Helmets maxi-FAQ

Cycle Helmets
Cycle helmet
Main list

Are you sure you want to ask this question?

OK, I assume you have come here for a reason. You have read the Mini-FAQ and you want more. More what? If you're here for the five pound argument then please look here first.

This page is not intended to be polemical, but nor is it guaranteed to be wholly free of bias since I find that my reading of most helmet reports is coloured somewhat by my understanding of the previous ones. If you don't trust me (and you'd be wise not to, after all I'm just some random bloke ni the internets) and you don't trust BHIT (wiser still!) then you have only one option: to find and read as much of the source data as you can. I give you fair warning: this will take a while. My helmet resource is far from complete and it contains over 700 documents of one sort and another. I am writing this in response to comments from uk.rec.cycling but this is me speaking, and if you don't like it - well, then you'll have to write your own.

And I have to be honest here, this is not so much a helmet FAQ as a helmet scepticism FAQ. It is an explanation of why scepticism is a valid view, as much as a look at the evidence. But then, I make no pretence to being anything other than a sceptic.

OK, hit me with it

Right, so the health warning has been ignored and you're still here. I'm going to divide the document up into chunks, in line with standard elephant-eating procedure. In many cases the situation for children is somewhat different to that for adults, and many helmet laws (and most helmet campaigns) cover only children, so the case for children is considered separately in some places.

You will see a lot of statistics. "Aha!" you will think, "Statistics are like bikinis: What they reveal is interesting, but what they hide may be vital." And you will be right. So I've tried to give some of the background. So this page violates another mildly sexist axiom, that web pages, like skirts, should be long enough to cover everything important, but short enough to remain interesting. This page is long. It's long because there are no short answers. Well, I suppose there is a short answer to the question "are helmets a good thing?" and that is "yes and no".

Risks of cycling

All discussion of the benefits or otherwise of helmets hinges on the perceived risk inherent in cycling. So, with shamelessly Anglocentric emphasis, here are some of the key figures.

You may note that the real figures do not always match up to the claims made by helmet promoters. I do not know why they exaggerate the risks - I have my suspicions, but Hanlon's Razor states that one should never attribute to malice that which can be adequately explained by incompetence.

Q: A:
How many people die cycling every year? The key source here is DfT transport statistics data - this shows that in 2003, 95 adult and 18 child cyclists died on Britain's roads. That is approximately 3% of all road fatalities.
How many children die cycling every year? Here are the figures from the Department of Health
Hospital admissions
Year Admissions
2001 - 2002 3,372
2000 - 2001 3,489
1999 - 2000 3,115
1998 - 1999 2,711
1997 - 1998 3,349
1996 - 1997 2,781
1995 - 1996 2,333
Year HI Fatalities As % of child cyclist deaths
2002 10 53%
2001 12 63%
2000 14 58%
1999 20 63%
1998 18 62%
1997 25 71%
1996 34 72%
1995 31 72%
1994 27 69%
1993 19 63%

Reply on 10th November 2003 by Dr Ladyman, Dept of Health Minister, to Parliamentary Question by Brian Jenkins MP. Hansard Written Answers, 10th Nov 2003 page 17-8W

You may have seen other claims. BHIT claim fifty children die of cycling-related head injuries annually, based on an anonymous citation of a Hansard report; this is doubly false since the year in question was 1996, a peak year for deaths in an otherwise reducing trend (see above), and the figure quoted was for all deaths. They have subsequently said that this is "an estimate based on under-reporting" but there is no known evidence to support the idea of substantial - or indeed any - under-reporting of child cyclist deaths. The number of off-road deaths is tiny - one for 2003, I understand - and under-reporting of fatal cyclist road accidents is 0%, according to the TRL.

Eric Martlew says: "It is estimated that about 30 children a year get killed on cycles and the lives of 12 of them could be saved if we had compulsory helmets." (BBC). His estimate for lives saved exceeds the total number dying of head injuries. There is no figure for how many of those had other mortal injuries as well, as would be common in cyclist v. motor vehicle crashes.

How many people are injured cycling every year? There are two common measures of numbers injured, that for serious injury and that for all injury. It is recognised that reporting rates are lower for less serious injuries, with non-reporting rare for fatalities but relatively common for trivial injuries, especially off road.

HASS/LASS gives a figure of around 132,000 injuries of all severities from leisure cycling treated in hospitals, GP clinics and minor injury units, and notes that up to four injuries may be recorded per treatment event.

DfT figures give 1,776 adult and 595 child cyclists killed or seriously injured on the roads, and 11,643 and 4,769 respectively for all severities. Some people assume a 1:20 rule for fatal to serious to slight; assuming this rule to be true, and assuming that the balance is accounted for by under-reporting, the majority of adult injuries recorded in HASS/LASS would seem to be due to road cycling. Urban roads are the most common place for injuries to happen outside the home.

TRL 220, discussed below, puts the number of injuries (all severities) at around 100,000, as do RoSPA in a 1996 booklet (this is the source of the figure used by Lee & Mann in their papers). Although these figures are speculative they are plausible - but remember that the vast majority of these injuries will be trivial.

The likely figure, then, would be: 113 dead 3000 serious traffic injuries, using Mills's 36% under-reporting for car v bike; 93% of recorded injuries on public roads & footways 100,000 trivial injuries, maybe.

Since over 3,000 children are admitted to hospital for cycling injuries of various sorts, and this exceeds the number of serious injuries recorded above, we might speculate that many of the admissions are not recorded as serious by the Police. For a more detailed picture, TRL220 is an excellent source.

