Cycle helmet debate

This article is designed to be an overview of the cycle helmet debate in reasonable but hopefully not excessive detail, and including enough of the facts and sources to get you started.

Please see also Cycle helmet debate/Terms. You may also want to read the overview

My bias
Now I know some people will accuse me of being anti-helmet and thus unable to present a balanced view, but this is false: I am not anti-helmet, I am a former helmet advocate who has become sceptical in response to the evidence. Nonetheless I don't doubt that I will have missed out some factors which fail my personal cultural filters, so I encourage you do your own research and think for yourself. And don't forget the bit about thinking.

There are now a significant number of helmet laws around the world, so it is entirely fair to ask those who promote new laws to demonstrate that those already in force have delivered what was promised. As far as I know - and I have asked - they cannot do this. The purported success of helmet laws is always presented in terms of increases in helmet wearing (begging the question) or reductions in the number of injured cyclists (which turns out to be due to a reduction in the number of cyclists). All of which just serves to make me more sceptical, I'm afraid.

The latest response is to point to the fact that a high proportion of helmets are apparently incorrectly fitted and worn. I recall a Dilbert cartoon:


 * Salesman: This is our most reliable computer, unless you try to use software. It'll freeze several times a day. But you can restart it by poking a spoon into a hole in the back.
 * Dilmom: Has that ever worked?
 * Salesman: We think people are doing it wrong.

So: we say a helmet law will save enormous numbers of deaths and serious injuries. Has that ever worked? We think people are wearing them wrong.

Convinced? No? Shame on you :-)

A balanced view
The thing is, a balanced view means looking at both sides, then making up your mind. I have done that, and have come out on balance broadly sceptical, and strongly against compulsion and aggressive promotion. I know others who are accepting of promotion, but not in favour of compulsion.

Remember, too, that the mean point in this debate is not between scepticism and zealotry, it's not even between those who advocate laws banning helmet use (if any such exist) and those who advocate compulsion - beware the fallacy of the false middle. The default in the scientific method is for those who advocate an intervention, to bear the burden of evidence, and to explain away any conflicting evidence, but helmet advocates seem to consistently fail to recognise the null hypothesis with the result that they fall into several of the traps inherent in observational studies.

I cannot readily improve on this thoughtful statement from the National Cycling Strategy Board:

Much of what you will read around here is also addressed on http://www.cyclehelmets.org and http://www.cycle-helmets.com. When you've read those, go to pro-helmet sites like BeHIT and BHSI and see if you can spot any allusion to the conflicting nature of the evidence (to be fair, BHSI alludes to it, but simply dismisses it). A good litmus test of any helmet site is: if they uncritically quote 85% effectiveness, engage suspicion mode immediately (see Thompson, Rivara and Thompson (1989)). By comparison with BeHIT and BHSI, of course BHRF look anti-helmet. Ask yourself this, though: who would you rather believe, someone who was promoting helmets so set out to prove they are a good thing, or a world-renowned authority on cycling safety who was called as an expert witness and found to his great surprise that the evidence was vastly weaker than had been made out?

Helmets and Safety: The Meat of the Argument
The first, and probably most counter-intuitive thing to learn about bicycle helmets is that they have very little to do with bicycle safety. You see, already I'm telling you things which can't possibly be true. But bear with me. Road safety minister David Jamieson was said to be pro-helmet, and yet he acknowledged in a letter that the Government knows of no case where cyclist safety has improved with increasing helmet use. And that's not for want of looking, even to the extent of commissioning some leading helmet advocates to write an "independent review" of the evidence (OK, now I know you're going to accuse me of bias, but it is true: several of the authors of DfT research report 30, 2002, have a previous publication history which includes pro-helmet work, and none have ever published against compulsion or promotion; indeed, almost all of them are linked through a course sponsored by a group which is pro-helmet - that kind of independent report).

