hard shell helmets protect from head injury (Hansen, Engesaeter and Viste, 2003), or that non-approved helmets offer no benefit (McDermott, Lane, Brazenore and Debney, 1993). It is difficult to see why non-approved and soft shell helmets should offer no benefit at all if protection from hard shell helmets is significant. In today's fashion-conscious market, most helmets on sale have only a thin microshell if any at all, and hard shell helmets are difficult to buy. Moreover, the estimates of protection given by hard shell helmets are very variable. In a single study, odds ratios of 1.2, 0.9 and 0.42 were calculated (##219'#). Other case-control studies also have wide confidence intervals, e.g. 1.16 to 9.06 (Maimaris, Summers, Browning and Palmer, 1994). This makes it very difficult to determine if they show that helmets have any real benefit at all.
Sometimes helmets have been found to protect from injuries to the whole face (Wasserman and Buccini, 1990), sometimes to offer no protection against facial injuries (Hansen, Engesaeter and Viste, 2003) and, in one study, children under 9 who wore foam helmets had four times as many face injuries as non-wearers (Hansen, Engesaeter and Viste, 2003). Claims have been made that helmets can protect from death (Dorsch, Woodward and Somers, 1987), that they are less effective in greater impact crashes (McDermott, Lane, Brazenore and Debney, 1993) and that they only protect against serious, but not more minor, injury (Shafi et al, 1998) which seems improbable. Most of the more serious injuries to cyclists are as a result of a collision with a motor vehicle, but case-control studies have sometimes found helmets to be ineffective in these circumstances (Larsen, 2002).
Some studies have noted how bareheaded riders not only suffer more head injuries, but also more serious non-head injuries in crashes with motor vehicles, suggesting that riding style and other factors may be more important determinants of head injury than whether or not a helmet is worn (Spaite et al, 1991). On the other hand, helmeted cyclists have sometimes been found to have more serious non-head injuries, such as injuries to the neck, the trunk, extremities and pelvic girdle (McDermott, Lane, Brazenore and Debney, 1993; Wasserman and Buccini, 1990; Maimaris, Summers, Browning and Palmer, 1994). It has been suggested that helmeted cyclists hit their heads more frequently than those without helmets (Wasserman et al, 1988).
Significantly, only studies by one team of researchers have concluded that all types of cycle helmet offer protection to all cyclists under virtually all circumstances, with and without motor vehicle involvement (Thompson, Rivara and Thompson, 1989; Thompson, Rivara and Thompson, 1996). It is this research that is most frequently cited in favour of cycle helmet effectiveness and helmet laws.
Scientific research usually takes pains to investigate discrepancies between its own conclusions, those of other research and data from other sources. As a result there is normally much closer agreement about the interpretation of studies than is to be found in cycle helmet research. The positive results from case-control studies may simply result from inability to compensate sufficiently for the many auxiliary differences between helmet wearers and non-wearers.
Most case-control studies on cycle helmets have been the subject of peer criticism and critical meta-analysis (e.g. Towner et al, 2002) for shortcomings that make their conclusions unreliable. For example, there has sometimes been no distinction between head and facial injuries or between life-threatening and minor injuries; data has been self-reported with no checks for reliability; and target groups have sometimes been atypical (e.g. club cyclists). The wider circumstances of crashes have never been taken into account.
Cycle helmet research is not alone in being controversial. In recent years an increasing number of papers in epidemiological journals have drawn attention to the unreliability of findings in case-control studies (BHRF, 1134).
Three literature reviews of the medical evidence about cycle helmets have been published (Towner et al, 2002; Thompson, Rivara and Thompson, 2002-9; Attewell, Glase and McFadden, 2001). Based almost exclusively on case-control studies, two focus on the same set of studies and the third on a subset of this set. One review (Thompson, Rivara and Thompson, 2002-9) has been criticised for a lack of independence, being dominated by the authors' own studies. Another (Towner et al, 2002) because its conclusion (in favour of cycle helmet effectiveness) is at odds with the extensive criticisms of the studies published in a technical annex.