How many children are injured cycling every year? Figures given to CTC by the Department of Health show just over 2,183 child cycling head injury admissions to hospital in 2002, 37.6% of all child cyclist admissions, of which under 500 are likely to have been genuinely serious injuries to parts of the head covered by a helmet.

Child cycling head injury admissions, then, account for 2.4% of child admissions, and 7.1% of child head injury admissions.

Injury risk is strongly age- and gender-dependent, with teenaged boys being most at risk. Walking alongside a road seems to be a major source of risk, and overall child pedestrian accidents are more likely to result in head injury than child cycling accidents, although when these are separated out it turns out that motor vehicle involvement is the big predictor of both occurrence and seriousness of head injury. In this respect cycling and walking are similar. There are around six times as many child pedestrian head injuries as child cyclist.

What is the relative risk? INRET (the French equivalent of TRL) studied 5459 casualties that were hospitalised after serious road crashes. The study noted the proportion of road users that suffered more than 24 hours' unconsciousness following the crash. The results were:
  • pedestrians 4.2%
  • motorcyclists 3.9%
  • cyclists 2.3%
  • car occupants 1.3%.

The study was in the late 1980s, before people were wearing helmets in France. (Ramet M, Vallet G. Typologies des accidentes du trafic routier a partier de 5459 dossiers. Rapports INRET/LCB Aug 1987 p106)

So what is the risk per unit exposure? CTC has calculated that the risk of serious injury for its members equates to around one injury per three thousand years of regular cycling.

Another risk estimate from someone involved in transport planning in Oxfordshire: "The number of accidents versus the number of opportunities is a big ratio. In Oxfordshire over a 5 year period there were around 1500 cycle accidents. In that same period one in ten of Oxfordshire's 600,000 population made roughly 2 cycle trips a day every day. Assuming a potential 1 reportable accident per trip, that's 219 million opportunities. So only 1 accident per 146 thousand chances of an accident. It is known that pedestrians from the poorest fifth of the population are 5 times more likely to be killed or injured than those from the wealthiest. Helmets may reduce the risk of injury in the event of a crash by 10% - 20%, but if those least likely to wear them are 5 times more likely to crash than those most likely, the reduction in total injuries is not going to be 10% - 20% but more like 1% - 2%. I don't think you'd notice the benefit in the noise."

The cyclist fatality rate in the UK per billion passenger km was 25.9, against an average for all modes of 12. Walking is rather more dangerous than cycling by this measure, but it is skewed by inaccuracies in the exposure measures.

Still confused! Is cycling dangerous? Not especially, and your chance of receiving a head injury should there be a crash is no different from any other mode - the thing that predicts head injury, especially serious head injury, is the involvement of motor traffic. The first, best thing you can do to reduce your chances of injury is to get trained. Adult cycle training is becoming much easier to find, and when the CTC staff and officers sampled the training recently they all came away having learned useful skills and tips.

If you can't find a training course - or if you can - you should also read Cyclecraft by John Franklin.

Helmets Work!

Yup, they definitely work. Up to a point. The $64 question is, of course, what point. Let us review some of the claims of efficacy:

Claim Source Comments
  • prevent 85% of head injuries
  • prevent 88% of brain injuries
A case control study of the effectiveness of bicycle safety helmets, Thompson RS, Rivara FP, Thompson DC. 1989. New England Journal of Medicine: 1989 v320 n21 p1361-7 This is probably the most widely-quoted figure for helmet efficacy, used by proponents of helmet laws in particular - scratch any helmet promotion campaign and this is the smell you'll get. The flaws in it are well documented and include:
  • an assumed helmet wearing rate in the "case" population taken from a homogeneous control group rather than co-author Rivara's street counts of the previous year
  • insufficient accounting for the different accident profiles of the case and control groups; the controls were less likely to suffer all kinds injuries, head or not
  • insufficient distinction between levels of severity of injury
  • acceptance of efficacy for some facial injuries without proposing a mechanism by which that might be achieved

The 85% figure is a litmus test: anybody quoting it is either uninformed or deliberately misleading you. It was one of the claims in the complaint to ASA about BHIT.

  • 69% efficacy in crashes involving motor vehicles
  • 68% efficacy in other crashes
  • 65% reduction in Injuries to the upper and mid facial areas
Helmets for preventing head and facial injuries in bicyclists, Thompson DC, Rivara FP, Thompson RS.. 2002. Cochrane Database Syst Rev: issue 4, 2002 The Cochrane Review is an influential and notionally impartial review of literature pertaining to various types of injury. This review was written by Diane Thompson, co-author of the Seattle study. Criticisms form part of the review in the Cochrane database, although they are sometimes edited and "spun". Criticisms of this particular review include:
  • Selection of studies is biased and the Review is not impartial. 4 of the 7 studies are by the reviewers themselves, and these dominate the overall findings.
  • Assessment of the selected studies is biased, with no mention of fundamental shortcomings; the 1989 Seattle study above is included as-is with no correction for widely-publicised confounding factors.
  • Only case-control studies are considered. No reference is made to population studies showing no change in head injury rates even with very large increases in helmet use. Time-series studies are less subject to this bias but these are not mentioned unless supportive of the authors' conclusions (most are not). All the selected studies are based on small samples (445 min, 3390 max) with no review or reference to whole-population effects (head injury rates increased by 40% in the USA as helmet use rose from 18% to 50% - source CPSC).
  • Case-control studies often show differences other than in head injury between helmet wearers and non-wearers (commented on in detail by Spaite) . Fundamental population differences make it difficult to separate these effects, but the Review makes no reference to such limitations - see also Two Old Women below.
  • The authors are dismissive of the possibility of risk compensation, and of counter-opinions generally. Risk compensation has been demonstrated in children by Pless and in the wider population by Wilde and Adams.
  • The authors are dismissive of discussion of the mechanisms of severe brain injury (Curnow).
  • Relatively small increases in helmet use in the selected studies cannot possibly account for the much larger reductions in head injuries cited even if helmets prevented all head injuries. (e.g. increase of 35% in cyclists wearing helmets led to 66% fewer head injuries– TR&T 1994)
  • There is misleading interpretation of 'odds ratio', using it interchangeably with 'percentage reduction in head injuries'. The paper showing the least benefit from helmets is omitted from computation of odds ratio. Both of these exaggerate benefit.
  • At least one study focusing on motor vehicle injuries was excluded because these are "not representative of the exposure experience of the population at risk" despite findings in other, included reports showing that motor vehicle involvement is strongly correlated with the most serious head injuries.
  • Conclusions are extremist and uncompromising ("helmets reduce bicycle-related head and facial injuries for bicyclists of all ages involved in all types of crashes including those involving motor vehicles", and "recommending or mandating helmet use is based on solid scientific evidence"). The authors urge the development of helmets with chin protection and face guards.