Where was I? Oh yes, so why don't helmets work? Well, I'm not actually saying that, and neither I think was Mr Jamieson. The thing is, the pro-helmet and anti-compulsion camps are a bit like the two old women shouting between top-floor windows on opposite sides of a mediaeval street: they are arguing from different premises.

Helmet promotion is often pushed by medics working in accident & emergency. They see "lots" of cycling head injuries. Of course, most of them are trivial. Scalps bleed profusely, and people are far more likely to attend A&E for a cut head than a cut leg, if only because it is harder to dress a wound in the scalp using domestic first aid kits. There are loads of reasons why head injuries are more likely to end up in hospital than injuries elsewhere. And all these are counted, and the number looks big. Anything up to 100,000, if you believe BHIT. Or it might be 28,000 this week, that's a bit of a moving target. It is these kinds of figures on which prospective studies are based, and these are the studies used to justify promotion and even compulsion of helmets. They show figures for efficacy up into the high 80s percent (but more on that in a minute).

Population-level figures generally include only serious and fatal injuries, because these figures are much easier to collect in a robust manner (and because who needs to know about non-serious injuries?). At that level, helmets apparently have no effect. Even where compulsion has been introduced and helmet use has doubled or more over a period of three years or less it is still not possible to see the effect of helmets on cyclist head injuries (provided you control out the massive reductions in the numbers of cyclists caused by the laws, of course - something the helmet promoters sometimes forget to do, or in one case even count as a benefit, having removed people from "danger"; this is extreme, most medics will acknowledge without hesitation that cycling is good).

This is not at all inconsistent: if helmets prevent a high proportion of cut heads but no serious injuries or deaths in crashes involving cars, for example, that would explain the disparity between the population-level figures and the prospective studies right away. It would also be consistent with the claims of the helmet manufacturers, which are much more modest than those of the campaigners, presumably for legal reasons, and with the published standards for helmets, which involve impacts equivalent to a very low speed crash or a fall from a stationary riding position. All this makes perfect sense. The logical disconnect comes when you say "helmets prevent 85% of injuries, therefore they would prevent 85% of deaths". I don't think it requires particularly acute critical faculties to see the flaw in that chain of logic. You can call cuts and bruises lacerations and contusions, to make them sound worse, and you can lump everything from a cut ear to acute neurological trauma into one basket of "scary head injuries", but trivial injuries remain trivial; a device which protects against such injuries cannot without substantial additional evidence be assumed to prevent more serious consequences.

So there you have the bones of the argument. Cycle safety campaigners see no justification for the monomaniacal focus on helmets which has characterised public debate on cycle safety for the last couple of decades. We would much rather see the effort and cash (over £100,000 of public money has been given to a helmet promotion charity in recent years) used on things which are known to produce definite improvements in overall cyclist safety, of which the first and best is getting more people on their bikes.

History

 * See also: Cycle helmet timeline

Bicycle helmets had a slow evolution. Early racers sometimes wore an arrangement of padded leather strips to prevent the scalp and ears from being abraded in crashes, but it wasn't until the late 1960s and early 70s that people started experimenting with sports helmets for cycling. This was a fairly extreme reaction to the perceived risk of cycling, since these devices were barely vented, if at all, and the first Snell standard issued in 1970 indicated a device not unlike a light motorcycle helmet (imagine wearing a 1kg unvented helmet while riding up a long hill on a hot day!). In the mid 1970s Bell responded to the perceived market by producing the Bell Biker, a hard-shell helmet which would probably exceed every standard currently used to test helmets.

By the mid to late 1980s there was a thriving helmet industry, and you started seeing cyclists wearing helmets around and about. These early helmets were mostly certified to Snell standards, and many of them had hard shells. They were rounded, had minimal ventilation, were fairly heavy - in short, selling them to a society which had never viewed bicycling as particularly dangerous was something of an uphill struggle. Now, if I were in charge of Bell Helmets back then I would have done a couple of things: I'd have funded some soccer-mom handwringers to campaign for helmet use with good old-fashioned "think of the children" emotional nonsense, and I'd have put some money into some studies which would "prove" that helmets were some kind of magic talisman. Strangely, that's pretty much how it panned out, only apparently without Bell having to put its hand in its pocket. What luck!