Powerful evidence for helmet effectiveness comes from the experiences of helmeted cyclists who have hit their heads in a crash and believe that their helmet has saved them from injury. However, there is no evidence that helmets save lives or prevent serious injury at all across cyclists as a whole.
For a full discussion of this paradox, see "A helmet saved my life!".
If cycle helmets are effective in reducing head injuries, and in particular if the more optimistic predictions for their effectiveness are true, then it is reasonable to expect to see a reduction in head injuries (relative to cycle use) across the whole population of cyclists where helmet use has become common. However, there is no whole population data from anywhere in the world to confirm these predictions.
Helmet laws in Australia, New Zealand and parts of Canada (BHRF, 1096) have resulted in the great majority of cyclists wearing helmets, but there has been no reduction in rates of head injury relative to cycle use. An analysis of enforced laws in these countries found no clear evidence of benefit (Robinson, 2006).
Casualty trends from other countries where helmet use has become significant also show no reductions in serious or fatal injuries attributable to helmets. In the USA, an increase in helmet use from 18% to 50% of cyclists over a decade was accompanied by a 10% increase in head injuries. There was no clear evidence of any increase in cycle use, which may have declined (BHRF, 1028). In England, an analysis of road traffic injuries found no association between differing patterns of helmet wearing rates and casualty rates for adults or children (Hewson, 2005; Hewson, 2005b).
More localised studies have also failed to find population-level evidence of a significant benefit from helmet use. For example, a study based in the Lothians in Scotland found that, although helmet wearers had fewer wounds or other minor injuries to the head and neck, there was no significant difference in outcome between helmet wearers and non-wearers for injuries serious enough to warrant hospital admission or follow-up (Scottish Exec, 2005). There were other differences between wearers and non-wearers. For example, adults were significantly more likely than children to be wearing a helmet (57% vs 24%). Also, although 39% of injured cyclists wore helmets, a much lower proportion (18% in 2001) of Scottish cyclists said they always wore a helmet, with a further 5% saying they wore a helmet on most trips. This suggests that helmet wearing is associated with a higher risk of injury.
Whole population data on cyclist fatalities is particularly robust as, unlike for lesser injuries, it is not subject to under-reporting or inaccurate classification of injury severity. Long-term analyses of fatalities in Canada (Burdett, Can), New Zealand (Burdett, NZ) and USA (Kunich, 2002; Rodgers, 1988) show no helmet benefit; indeed, one study (Rodgers, 1988) suggests helmeted cyclists to be more likely to be killed. Although fatality rates have generally declined, cyclists have fared no better than pedestrians.
Research suggests that rotational force causes most of the fatal and disabling brain injuries in road crashes, but there is no assurance that cycle helmets can mitigate it. Indeed, there is some evidence that cycle helmets may increase it.
For further discussion of this topic, see: Cycle helmets and rotational injuries.
In all countries where helmet laws have been introduced and enforced, there has been a substantial reduction in cycle use (BHRF, 1096). Helmet promotion has also been shown to reduce cycle use (BHRF, 1020). Cycling has substantial health benefits (BHRF, 1015) and people who cycle regularly live, on average, longer than non-cyclists with less illness and poor health (Andersen, Schnohr, Schroll and Hein, 2000). Concern has therefore been expressed that any benefit from helmet wearing is outweighed by the loss of health benefits to those deterred from cycling. This is particularly an issue at a time when reductions in cycle use and other physical activity have given rise to increased levels of obesity, especially among children.
The greatest influence on cycling safety is the number of people who cycle; cycling gets safer the more people who do it (Jacobsen, 2003; Robinson, 2005b; Turner, Roozenburg and Francis, 2006). Conversely, any reduction in cycle use, due to helmets or any other factor, results in reduced safety for cyclists as a whole, including those who decide to wear helmets.
Just as most pro-helmet analysis is characterised by a restricted evidence base, so it is the case that analyses based on a wider range of research are invariably more sceptical about the benefits of helmets.
