Cycles of all sorts

Research evidence

Evidence for the efficacy of helmets in preventing serious injury is contradictory and inconclusive. In general, analyses of the relative merits of different bike safety interventions put helmets low down, because no helmet will reduce the probability of crashing (and there is some evidence that helmets may increase this likelihood). Proactive measures including bike maintenance and riding skills are far more important. Although the link is not causal, it is observed that the countries with the best cycle safety records (Denmark and the Netherlands) have among the lowest levels of helmet use. Their bicycle safety record is generally attributed to public awareness and understanding of cyclists, education, and to some extent separation from motor traffic.

The evidence comes from two main types of observational study:

• case-control studies, in which cyclists who have injured their heads ("cases"), and cyclists who have not ("controls"), are compared. Such studies consistently find that cases report a lower rate of helmet-wearing than controls. This has been taken as strong evidence that cycle helmets are beneficial in a crash and that all cyclists should be compelled to wear them. Known problems with this study design include confounding (attributing benefits from differences in behaviour to differences in helmet choice), and reporting errors (people falsely reporting helmet use).
•  The most widely-quoted case-control study, by Thompson, Rivara, and Thompson, reported an 85% reduction in the risk of head injury by using a helmet. There are many criticisms of this study, including use of an unrepresentative control group. This study includes clear evidence that the injury profile of helmeted and unhelmeted cyclists in a population with voluntary wearing is significantly different, an effect first documented by Spaite et al.:
• A striking finding was noted when the group of patients without major head injuries (246) was analyzed separately. Helmet users in this group still had a much lower mean ISS (3.6 vs. 12.9, p less than 0.001) and were much less likely to have an ISS greater than 15 (4.4% vs. 32.1%, p less than 0.0001) than were nonusers. In this group, 42 of 47 patients with an ISS greater than 15 (89.4%) were not wearing helmets. We conclude that helmet non-use is strongly associated with severe injuries in this study population. This is true even when the patients without major head injuries are analyzed as a group
• Population studies compare changes in helmet use and injury rates in a single population over time, most notably where helmet laws have resulted in large changes in a short time. A review of jurisdictions where helmet use increased by 40% or more following compulsion showed no measurable change to head injury rates. The largest study, covering eight million cyclist injuries over 15 years, showed no effect on serious injuries and a small but significant increase in risk of fatality. Weaknesses of this type of study include: simultaneous changes in the road environment (e.g. drink-drive campaigns); inaccuracy of exposure estimates (numbers cycling, distance cycled etc.).

Different analyses of the same data can produce different results. For example, Scuffham analysed data on the New Zealand helmet law in 1995 and concluded that, after taking into account long-term trends, the laws had no measurable effect. His subsequent re-analysis without accounting for the long-term trends showed a small benefit. Re-analysis of the Thompson, Rivara and Thompson data substituting helmet wearing rates from co-author Rivara's contemporaneous street counts, reduces the calculated benefit to below the level of statistical significance. Another analysis of the source data from this study showed a 70% reduction in lower limb injuries from helmet use. One problem with all analyses is that the population of injured cyclists is generally very small, and it is difficult to collect sufficient incidents to form a statistically significant sample.

The definition of injury is also open to debate, and injury figures are acknowledged to be inacurate. Research by TRL and others shows that reporting of injuries is related to severity: fatal injuries are almost always reported, in the developed world, but 90% or more of lesser injuries go unreported. Helmets are most likely to be effective against lesser injuries. Pro-helmet studies routinely refer to prevention of traumatic brain injury, which has connotations of permanent intellectual disablement, but where sufficient data is provided it is found that the majority of the brain injuries in these studies are concussion. A study of fatally injured cyclists found injuries of fatal severity to multiple organ systems were in sixteen of twenty riders, including six with no significant head injury. Four riders died of fatal injury to head alone and one of these was the only rider known to be wearing a safety helmet. His death resulted from a fall from a bicycle at moderate speed rather than collision with a motor vehicle.

Recent research on traumatic brain injury adds further confusion, suggesting that the major causes of permanent intellectual disablement and death may well be torsional forces leading to diffuse axonal injury (DAI), a form of injury which helmets cannot mitigate. Helmets may increase the torsional forces by increasing the distance from the extremities of the helmet to the centre of the spine, compared to the distance without a helmet.

Much of the research is partisan in one way or another. Thompson, Rivara and Thompson were already committed advocates of helmet legislation before publishing their first study; their report for the Cochrane review has also been criticised for being dominated by their own work. Rodgers, who showed helmets to be associated with increased risk of fatality, was replying to criticism of CPSC for focusing on bicycle design and manufacture standards. One report concluding a 60% reduction in injuries was found to be in error due to a simple statistical error; correcting the error results in a claimed efficacy of 186%; despite this the authors continue to assert that the results stand. A report commissioned by the UK Government was supportive of cycle helmet promotion but dismissed out of hand much of the contradictory evidence, and the principal authors were associated with a programme of the Child Accident Prevention Trust (CAPT), which is strongly pro-helmet. Curnow, author of papers on helmets and traumatic brain injury, has also published criticism of pro-helmet research.