The granddaddy of all helmet studies, referenced above, was produced in 1989 by Thompson, Rivara and Thompson. Their report, with its puzzling conclusion that helmets are more effective against brain injury than cuts and bruises, launched the laws in Australia and New Zealand which started the whole compulsion bandwagon rolling. Later re-analysis of the raw data showed some anomalies: although they claimed that they had controlled for differences between their "case" group of mainly solo urban poor road cyclists and their "control" group of predominantly white middle-class families riding on off-road trails, Dorothy Robinson, a statistician from Australia, showed from their data that you could equally show that helmets had prevented 75% of broken legs. The authors have since published revised, lower estimates, but the 85% and 88% figures you see quoted everywhere are from the original 1989 study. Anyone using those figures should know better!

Modern helmets are rarely certified to Snell standards. There are several competing standards, EN1078 in Europe, ANSI Z.90 (presumably designed to sound like Snell B90) and CPSC standards in the USA. Snell are generally the most stringent, requiring drops onto kerbs of various sorts. Few helmets certified to EN1078 would meet the Snell B90 standard (this was current around the time TR&T did their work). So modern helmets are less good than the ones on which the most prominent study is based.

Australia and New Zealand These two countries became the poster children for helmet compulsionists the world over. They pointed at the reduction in head injuries in Australia in particular, ignored the fact that the reduction in cyclist numbers had been substantially greater, and declared the laws a success. Ten years later, with a rapidly growing population and with an Olympic Games to promote all kinds of fitness activity, Australian cycling is still struggling to match pre-law levels. Utility cycling has recovered slowest, teenagers are least likely to have taken up cycling - it's been relegated to a leisure activity, marginalised by a law which saw cyclists head injuries stay the same, proportionally, as those of pedestrians.

The table below shows adult cyclist hospital admissions not involving motor vehicles, the figures which might be expected to show exactly the change predicted by helmet advocates. No motor vehicle is involved, so these crashes are theoretically at least within the envelope of the design performance of helmets. These figures are for New Zealand.

Looking at these figures, and remembering that helmets are supposed to prevent 90% of all head injuries, can you spot the year when helmet usage doubled? Here's Dr. Nigel Perry's graph of cycling head injury percentage against helmet usage and head injury percentage in the general population (shrunk, please see here for the original):



Helmet use (dark green) vs. head injury rate for cyclists (pink) and all population (blue)

Can you see anything here to suggest that helmets are preventing a meaningful number of injuries? It looks to me as if the trend for cyclist injuries is not noticeably different from the trend for the whole population - and surely we can't credit helmets for saving those who aren't wearing them? (Actually we can try, but we won't get very far).

How Helmets Work
For the last decade and more the standard helmet has been made of polystyrene with some kind of abrasion-resistant cover. The helmet works by crushing in a controlled way, rather like crumple zones in cars. Collision energy is absorbed by the destruction of the helmet. This leads to one of the absolute fundamentals of modern cycle helmets: if your helmet has been in an accident throw it away immediately. The chances are it will have been fatally damaged, even if you can't see the damage.

My thanks to "Mike", who sent me feedback to say this is wrong. Actually I believe it is right: I have looked into it, but he is made a valid point in that old-fashioned hard-shell helmets were sold as spreading an impact over a larger area. More modern thinking has it that serious and fatal brain injuries are predominantly not caused by focal injuries anyway but by diffuse effects resulting from extreme accelerations. The tests demand that the helmet reduce acceleration from a given impact to within certain limits, which supports the view that they are designed to decelerate the head rather than spread the load. As far as I can tell there is no measurement in the standards tests of point loadings or other parameters related to a helmet's ability to spread load. If the helmet were genuinely designed to work primarily by spreading load, I don't think we would see the current designs with their deep, narrow ribs, and hardshell helmets would still dominate since they do this much better. As it is I believe the human skull is already quite well designed for low-speed impacts.