After considering the evidence for and against helmets, the Royal College of General Practitioners concluded in 2005 that the evidence was too ambiguous for it to take a stand one way or another. In 1999 the British Medical Association, though supporting helmet use, decided firmly against compulsion after undertaking a comprehensive review of the evidence (BMA, 1999). (In 2005, the BMA reversed this decision, in favour of compulsion, on the basis of a 4-minute debate and a single paper, produced in response to a letter-writing campaign, that did not consider any evidence unsupportive of helmet use and which has been much criticised for bias and inaccuracies (BMA, 2004)).
A review for the National Childrens Bureau concluded that the strong claims of injury reduction made by helmet proponents have not been borne out for fatalities in real-life settings. It warned that those who cycle should be under no illusion that helmets offer reliable protection in crash situations where lives may be in danger (Gill, 2005).
The detailed analyses carried out by the Bicycle Helmet Research Foundation have also not found the balance of the evidence to be supportive of helmet efficacy.
Solicitors who specialise in cyclist injuries have written that, in their experience, the use of cycle helmets has not reduced the likelihood of serious injury (BHRF, 1054; BHRF, 1173; Fulbrook, 2004). In the UK, the courts have so far not supported claims that wearing a helmet would have made any difference to injuries suffered by cyclists in the cases they have considered. Senior neuro-surgeons have given evidence that cycle helmets afford very limited head protection that would make no significant difference in cases involving serious injury (BHRF, 1054). Experience suggests that doctors are much more cautious in their assessment of cycle helmets when they give evidence on oath and are subject to cross-examination and the high standards of evidence required by the courts.
One of the world's most prominent helmet test experts has stated that most helmets are physically incapable of sustaining impacts of the type associated with serious crashes; helmets provide protection only in low impact crashes under favourable circumstances (Walker, 2005). Consumer tests of cycle helmets have shown that many helmets do not meet the standards to which they are accredited and only a very few helmets meet the higher standards most relevant to real-life crashes (Which, 1998).
Helmeted cyclists have been shown to be more likely to hit their heads if they crash and may be more likely to crash in the first place (Wasserman et al, 1988). The disproportionate number of helmet wearers who believe that a helmet has saved their life (see above) is further evidence that helmet use might adversely affect crash involvement or outcome.
Risk compensation by cyclists who wear helmets has been confirmed in research (Mok et al, 2004; Halliday, White, Finch and Ward, 1996). It is also reflected, where helmet use is voluntary, in the much higher levels of helmet wearing by cyclists riding on busy roads (Inwood, Whitley and Sexton, 2005).
At least 14 young children have died world-wide through strangulation by their cycle helmets when playing off their bikes. Others have been saved only by timely intervention (BHRF, 1227).
Andersen LB, Schnohr P, Schroll M, Hein HO, 2000. All-cause mortality associated with physical activity during leisure time, work, sports, and cycling to work. Arch Intern Med 2000 Jun 12;160(11):1621-8.
Attewell RG, Glase K, McFadden M, 2001. Bicycle helmet efficacy: a meta-analysis. Accident Analysis & Prevention 2001-05 v33 n3 p345-52.
Barnes JE, . A bicycling mystery: head injuries piling up. New York Times, 29 July 2001.
Cycle Helmets. BMA Board of Science and Education ISBN 0-7279-1430-8.1999.
BMA, 2004. Legislation for the compulsory wearing of cycle helmets. BMA .
Burdett A, . Cyclist fatalities in Canada. OCBC .
Burdett AJ, . Cyclist fatalities in New Zealand. .
Dorsch MM, Woodward AJ, Somers RL, 1987. Do bicycle safety helmets reduce severity of head injury in real crashes?. Accident Analysis and Prevention 1987 Jun;19(3):183-90.
Finvers KA, Strother RT, Mohtadi N, 1996. The effects of bicycling helmets in preventing significant bicycle-related injuries in children. Clinical Journal of Sports Medicine 1996 Apr;6(2):102-7.