As a piece of op-ed summarising the most aggressive case possible for helmet use, this paper has some merit, but it is a long way from being either a complete or an impartial review of the available evidence.

65% of head injuries prevented Injury patterns in cyclists attending an accident and emergency department: a comparison of helmet wearers and non-wearers, Maimaris C, Summers CL, Browning C, Palmer CR. 1994. BMJ: 1994 Jun 11;308(6943):1537-40 This is an influential and very detailed study. It provides sufficient detail to draw some conclusions about the conclusions it draws, as it were.

The claims are essentially: 8.1% head injury among non wearing helmet owners; 9.2% among non-owners; 3.5% head injury among helmet users. Looking at the authors' Table III:

Leg || 31 (27) || 231 (25) || 0.90 (0.57 to 1.42)
Injury site Helmet worn Odds Ratio (95%CI)
Yes (%) No (%)
Head 4 (4) 100 (11) 3.25 (1.16 to 9.06)
Face or neck 22 (19) 212 (23) 1.32 (0.77 to 2.26)
Trunk 11 (10) 48 (5) 0.54 (0.26 to 1.11)
Arm 56 (49) 418 (45) 0.90 (0.61 to 1.33)

So we can see that (a) the confidence intervals are very wide, not surprisingly given the small sample sizes, and (b) there are some evident differences in injury profile between the populations (unhelmeted cyclists are much more likely to suffer trunk injury, but much less likely to suffer face and neck injuries, which helmets don't prevent). The differential effect of helmets owned but not worn over those not owned at all is also puzzling!

In the Conclusions the authors state: "However, the incidence of helmet wearing is low in Britain - 10.9% in our study. Educational programmes have been effective in increasing helmet use to over 50% in some American cities, and this has been accompanied by a 75% reduction in the number of bicycle related head injuries requiring hospital care (F Rivara, personal communication). The early results of our educational programmes in Cambridge are encouraging: there has been an overall increase in helmet use for the past two years, from 10% at the start of 1992 to over 16% in 1994"

And during that period (1992-1994) when helmet use increased by 2/3 in Cambridge, what happened to reported head injury rates?


So once again the claimed benefit is not observed in the whole population. But the authors dispute this, citing a letter from Rivara (remember that name?). Look again at that claim: an increase to 50% helmet wearing has resulted in a 75% reduction in hospitalisations for head injuries. The reduction in injuries is much greater than the increase in helmet use!

19% of serious head injuries prevented Head injuries to bicyclists and the New Zealand bicycle helmet law, Scuffham P, Alsop J, Cryer C, Langley JD. 2000. Accident Analysis and Prevention: 2000 Jul;32(4):565-73 This conclusion was reached by taking a subset of data the author had used in a previous paper, and contradicts his earlier conclusion that "increased helmet wearing percentages has had little association with serious head injuries to cyclists as a percentage of all serious injuries to cyclists [...] with no apparent difference between bicycle only and all cycle crashes" (Trends in cycle injury in New Zealand under voluntary helmet use, Scuffham PA, Langley JD. 1997. Accident Analysis and Prevention: 1997 Jan;29(1):1-9).

Problems with this paper include:

  • The HI rate in the first year considered was somewhat above the trend for the whole population, and the return to trend is the major source of the claimed benefit
  • Scalp lacerations are included as head injuries, although these are of little public concern and were not an area of focus in promoting the law

But these are minor points compared with what was found when Nigel Perry re-analysed the authors' own data, which they provided for him. The following chart was produced from that data, including data points not mentioned in the report, and shows the change in head injury rate for cyclists and for the whole population, and the change in helmet wearing rate over the same period:


From this it is obvious that:

  • The general fall in the likelihood of head injury for cyclists coincides to a high degree with a similar fall in head injury for non-cyclists (who make up the great majority of the whole population control group).
  • There is a fall for cyclists and not the control group when the law was introduced, but the cyclist head injury rate had previously risen above the overall one, and the fall is soon followed by an increase. The cyclist head injury rate does appear to fluctuate a little more than the overall rate (probably due to small numbers).
  • There has been no additional benefit for cyclists through the wearing of helmets that has not been enjoyed by the population as a whole without helmets. This is despite the fact that the survey on which the graph is based included head injuries sustained in simple falls (for which cycle helmets are designed to offer some degree of protection) as well as more serious instances of collisions with motor vehicles.