Many people hold up broken helmets as evidence that the helmet has "saved their life." This is almost certainly false. Helmets are designed to absorb impact by plastic deformation (squashing). In brittle failure mode, polystyrene foam absorbs very little energy. A broken helmet is actually evidence that the helmet failed to work as designed. Think of it like dropping a stone on a polystyrene ceiling tile: if the tile stays in one piece the stone lands with a soft thud, but if the polystyrene breaks (for example, if it's suspended even a little way off the ground) the tile will snap and the stone will hit the ground with a loud bang. That's one reason you have to be sure a helmet fits properly, if you wear one.

Helmets are also supposed to spread point loads over the surface of the skull. I'm not so sure that modern highly ventilated helmets do this too well. There is also no doubt that in order to build in vents without failing the drop tests, the shell of the helmet has become thicker; there is no more foam there, but it's in radial bars instead of a hemispherical shell. That means your head is bigger than it would otherwise be -especially for children's helmets. That might be one factor accounting for the fact that helmeted cyclists are more likely to hit their heads than unhelmeted cyclists. Another could simply be the kinds of cyclists who are likely to wear helmets, of course...

Choosing and Fitting
I've already talked a bit about standards; Brian Walker of Head Protection Evaluations makes a living testing helmets. He recommends Specialized, as being the only ones widely available in the UK which would meet the Snell standard. A CE mark is merely a license to sell in the EU, and means nothing, and CPSC is "self-certified" with no independent testing, so Snell is your best bet. Go here for a list of certified lids - and don't bother looking for Bell or its brands (e.g. Giro), they are not in there.

Your helmet must fit properly without slipping about (as this report shows). Better models have an adjustable cradle at the back of the head which allows very precise fitting. Make sure that the helmet doesn't press against your skull, as this rapidly becomes uncomfortable. Adjust the chinstrap so it's as close to your neck as possible - you need to be able to open your mouth to yell at motorists. Remember: if your helmet slips on impact, it won't protect you.

If you are genuinely concerned about serious head injuries you should buy a helmet which is smooth, has a minimum of vents and a hard shell. Avoid the duck-beak back ends which are very popular at present, they have been said to cause injury in off-road crashes.

The helmet will probably come with various foam pads which can be stuck or velcroed in place to improve fit. If the helmet does not fit reasonably well without these, it doesn't fit. The pads are just to fine-tune for comfort, not to rely on for fit. Distrust immediately any helmet which only comes in 'Large' and 'Small' - or, worse still, one size with a plethora of various thickness pads.

What Helmets Are Good For
Helmets are primarily designed to protect against injury in simple falls at speeds up to around 12mph. The limitation is primarily due to weight and ventilation considerations: motorcycle helmets are better, but far too bulky, heavy and restrictive for use on bicycles.

Helmets are not designed to protect in crashes involving motor vehicles and are not associated with reductions in serious head injury; only 8% of reported injury accidents involving bicycles are single-vehicle accidents and 50% of head injuries sustained are on areas which are not, or would not have been, covered by a helmet. Crashes involving motor vehicles typically involve forces well above the rated capacity of Formula 1 helmets. If you ride recklessly in traffic confident that your helmet will spare you harm, then you are asking for trouble. Much better to read Cyclecraft, and ride as if your helmet is made of meringue covered in eggshell. Which, in the context of a road traffic crash, it is. The worst cyclist crashes are those where the rider is crushed by a (usually) left-turning lorry. No helmet in the world will prevent a 38T truck from killing you, much better to learn not to ride up the inside of them in the first place. In fact, much better to learn not to crash in the first place. So, a copy of Cyclecraft is a much better investment than a helmet if you've only got ten quid to spend.

