Effectiveness of cycle helmets and the ethics of legislation

Abstract

Professor Sheikh and his colleagues (June 2004 JRSM1) argue for compulsion in the use of helmets by cyclists, referring to a previous paper of theirs2 claiming that the rate of head injury amongst serious casualties fell 3.4 percentage points (PPs) more for cyclists than pedestrians during a period in which measured on-road helmet use increased by 5.8 PPs, a rate of increase of less than 1 PP per annum. It was assumed that all of the advantage for cyclists was due to increasing helmet use. They concluded that cycle helmets prevent 60% of serious head injuries. Clearly they miscalculated. With a prevailing rate of head injury amongst serious casualties of about 30%, as in this case, and helmets 60% effective, a 6 PP increase in helmet wearing would reduce the head injury rate by only 6×0.18 PPs=1.09 PPs, not the 3.4 PPs assumed to be a 'helmet effect'. If all the observed improvement were due to helmet use, then the effectiveness would be around 190%. The authors' assumption, in their JRSM paper, of a linear relationship between fractionally rising helmet use and population level injuries is speculative and is contradicted by experience in countries where helmet use increased at more than ten times the rate in Britain. In both Western Australia and New Zealand, helmet use increased from negligible levels to more than 80% in around eight years, yet follow-up studies3,4 did not show long-term benefits for the cyclist populations relative to control groups. Study of injury trends in each state of Australia for the period when helmet laws were passed shows stable characteristics, revealing no evidence of extra prevention due to legislation coming into force.5 Thus international evidence indicates that the authors' interpretation of British data is in error. It must be stressed that hospital-based injury data include both off-road and on-road injuries. Road casualty data specifically show that rising helmet use is associated with cyclists' injuries getting more severe relative to other road users. An alarming association with increased risk of death has twice been reported.6,7 Risk compensation by helmeted cyclists is the most plausible explanation. The case for only cyclists to wear helmets is weakly founded. Estimates of risk8 may be calculated from routinely collected casualty and use data. In pedestrians, risk per mile travelled is about 60% higher than in cyclists. Pedestrians are far more vulnerable than cyclists, facing a 2% risk of death in a reported road accident, as against 0.7% for cyclists.9 A scenario of the disastrous consequences of promoting walking helmets has been proposed.7 Comparison with risk in driving requires certain adjustments to enable a semblance of like for like comparison. Despite the marginalized condition of cycling in Britain, the risk per hour travelled may be as low as the EU average for drivers. There is no case for distinguishing cyclists as a high-risk group. The one clear population-level effect of helmet laws that has been widely reported is the deterrence of cycling. In every case where data are available, cycle use has fallen by 25–50% when a helmet law was enforced.10,11 This has a direct consequence on the risk of death in cycling. Study of international evidence points to a reliable relationship between the amount of cycling and the risk in cycling12—a power–law relationship with an index value of around 0.4. A fall in cycle use of 50% would increase risk per cyclist by more than 50%, whereas an increase in cycling of 100% would reduce the risk by almost 40%. Public health would benefit substantially.13 A report by the Commons Select Committee on Health specifically cited a resurgence in cycling as 'probably the most effective response' that could be made to address the obesity 'time bomb'. It is most likely that road deaths would fall overall; even in Britain one hour of cycle use is not more likely to result in a road death than one hour of driving, because the third-party risk from cycling is so low.8 With an increase in cycling, the advantage would swing to the bicycle. Tripling the level of cycle use by 2012 in line with Government policy would dramatically reduce the risk in cycling, improve public health and most likely reduce road casualties overall. Helmet laws never achieved anything positive elsewhere; why should the British experience be any different?