This is compelling evidence to support the idea that, whatever the efficacy of helmets themselves, helmet laws are not an effective way to reduce cyclist head injuries.

  • 39% of head injuries
  • 28% of face injuries
The effectiveness of bicycle helmets: a study of 1710 casualties, McDermott FT, Lane JC, Brazenore GA, Debney EA, Journal of Trauma: 1993; 34(6):834-845 From the Abstract: "The efficacy of helmet use was evaluated by comparison of crashes and injuries (AIS-1985) in 366 helmeted (261 Australian Standard approved and 105 non-approved) and 1344 unhelmeted casualties treated from 1987 through 1989 at Melbourne and Geelong hospitals or dying before hospitalization. Head injury (HI) occurred in 21.1% of wearers of approved helmets and in 34.8% of non-wearers (p < 0.001). The AIS scores were decreased for wearers of approved helmets (p < 0.001), face injuries were reduced (p < 0.01), and extremity/pelvic girdle injuries increased (p < 0.001) and the overall risk of HI was reduced by at least 39% and face injury by 28%. When casualties with dislodged helmets were excluded, HI was reduced 45% by approved helmets. Head injury reduction by helmets, although substantial, was less than that found in a similar study in Seattle, Washington."

Note that there is no comment on the relative severity, or whether or not a motor vehicle was involved. The claim regarding facial injuries may indicate confounding factors. Comments like: "Wearers of non-approved helmets sustained a blow to the head, face, or helmet more frequently (66.0%) than wearers of approved helmets (50.6%) or unhelmeted casualties (53.4%)" also suggest differences between the populations under consideration.

The claimed benefit is under half that of the Seattle study, and the sample group is much larger, in keeping with the common trend that larger studies produce lower figures.

  • 30% of slightly injured would be uninjured
  • 18% of the serious casualties would be slight
  • 11% of serious casualties would be uninjured
Pedal cycle accidents – a hospital based study, Paula Mills, TRL research report 220, 1989 These claims are noticeably lower than those usually made by helmet proponents. The study provides a lot of useful data, including nailing the idea that there is any significant under-reporting of fatal cyclist injuries as claimed by BHIT. The calculated under-reporting rates for slight, serious and fatal injuries are 74%, 61% and 0% respectively, with under-reporting of serious road injuries probably lower (36% in the case of car v bike injuries). From this we can extrapolate the DfT road casualty data to show something like a 20:1 ratio between slight and serious, and between serious and fatal injuries. The numbers of fatal injuries are extremely reliable, so this may be a helpful rule of thumb for checking claims of serious and slight injuries.

The locations of injuries as recorded were:

Location %
Road 81
Footway 7
Cycle Track 3
Footpath 2
Park / play area / BMX track 2
Other 5

It also provides the following table for fault in bike v car crashes:

Fault %
Cyclist 25
Motorist 57
Both 8
Environment 2
Mechanical Fault 2
Other 6

Fault was more likely to be assigned to younger cyclists.

Less than 5 per cent of the cyclists who had their accidents on a road or cycle track had been wearing conspicuity aids, however forty-two per cent of the cyclists had at least one conspicuity aid, not including rear reflectors, fitted to their bicycles, and lack of conspicuity was considered a factor in around 20% of cases.

The incidents giving rise to injury in adults were:

Source %
Bike on main road, passing side road 14
Being overtaken 9
Turning right 7
Passing parked cars 7
Wet surface 6
Bike v Bike 6
Bike fault 5
Entering / leaving roundabout 5

Note that this is for all crashes, including off-road. Clearly the major source of danger is the classic SMIDSY!

Unfortunately despite the wealth of data the report provides, there is insufficient detail to assess the reliability of the claims regarding potential injury savings. This is not a surprise, since the main thrust of the document is not to investigate helmet efficacy but to quantify the number and sources of injury, and in particular to establish the levels of under-reporting.

Age profile

Both helmet efficacy and accident involvement appear to be strongly age-dependent. It appears that the greatest benefit is experienced by the youngest children, but the highest number of accidents is experienced by teenaged boys. Again, it would be a mistake to suggest that the differential efficacy between younger and older children is a function of the helmets themselves - older children have more fully-formed skulls, and of course they are bigger and stronger. Teenaged boys are an injury risk just sitting still - and you know what happens when they get into a car!

There appears to be some merit in encouraging young children when learning to ride to wear something on their head - as well as knee pads and elbow pads. Once confidence and competence are gained this benefit is much harder to establish.

Helmets don't work!

For sure, helmets definitely don't work. For certain given values of "work" of course...

Reducing Bicycle Accidents: A Reevaluation of the Impacts of the CPSC Bicycle Standard and Helmet Use, Rodgers GB. 1988. Journal of Products Liability: 1988,11:307-317 This paper was a response by the Consumer Products Safety Commission (CPSC) to a paper by Petty claiming that its cycle manufacturing standards had been associated with an increase in risk to cyclists, while use of hard-shell helmets had been effective in reducing injuries. Rodgers reviewed all eight million cyclist crashes in the USA over a fifteen year period and concluded:
"First. there is no statistical evidence that the bicycle standard is positively and significantly correlated with the injury rate. Second, there is no evidence that hard shell helmets have reduced the head injury and fatality rates. The most surprising finding is that the bicycle related fatality rate is positively and significantly correlated with increased helmet use.
Although the helmet finding may seem surprising. helmet wearers could be less careful in automobile traffic, believing that their helmets give them an increased measure of protection. This would be consistent with the "lulling effect" discussed by Viscusi. It is also possible.although we have no evidence, that some helmets may increase the chance of an accident. by blocking visibility or sound. If helmet fit is not good. accidents could also occur due to the distraction of having to adjust helmets in traffic."
The efficacy of bicycle helmets against brain injury, Curnow WJ, 2003. Accident Analysis and Prevention: 2003,35:287-292 This is in part a critique of a meta-analysis by Attewell, Glass and McFadden (a singularly poor piece of work which double-counts at least one study).