Helmet advocates make claims for up to 90% head injury reductions where helmets are worn. These claims are problematic. They tend to be accompanied by gratuitous scaremongering; cycling is not a dangerous activity and Mayer Hillman, author of the British Medical Association's report on this issue, concludes that the benefits outweigh the risks at least twenty-fold. The claims tend to be accompanied by puzzling anomalies, like conclusions which don't match the data, greater effect against serious injuries than against trivial ones, or sometimes glaring statistical errors, as in Cook & Sheikh's 2003 paper. And they are not matched even slightly by real-world whole population data - that's the kind which proved the link between smoking and cancer, set against the kind of small-scale often industry-funded trials which set out to prove the opposite. Again, I'm showing my prejudices, but there is no doubt that whole population evidence is more robust than small-scale studies. I think it is no surprise that one of the leading opponents of compulsory helmets in the UK is an epidemiologist by profession.

Problems with Promotion
Helmet advocates love to claim that helmet use can be promoted without deterring cycle use. Angela Lee of BHIT has stated that even the idea of deterrent effect is ridiculous (hear that, all you Aussies? You're ridiculous, Angie says so). The figures do not support that assertion. In every country where compulsion has been introduced, there has been a substantial and sustained reduction in cycling. Without claiming cause and effect, you can see from the chart below that helmet use is highest where cycling is least common. You can also see that helmet use is highest where deaths are most common. The received wisdom is that helmets are actually irrelevant in this case: that cyclist safety increases with cycle use - that is certainly what Occam's Razor would suggest here. Source: Cycle helmet wearing Netherlands, Denmark, Finland, Sweden EC 1999; Germany Pucher 2000; UK Bryan-Brown 1999; USA & France (Paris) Osberg 1998. Cycle deaths EU CfIT 2001; USA Pucher 2000. Cycle Percentage of trips: ECF 1997 for estimates of cycling, Flanders cities; Pucher & Dijkstra 2000 for USA, Canada, France, Italy, Austria; EC 1999 for UK, Norway, Switzerland, Sweden, Denmark; Bracher 2003 for Germany; EC 2002 Wallon region; Welleman 2000 Netherlands.

Children
Inexperience, and the use of bicycles for play, increases the risk of the kinds of accidents for which helmets are designed. I said in the original version of this page that "Between 80% and 90% of these accidents take place off the road, so would not be covered by any helmet legislation." How wrong I was! The Martlew Bill sought to make it a crime to allow any child to ride any pedal cycle in any public place without a helmet. Yes, that means riding a tricycle in the park without a plastic hat would be a no-no. But don't worry, those rather natty electric trikes and jeeps would be fine, so you can save the little darlings the effort of pedalling. Compulsory helmet wearing by children when riding on the road would result in an estimated maximum of two lives saved per annum compared with a situation where no children used helmets.

We need to be careful that we don't give out the message that the helmet is some kind of magic talisman which will protect you from the consequences of your own actions. Good roadcraft, maintenance, correct use of lights and visible clothing all have a part to play - a much bigger part than injury mitigation devices, as most cyclist injuries are not even theoretically preventable by helmets. One of the more unsavoury aspects of the Martlew campaign was the use of grieving mother Carlie Annetts, whose son Troy died after being hit by a car. Helmet promoters had apparently persuaded her that her son would be alive today if he had been wearing a helmet. The Coroner's report puts the blame for his death not so much on his lack of a helmet as his lack of working brakes, and the fact that this led him to ride out off the pavement into the path of the car. Is this an argument for helmet compulsion? Or better enforcement of bike maintenance and pavement riding laws?