Fulbrook J, 2004. Cycle helmets and contributory negligence. Journal of Personal Injury Law 2004:171-191.
Gill T, 2005. Cycling and children and young people: a review. National Children's Bureau .
Grayling T, Hallam K, Graham D, Anderson R, Glaister S, 2002. Streets ahead - Safe and liveable streets for children. Institute for Public Policy Research .
Halliday ME, White C, Finch H, Ward K, 1996. Attitudes to cycle helmets - a qualitative study. Transport Research Laboratory Report 154.
Hansen KS, Engesaeter LB, Viste A, 2003. Protective effect of different types of bicycle helmets. Traffic Inj Prev 2003;4(4):285-290.
Hewson PJ, 2005. Cycle helmets and road casualties in the UK. Traffic Injury Prevention 2005;6(2):127-134.
Hewson PJ, 2005. Investigating population level trends in head injuries amongst child cyclists in the UK. Accident Analysis & Prevention 2005;37(5):807-815.
Inwood C, Whitley C, Sexton B, 2005. Cycle helmet wearing in 2004. Transport Research Laboratory Report 644.
Jacobsen PL, 2003. Safety in numbers: more walkers and bicyclists, safer walking and bicycling. Injury Prevention 2003;9:205-209.
Jacobson GA, Blizzard L, Dwyer T, 1998. Bicycle injuries: road trauma is not the only concern. Austr NZ J Public Health 1998 Jun;22(4):451-5..
Kunich TH, 2002. Latest CPSC helmet standard and US fatality trends. OCBC .
Larsen LB, 2002. Betydningen af brug af cykelhjelm for hovedskader hos tilskadekomne 0-15-årige cyklister. Ugeskr Læger 2002;164(44):5115.
Maimaris C, Summers CL, Browning C, Palmer CR, 1994. Injury patterns in cyclists attending an accident and emergency department: a comparison of helmet wearers and non-wearers. BMJ 1994 Jun 11;308(6943):1537-40.
McDermott FT, Lane JC, Brazenore GA, Debney EA, 1993. The effectiveness of bicycle helmets: a study of 1710 casualties. Journal of Trauma 1993; 34(6):834-845.
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, 2005. Safety in numbers in Australia: more walkers and bicyclists, safer walking and bicycling. Health Promotion Journal of Australia 2005;16:47-51.
Robinson DL, 2006. Do enforced bicycle helmet laws improve public health?. BMJ 2006;332:722-725.
Rodgers GB, 1988. Reducing bicycle accidents: a re-evaluation of the impacts of the CPSC bicycle standard and helmet use. Journal of Products Liability 1988 ,11:307-317.
Extent and severity of cycle accident casualties. Scottish Executive Social Research .2005.
Shafi S, Gilbert JC, Loghmanee F, Allen JE, Caty MG, Glick PL, Carden S, Azizhan RG, 1998. Impact of bicycle helmet safety legislation on children admitted to a regional trauma center. J Paediatric Surgery 1998 Feb;33(2):317-21.
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..
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., 1996. Effectiveness of bicycle safety helmets in preventing head injuries: a case-control study. JAMA 1996 Dec 25;276(24):1968-73.
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.
Turner SA, Roozenburg AP, Francis T, 2006. Predicting accident rates for cyclists and pedestrians. Land Transport New Zealand Research Report 289.
Walker B, 2005. Heads up - the science of helmets. Cycle Jun/Jul 2005.
Wasserman RC, Buccini RV, 1990. Helmet protection from head injuries among recreational bicyclists. American Journal of Sports Medicine 1990 Jan-Feb;18(1):96-7.
Wasserman RC, Waller JA, Monty MJ, Emery AB, Robinson DR, 1988. Bicyclists, helmets and head injuries: a rider-based study of helmet use and effectiveness. American Journal of Public Health 1988 Sep;78(9):1220-1.
Get a head start. Which? October 1998 pp 28-31.