"This examination concentrates on injury to the brain and shows that the meta-analysis and its included studies take no account of scientific knowledge of its mechanisms. Consequently, the choice of studies for the meta-analysis and the collection, treatment and interpretation of their data lack the guidance needed to distinguish injuries caused through fracture of the skull and by angular acceleration. It is shown that the design of helmets reflects a discredited theory of brain injury. The conclusions are that the meta-analysis does not provide scientific evidence that such helmets reduce serious injury to the brain, and the Australian policy of compulsory wearing lacks a basis of verified efficacy against brain injury. "

Vicki McCreery killed on Blackfriars Bridge Angie Lee jumped on the bandwagon of a highly publicised cyclist death to say: "Roads like this are leaving cyclists on a very thin line - this accident illustrates just how much danger they have to deal with every day. We are trying to protect ourselves and our children from suffering

serious head injuries but it seems that no one is listening to us. We just do not know what to do."

Vicki was wearing a cycle helmet, as she always did. BHIT were at sea when it came to dealing with a genuine cycle safety issue. CTC, by contrast, made an informed response. the Guardian reported as follows:

Roger Geffen, campaigns manager at the national cycling body, the Cyclists' Touring Club, said a cultural shift was needed so that local authorities considered cycle lanes more carefully. They had 'been left to the most junior planning officers, and we need better guidance on dealing with major junctions.'
Tony Russell, who advises councils on safer cycle lanes for the club, said: 'There are situations where designs put the cyclist in a more dangerous position. Most accidents, though, are caused by motorists not being careful.'
Fatal injuries to bicycle riders in Auckland, Sage MD. 1985. NZ Med J: 25 Dec 1985 Vol 98 No 793 The Abstract says: This study indicates that compulsory wearing of suitable safety helmets by cyclists is unlikely to lead to a great reduction in fatal injuries, despite their enthusiastic advocacy. Injuries of fatal severity to multiple organ systems were seen in sixteen of twenty riders, including six with no significant head injury. Only four riders died of fatal injury to head alone and one of these was the only rider know 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. Nonetheless, this does not exclude the possibility that a significant reduction in severity of non-fatal injuries to the head might result if helmets are worn, or some fatalities have already been avoided by their use.
The Standards The standard test for cycle helmets under EN 1078 involves dropping the helmet, containing a dummy head weighing not more than 6Kg, onto a flat surface from a height of 1.5 metres. I am a shade over 1.8m tall and I weigh around 82kg. If I fall over from standing upright I already exceed the impact which cycle helmets sold in Europe are tested to protect against by a substantial margin. And that's before I've even turned the cranks. Mind you, in my case that places me lower down. But then, when I come off the 'bent I always land arse first, so maybe I need polystyrene shorts...

Two Old Women

Long and terrible arguments have raged between road and off-road cyclists, often both opposed to compulsion but with a completely different perspective on what constitutes "reasonable precautions". Like the two old women holding a slanging match between upper windows on opposite sides of a cobbled street, they are arguing from different premises. This is eloquently summed up here:

There is no evidence [...] that helmets have had a marked safety benefit at the population level for road-using pedal cyclists. Clearly this cannot be extrapolated to non road-using pedal cyclists, indeed it may well be that leisure cycle use is unusually well protected by helmet wearing. Health practitioners, who have additional concerns about leisure use of bicycles may have justified interest in cycle helmets across the whole spectrum of cycling activity. However, it would appear that helmet wearing has at best a modest role in terms of Transport policy. Transportation practitioners therefore need to concentrate on primary safety, to prevent collisions occurring. Whilst it may be desirable for individuals to wear helmets, the circumstances when helmets may be effective are not clear. In particular, the evidence that cycle helmets are effective in traffic collisions is much weaker.

Cycle Helmets and Simpson's Paradox: where are the casualty savings? Hewson PJ, 2004

This also goes to the heart of the disparity between observational studies and population and time trend statistics. It is not alone in identifying fundamental differences between the helmeted and unhelmeted populations:

Discounting direct head injuries, helmet use has been associated with a much lower incidence of severe injury (injury severity score >15) to the whole body.17 It has been suggested that non-use is a marker of a population of cyclists more likely to be involved in high impact crashes with motor vehicles, perhaps as a result of less safe cycling practices. This study supports the view that use of a cycling helmet is associated with higher use of pre-event phase equipment measures to avoid collision."

Cycling safety: injury prevention in Oxford cyclists, L McGuire, N Smith, Injury Prevention 2000;6:285–287

Helmet non-use was strongly associated with severe injuries in this study population. This was true even when the patients without major head injuries were analyzed as a group; a finding not previously described to our knowledge. This implies that helmet non-users tend to be in higher impact collisions than helmet users since the injuries suffered in body areas other than the head also tended to be much more severe. Helmet use may be a "marker" for a population of cyclists who tend to be in less severe crashes. It is possible that some of the "protection" against head injury ascribed to helmets in previous studies is actually a reflection of the severity of the impact sustained by helmet users compared with that sustained by non-users."