Adults
Despite the fact that they do not protect against the most common types of accident, helmet wearing is becoming more common among adults. This is partly due to the feeling of safety which helmets provide (although this is to an extent illusory), partly because modern lightweight helmets are not too intrusive and may have a sun visor which is handy in bright sunlight (ie, twice per year in the UK) and partly because if you go over the handlebars of a safety bicycle, the first thing to hit the ground is your head. However unlikely such an accident might be, the consequences are so painful as to make wearing a helmet seem attractive. It's also a measure of the success of the scare campaigns. The introduction to Haynes' Bicycle Book says something along the lines of "few people would dare to ride without one" - this is flat wrong, wearing rates are nowhere near 50% anywhere in the UK.

Sports Events
Many sporting events require the use of helmets. This may include races, organised time trials and mountain bike events. Racing drivers are required to wear helmets and fireproof suits. We don't consider these necessary when popping down to Sainsbury's for a loaf of bread. Nor has the compulsory use of cycle helmets stopped a steady trickle of pro cyclists from suffering serious and even fatal head injury. This will not be a surprise to you by now, as you will have realised that the benefits of helmets are limited. The mandation of helmets by racing bodies is driven by insurance companies, and not even based on actuarial judgements, as there is insufficient data to form valid comparisons. Unless you regularly ride down mountains at 50mph hanging on the back wheel of another cyclist, the Tour de France's helmet requirement has nothing to say to you!

Insurance
You may be aware that insurers have also tried on several occasions to reduce the scale of compensation to cyclists injured by negligent motorists, on the grounds that the cyclist was not wearing a helmet. This particularly unsavoury bit of victim-blaming prompted CTC to set up the Cyclists' Defence Fund, which has successfully fought a number of such cases. If you go cycling on holiday you may need to check the small print, some travel insurance has reportedly written in a helmet clause (but this is uncommon).

The Case For Compulsion
Cyclists are vulnerable, but giving them separate space on the roads is expensive. Compulsory helmet wearing is a 'safety measure' whose costs fall entirely on the cyclist; no government spending is required so it is a superficially attractive "quick fix". And. Shroud-waving is prominent in the world of helmet advocacy - the deaths of two helmeted cyclists in races in the USA last year are not mentioned.

Comparison is often made with motorcycle helmets. Compulsory wearing of helmets has, according to some reports, reduced deaths of motorcyclists due to head injury by a significant amount (the fact that the fatality rate in US states which repealed helmet laws declined by comparison with those which did not, is as irrelevant as the fact that the fatality rate in the UK dropped by only 2% when the overall rate dropped by 9% in the year compulsion was introduced here - irrelevant, that is, if you are a helmet promoter). Even given the easily provable fact that compulsory motorcycle helmets have not reduced motorcyclist deaths, the comparison with cycle helmets still doesn't stand up. Cyclists are killed in Britain at a rate of 53 per hundred million passenger hours, but the rate for motorcyclists is 382 fatalities per hundred million passenger hours - seven times higher. Motorcyclists wear much larger and stronger helmets, and there is some evidence that these actually contribute to injuries in certain circumstances - and motorcyclists are typically travelling many times faster than cyclists. Cycle helmets are not designed to withstand impacts involving motor traffic.

One might just as well suggest that helmet wearing be made compulsory around the home, where most head injury accidents happen.

The Case Against Compulsion
The case against compulsion is often represented as primarily libertarian, but as you will no doubt appreciate by now, there is a strong evidence base to suggest that compulsion simply does not work. The kinds of accidents which worry the politicians and are used to scare up support - cyclists being knocked off by cars - are not the kinds of accidents in which helmets provide any protection. In addition to the lack of benefit, there is also a positive disbenefit to compulsion: compulsory helmet wearing leads to a reduction in cycling, which is profoundly undesirable in a society where obesity is rapidly approaching epidemic proportions;

In any case the case for compulsion misses the point. It's not that cyclists are uniquely vulnerable, or that they require separate provision, but that motor traffic is uniquely dangerous and threatening to users of benign modes. Pedestrians and horse riders are also vulnerable to motor danger, but these three groups cause little or no danger to each other because they are moving at human-scale speeds and are not isolated from their surroundings.

Ethics of compulsion
The ethical tests for compulsion have been posited as: Is the goal of the programme good? Does the programme achieve the goal effectively? Does it do so efficiently? Does it do so in a manner consistent with the values and liberties of the target population?

Is the goal good? Well, it depends how you understand the goal. The goal of helmet promotion seems to me to be to make people wear helmets and who cares if it makes any difference but that may be just a tad cynical. Given the evidence that compulsion deters cycling and thus increases overall danger, it would be difficult to say that the goal is to make cycling safer, and there are far more head injuries admitted to hospital as a result of simple falls, and of course pedestrian accidents. But to be charitable (ish) I guess the actual goal is to try to reduce the number of cycling head injuries, a small proportion of the injuries suffered by a small proportion of road users (off-road cycling is generally an order of magnitude safer, except for extreme technical riding). As such, I have to say the goal lacks ambition.

Does the programme achieve the goal effectively? Hell, no! Look at Australia: cycling numbers reduced by a third, cycling head injuries proportionally higher than before the law? Terrible. And think of all the head injuries sustained by pedestrians on the roads - none of them addressed by the measure. The measure is poorly targeted and it has failed to deliver wherever it's been tried to date. Nobody has even thought about testing other methods to improve cyclist safety, although Ken Livingstone has quite by accident achieved the best recorded year-on-year improvements in safety in London through the introduction of the Congestion Charge. There is no evidence that this has affected helmet wearing rates (anecdotal evidence suggests they've gone down but that can only be happenstance). So on the one hand you have a measure which has always failed, and on the other one which quite by chance succeeded spectacularly. I know which I'd be seeking to repeat.

Does the programme achieve the goal efficiently? Hell, no! Dorothy Robinson showed a cost-benefit deficit of the New Zealand law of some millions of dollars, and that was not including fines or costs of enforcement. You can't achieve cost savings without saving lives and serious injuries, and helmet laws simply don't do that. The result is lots of enforcement cost, lots of helmets hanging on hooks unused most of the time, lots of bikes left in the shed because the owner can't legally ride them without going out and spending money on a helmet, and lots of people not riding their bikes either because they've been put off by the lids or because the message that cycling is dangerous has got to them. So they revert to motor transport, generating extra journeys, and storing up more health problems due to physical inactivity.

Does the programme achieve its goals in a manner consistent with the values and liberties of the target population? Hell, no! The values and liberties of the target population are essentially ignored. The proposed law in the UK made adults liable if their children took their helmets off without permission, the Australian law led to one 15-year-old girl being strip-searched and detained overnight. Enforcement has fallen more heavily on ethnic minorities, who in those countries are also least able to afford the helmets. Cycling is supposed to be carefree and convenient. If helmets are such a great idea, why are none of the cycling bodies in favour of them? Nobody wants to die or end up a vegetable. None of us are in it for the glory of being martyred for the cause - if we didn't think cycling was a safe and rational transport choice we would not be doing it in the first place! And if the helmet promoters can't see the difference between popping down the road to the shops and descending the Col du Tourmalet that is surely their problem not ours.

Show Me The Money
If helmets prevent such a large proportion of head injuries, why is this not reflected in the real world? Cyclist and pedestrian casualty rates historically track each other closely. An investigation of these rates over time shows that this parallel trend is not changing despite the increasing prevalence of helmets. When Australia introduced mandatory helmet legislation the cyclist injury and head injury levels dropped - but so did the number of people cycling. So the rates stayed the same. Australia is now the second fattest nation on Earth. In New Zealand, as we have seen, the effect of the helmet law is not visible in the data for cyclist casualties. Where is the evidence to support the grandiose claims of the helmet advocates?