A prospective analysis of injury severity among helmeted and non helmeted bicyclists involved in collisions with motor vehicles, Spaite DW, Murphy M, Criss EA, Valenzuela TD, Meislin HW. 1991. Journal of Trauma: 1991 Nov;31(11):1510-6

Logical Fallacies

Logical fallacies abound in the helmet debate, and can be seen from those on both sides of the argument - and even those in the middle. Logical fallacies are actually a common feature of all debate, and the more heated it becomes, the more entrenched the positions adopted, the greater the likelihood of their coming up. There are some good sites on the web regarding logical fallacies and how to recognise them. Nizkor is good, and so is Stephen Downes' guide to the logical fallacies. To a large extent they rely on a common fundamental flaw: the confusing of rationalisation with reasoning.

Logical fallacies will also creep in when the writer is looking to inject certainty. The helmet issue is fraught with contradictions and any truly authoritative statement is likely to require precision of wording, or many caveats. Usenet is a forum well suited to lengthy discourse, but (to show my prejudices) web forums tend to attract readers with less appetite for detailed argument, and a greater tendency to try for soundbytes. Soundbytes, in helmet wars, are almost invariably fallacious - in my view, anyway.

I have a whole page on logical fallacies - see how many you can spot. Bonus points for catching me out.


For any complex problem there is a solution which is simple, neat and wrong.

HL Mencken.

Consider the following example from Target Risk:

All human beings are mortal.
Socrates is a human being.
Therefore, Socrates is mortal.
A river empties into the sea through a delta.
The delta has three channels, all of equal size.
Therefore, damming two of the channels will reduce the flow of water to the sea by two-thirds.

One of these is self-evidently true, the other patently absurd. So here is another statement:

Cyclists can suffer head injury in a crash
Helmets reduce head injury in the event of a crash
Therefore cyclists should wear helmets

This is clearly analogous to one of the two statements above. But which one? Wilde's comment on the second statement is this: "One cannot stem the flow as long as there remain alternative routes to the destination. One cannot reduce mortality due to accidents and lifestyle-dependent disease unless all opportunity for premature death were eliminated by law or made impossible through technological intervention. And that, of course, can never be fully achieved. In the case above, the river would simply develop a fourth channel."

This is, of course, not a precise analogy. But it is a fair point. And especially when you consider that risk is not a measurable absolute, but a cultural construct, and one which varies from person to person. One person will be happy to go bungee jumping and skydiving, another will ferry Tarquin and Jocasta to school in the Range-Rover to avoid the danger posed by - well, by people ferrying Tarquin and Jocasta to school in the Range-Rover. Risk aversion varies with social class, gender and age. Middle-aged, middle-class women are very risk averse, teenaged boys from poorer social backgrounds are the least risk-averse. This is hardly a big secret. And the reason that helmet laws fail to live up to expectations may well be that they are trying to coerce people into someone else's pattern of risk-taking.

And there is no possible doubt that laws have failed to deliver anything close to the claimed levels of injury reduction. Far from preventing 85% / 75% / 69% / 65% / insert random percentage here of injuries, actually they don't even prevent enough injuries to pay for all the helmets in reduced healthcare costs. And even that is not actually a surprise - cycle injuries are rare, and you have to buy an enormous number of helmets to prevent any significant number of injuries.

Have a look at percentage head injury versus helmet wearing rates for Western Australia and New Zealand, probably the two most closely studied helmet law jurisdictions:


The NZ chart does not show the figures for the whole population - these follow pretty much the adult trend. Reductions in head injuries seem to follow enforcement of traffic laws, and modal shifts away from road cycling. This is obvious, since cars are the major source of serious head injury.

So: helmet laws fail for a number of reasons, including:

  • they are sold on traffic danger, but helmets are not designed for traffic crashes
  • they are sold with grossly exaggerated efficacy figures
  • cycle injuries are rare so enormous numbers of helmets are required to yield measurable reductions in
  • aggressively enforced laws deter cycling and reduce "safety in numbers" benefits
  • risk-averse cyclists are already wearing helmets, and the rest may risk compensate away any benefit.


Now we come to the vexed issue of helmet wars. Helmet threads are inevitable, the consequence of a clash of deeply held beliefs. They are also - contrary to what many might think - productive, at least on uk.rec.cycling. It has been suggested that nobody changes their mind as a result of a helmet war, but actually that is not true, a lot of us started as strongly pro-helmet and have had our eyes opened by the presentation of evidence which - let's be honest - helmet proponents don't so much as address, as sweep under the carpet. It often comes as a shock to people, even experienced cyclists, to find that there is any evidence at all to contradict the 85% figure they've been presented with for over a decade.

They are productive up to a point, the point where a helmet thread becomes a helmet war. At this point positions have been taken, trenches dug, and anybody going "over the top" is likely to come under heavy fire from the opposite side. The debate tends to become robust (read: seriously bad tempered). It is only a matter of time before someone posts one of the classic Liddite throwaways:

Anybody who doesn't wear a helmet has nothing to protect A variation on proof by assertion, the poster is essentially making a statement of faith. This is no different to an evangelical preacher telling unbelievers that they are damned to hell. The existence of other faiths which say the same thing is never seen as relevant in this context...