Subliminal Messages
Helmet promotion (without compulsion) would seem on the face of it to be sound, but even here there is a risk: the promotional materials create an unwarranted perception of cycling as a risky activity. It isn't. Yes, cyclists do occasionally get knocked off and killed, but on average they live around ten years longer because they get regular cardiovascular exercise. You will live longer - provided you live. And once again the danger comes mostly from drivers, not from the activity of cycling itself. The logical solution is to promote safe driving rather than helping cyclists to mitigate the risk caused to them by unsafe driving.

There is also a question as to whether it is ethical to suggest that wearing a helmet will protect a cyclist in an accident. The range of accidents in which a helmet is of use is narrow and specific, so unless there is a large amount of small print there is a danger of false reassurance, leading to potential risk-taking. The best protection, after all, is to avoid the crash in the first place.

Risk Factors
Lorries in particular are dangerous to cyclists, there being a 25 times greater likelihood of death in a cyclist versus-lorry-crash than in cyclist-versus-car. In accidents involving motor vehicles and cyclists, the motor vehicle is to blame in two thirds of cases - but the cyclist is, unsurprisingly, 37 times more likely to be killed or seriously injured than the motorist. Even then, cyclists are statistically less at risk than motorcyclists. This suggests that driver education would be a good place to focus attention.

The best protection against cyclist vs. lorry crashes is the provision of advanced stop lines (ASLs) and cycle lanes on the approaches to junctions. This simple and relatively inexpensive measure, combined with driver and cyclist education, may well be effective in preventing the worst lorry-versus-cyclist accidents. Cyclists are out of the danger zone to the left of the lorry (where they may be out of the field of vision of the driver's mirrors, and where the road can suddenly vanish as the lorry turns left). Education is also an important factor, both for cyclists and lorry drivers. Such deaths are in any case often caused by crushing injuries to the torso, with helmets therefore playing no significant part.

Factors Increasing Safety
Cyclists' death and injury rates decrease with increasing prevalence of cycling. The UK death rate of 53 per hundred million passenger hours is on a base of an average 62km cycled per person per year. In Holland the death rate is 23 per hundred million on a base of 850km per person per year, and in Denmark the rate is even lower - 18 deaths per hundred million hours on a base of 1,050km per person per year.

These ratios have been reasonably consistent over time, and indicate that the very best way to improve road safety for cyclists is to encourage more people to cycle. The reasons are obvious: more drivers will also be cyclists, and therefore be aware of the existence and needs of bikes on the roads; and the more bikes there are, the more drivers will expect to see them.

There is also good evidence that cycle lanes at least 2m wide, advanced stop lines, cycle-only phases at traffic lights, and good junction and roundabout design, are effective in increasing safety. After all, the most survivable crash is the one that doesn't happen in the first place. Helmets, by contrast, are a "sticking plaster" solution which fails to address in any way the underlying causes of accidents.

Helmets Reducing Road Safety
The most quoted example of the effect of compulsory helmet wearing is that of Australia, which passed legislation in 1991. The effect was immediate and dramatic: a 36% decline in cycling in the year after enactment. Research shows that recovery is slow. In Perth, in 2000 the reduction was still 15% against 1991 levels, although population had increased by 140,000 over the same period.

And, as the figures above suggest, fewer bikes on the road meant that cyclist injuries rose to their highest recorded levels. As a road safety measure it was - and remains - an own goal greater even then compulsory seat-belt wearing for car drivers.

The Case for Wearing a Helmet, Even Though You Don't Have to
All the above makes it look as if cycle helmets have a negative impact on road safety. This is not necessarily so, it is just that the accidents against which they protect are not the kinds of accidents which are increasing.

They can offer protection in one important class of cycle crash: those due to poor road surfaces. Any cyclist who has bounced along the road on his or her head following a pothole-related wheel circularity modification incident will probably become a dedicated helmet evangelist. If you don't like road rash on your head, wear a helmet (or a stout canvas cap). Just don't expect it to save your life.