For what it's worth I know a lot of very intelligent, very well-informed people who do not wear helmets. Some of them even think that helmets might prevent, or reduce the severity of, a meaningful number of injuries. It's just that none of them are convinced that the risk is that great to start with.

My helmet Saved My Life™! If you believe the stories, helmets have saved more lives than were ever lost before they were invented. There are two fallacies in one here: first, that the crash in question would have been fatal; and second, that the magic hat alone made the difference. Time trends and population studies indicate that this is simply untrue, at both levels.

But if it is true then you have also to accept that in a crash on Par Street Roundabout in 1984 my Milletts knitted acrylic balaclava Saved My Life™, and because it was not made of polystyrene I still have it and it might do the same again one day.

"All the studies agree" It is true: a large umber of former medical students agree that, all other things being equal, you will be less likely to be injured if you wear a helmet. And a lot of risk experts agree that all other things are not equal. But if that particular penny has not dropped by now we are probably wasting our time here, aren't we? ;-)

For an example of how twenty billion flies can be wrong, read the page on observational studies.

There is no downside Have you ever wondered why cycle helmets are full of holes, and why the standards have been diluted over the years? Heat and weight. Cycling is a physical activity. Riders in the major stage races are allowed to remove their helmets for the last climb, to avoid overheating as they race for the finish. Heat is one downside. Discomfort is another (and yes, you will often not notice this until you have tried riding without one for a while and then put it back on). Risk compensation is a third. I'm sure there are others. As usual with simplistic arguments, it is simply wrong.
You are anti-helmet! I do know one or two people who are anti-helmet, in the sense that they foster strong resentment of the way helmets have come to dominate the cycle safety agenda, against all sense - they are, after all, a device designed to reduce the severity of low speed impacts, they prevent precisely no crashes at all. But I have never come across anyone who would go up to a helmeted cyclist and berate them for wearing one. I have, on the other hand, been berated myself for not wearing a helmet (and guess what? the person doing the berating was a car driver).

This displays a True Believer's inability to distinguish between scepticism and atheism. And it usually precedes a straw man argument: "you are anti-helmet, helmets prevent injuries, therefore you are wrong..."

Doctors call unhelmeted cyclists "donors" Cobblers. Not least because the major source of cyclist fatality is road traffic crashes, in which cyclists will often sustain massive internal damage. The leading cause of cyclist death in London is left-turning lorries, leading to crushing injuries of the body. Motorcyclists - of all types - make much better donors because they are more likely to be involved in nice, clean single-vehicle crashes, and to die of a broken neck.

For some reason those who post these arguments seem to get upset when the response is widespread withering scorn. Who could have predicted that? That is such an atypical reaction to the assertion that one person has the monopoly on truth!

Here is a point to ponder: the people putting the sceptical viewpoint have often read dozens, sometimes hundreds, of helmet papers. Those arguing the points above usually turn out not to have read any, beyond the carefully-edited abstracts recycled by pro-helmet campaigners. Hmmmm.

Shoot the Messenger

One reason why I am suspicious of pro-helmet arguments (and I know others share this) is that the most vociferous proponents of helmets tend not to be too scrupulous. In the UK, the leading pro-helmet group is BHIT. They have been responsible for a number of dodgy claims, which I detail in this table.

Claim Facts
One child under the age of 16 dies every week in the UK of head injuries From 2000 to 2002 the average number of child cyclists killed due to head injuries was 12 per annum - very much less than the 52 suggested.
... and a further 60 are admitted to Accident and Emergency departments with serious head injuries In 2002, the number of serious child cycling head injuries admitted to A&E departments in England was 384 [3]. Pro-rata population, that would be approximately 460 for the UK. This equates to an average of under 9 per week. Most of these children made a full recovery.
Over 70% of child cyclist deaths involve head injury Over 70% of all impact deaths involve head injury. In fact, 82% of cyclists and 86% of pedestrians and motor vehicle occupants who die in crashes suffer lethal head injuries. 71% of cyclists die primarily from head injuries, more than the other groups. However, this is not because a fatality is more likely to involve head injury for a cyclist (it is not), but because cyclists suffer fewer lethal injuries to the thorax and abdomen.
30 per cent of children's head injuries admitted to hospital are due to cycle injuries In 2002, cycling was implicated in only 7.1% of all child head injury admissions in England [3]. Similar data from Australia, pre-law, shows that child cyclists account for only 8% of child head injuries from all causes.
Over 100,000 people under 16 are treated in hospital each year due to a cycle accident.

And 60% of those sustained an injury to their head or face

In 2002, there were 5,804 child hospital admissions in England associated with cycling (6,965 pro-rata population for UK). Of these, 2,183 (2,620 UK) involved head injury. The proportion of head injuries, 37.6%, is lower than for child pedestrians (43.7%) and only a little above the average for all child admissions (34.2%).
In real terms [helmet compulsion] equates to 20,000 young people being spared such tragedies each year. As stated above, and adding the 10 head-injury deaths in the UK the same year, gives a figure for might justly be called 'tragedies' of 470. Of these, one-third were associated with parts of the head for which helmets offer no protection. Other casualties involved multiple serious injuries, and an unknown number of the remaining children wore helmets. An unknown proportion exceeded the limited protective capabilities of helmets anyway, but even if helmets were 100% effective, fewer than 315 'tragedies' (2/3 of 473) could be avoided.
The savings in healthcare costs alone would approximate to £2,000,000,000 annually There is no real-world evidence that cycle helmets are effective in preventing serious or fatal injuries. However, if they were as effective as claimed, the total costs saved in preventing the 2001 count for England, 12 fatalities and 300 serious injuries, would be £47,867,880. The amount claimed, £2,000,000, is around a third more than the NHS spends on all injury treatments for children in the age group 5-15 for all causes, according to figures from the DoH website. The dominant causes of childhood mortality are congenital disease and cancer, the major cause of traumatic death is road traffic crashes, of which the great majority do not involve bicycles.
In Australia, teenage cycle use fell because the helmet law was introduced at the same time as the driving age was lowered The driving age was lowered, from 18 to 17, in only one Australian state - Victoria. In all other states there was no change. In Victoria at the time the helmet law was introduced, 17 year olds accounted for 6% of all traffic. Even if every cycling 17-year old had changed over completely from cycling to driving, that would account for only a small part of the 43% drop in cycling by Victorian teenagers that followed the law. Nor would it account at all for the decline of up to 60% in cycling amongst teenagers seen in other Australian states.

In a bizarre twist on the BHIT ASA complaint ruling, BHIT complained to ASA about a briefing document for MPs which contained a section much like the above headed "myths" and "facts". ASA decided that the Parliamentary briefing was a marketing communication (selling what?) and ruled that although the myths were indeed false claims made by the helmet lobby, and the facts were indeed accurate, because there was dispute about the extent to which the (acknowledged true) facts rebutted the (acknowledged false) claims, this was misleading. Er, right.

BHIT has derived much income from grants from the Department of Health and Department for Transport - that's you and me. Thus far what they have delivered in return for around £100,000 of public money is a couple of reports claiming that helmet promotion can increase helmet use (really? who would have thought it?) but containing too little data to evaluate the other claims contained, a conspicuously inaccurate Early Day Motion, and campaigning for a helmet law. These political campaigning activities must be done out of their unrestricted funds, but much of their income (e.g. Government grants) is restricted.

BHIT have also engaged in some seriously distasteful emotive campaigning. When Troy Parker was killed riding off the pavement into the path of a car on a bike with defective brakes, they brought his grieving mother to Westminster to press the case for a helmet law. Leaving aside for a moment the fact that there is no proof a helmet would have saved his life, as the Coroner said, it was riding on the footway and having defective brakes which caused the crash. And they have a video about a boy who is reduced to a vegetable after crashing his birthday present bike (unhelmeted). Curnow shows that this kind of brain injury is most likely to be caused by a type of force against which helmets provide no protection, and could even exacerbate. But BHIT make no claims to be a cycle safety charity - their sole interest is in promoting helmets.

In short: the Government is paying them a lot of our money to provide expert information on cycle helmets. They do not do this. They are propagandists. In addition to the dodgy claims above, they also continue quote, uncritically and often without even the "up to" qualifier, the 85% figure - even after accepting to the ASA that they should not make such claims.

In the USA helmets are strongly promoted by the National SafeKids Campaign. Would you be surprised to learn that they are sponsored by Bell? Probably not. Especially when you hear about recent moves in the USA towards helmets for soccer... Guess what figure they quote in their helmet promotion campaign? Yup, 85%.

Which invites the question: if helmets are such a good idea, why is it necessary to lie in order to promote them? I'll leave you to ponder that one at your leisure.

Can we stop now?

Right, so now you can quote at least three studies to support your pre-existing point of view, whatever that might have been. So how has that helped, exactly?

It has helped in this way:

  • you are better informed
  • you understand that there is no simple answer to the question of whether helmet use is good or not
  • you realise that simplistic arguments - on either side - are almost invariably incomplete, usually to the point of uselessness.

And now you are ready, Grasshopper, for the Great Truth: When cycle safety interventions are ranked by likely effectiveness, helmets are always placed last. The reason is simple and obvious. It is risk management 101: first reduce risk at source, then reduce exposure to risk, and only when this fails apply personal protective equipment.

CS Downing's "Pedal Cycle Accidents in Great Britain" (TRRL, 1985) rated the efficacy of interventions in reducing cyclist casualties thus:

Continuation of existing measures Reduction
Road engineering (Cycling facilities, blackspot treatment) 10%
Education, training, publicity, and enforcement (eg increased use of 'Cycleway', better penetration of publicity) 5%
Measures developed but not yet widely introduced
Wide spread application of area wide low cost engineering measures 5%
Bicycle engineering/construction (including pedal cycle helmets) 1%
Car external softening 3%
Heavy goods vehicle side guards 2%
New initiatives in education, training, enforcement and publicity aimed at improving road user behaviour (drivers and cyclists) including better use of new conspicuity aids and lights by cyclists 5%

Note that this was before helmets really started to take off, and risk-averse cyclists had just started to wear hard shell helmets.

The fundamental flaw with helmet promotion is that it ignores this principle, and elevates cycle helmets - a palliative measure, and one designed only for a small subset of the accidents which cause serious injury - to first place. In fact, only place. Helmet lockers are marketed as a "cycle safety programme"; helmets "help children cycle safely" But this is arrant nonsense! Cycle safety is not about planning for when you hit the ground, still less about making it less unpleasant to do so - it's about not hitting the ground in the first place!

So you can join the urc regulars in singing our variation on the Blue Öyster Cult classic:

You see me now, a veteran
Of a thousand helmet wars
I've been living on the edge so long
Where the winds of limbo roar
And I'm young enough to look at
And far too old to see
All the scars are on the inside
I'm not sure that there's anything left of me

About the author: Unlike some people Guy Chapman does not consider himself one of Earth's foremost authorities on helmets, although he knows some people who are. He has, however, read an awful lot of helmet papers.