Helmets serve several purposes:

1. They provide protection against serious head injury.
2. They make the cyclist more visible.
3. They have aerodynamic advantages. An old-fashioned "hairnet" type of helmet is aerodynamically superior to riding with no helmet; modern helmets are better still.
4. Although some people (usually those who never wear one!) claim helmets are hot, they can in fact be more comfortable in hot sunny conditions than riding without a helmet.  A light-colored helmet will reflect energy that would otherwise be absorbed by the rider's head. Most modern helmets are well-ventilated; some actually provide better airflow than riding without a helmet.

Recommended Reading:

• Effective Cycling, John Forester, 1993 (sixth edition), ISBN 0-262-56070-4 (paperback), 0-262-06159-7 (hardcover)
  • Chapters 2 and 28 have important information regarding helmets and the type of fatalities and permanent disabilities that they can prevent. The following quote from Chapter 2 is particularly relevant:

    Wearing an adequate helmet is strongly recommended. Riding so you are less likely to get into accident situations is the first safety measure; watchfulness and skill in escaping them is the second; but when all else fails, you need to reduce the injury. Helmets are the best investment in injury reduction.

    In Chapter 28, Forester informs the reader that

    All told, three-quarters of the cyclists who are killed die of brain injuries, and probably three-quarters of the permanent disabilities are from brain injuries. Wearing a good helmet is the best way of increasing your chances of surviving an accident without permanent disability.

• Serious Cycling, Edmund R. Burke, PhD, 1995, ISBN 0-87322-759-X
  • Chapter 11 Contains an informative section entitled Why Wear a Helmet?, and another entitled Hot Heads; the latter references research which addresses the myths about helmets and heat.
    
    
• Training For Cycling, Davis Phinney and Connie Carpenter, written with Peter Nye, 1992, ISBN 0-399-51731-6
  
  
  • Good advice from world-class cyclists. Section 4 has some Heat Acclimatization Tips, which quotes team doctor Massimo "Max" Testa's recommendation to wear light-colored head cover to shield the back of your head from the sun. 
    
    
• Science of Cycling, Edmund R. Burke, PhD, editor, 1986, ISBN 0-87322-181-8 (paperback), 0-87322-048-X
  
  
  • Currently out of print, this book has some information regarding aerodynamic advantages of helmets (in Chapter 6) as well as their value for injury prevention.

Web Resources:

• Bicycle Helmet Safety Institute
• [U.S. Centers for Disease Control and Prevention] documents regarding bicycle helmets
• [U.S.] National Bicycle Safety Network
• Snell Memorial Foundation
• World Health Organization Helmet Initiative
• Bicycle Helmet Data
• Some Helmet Manufacturers:
  • Bell
  • Cratoni
  • Louis Garneau
  • Giro

Considering the advantages of helmets mentioned above (note that the improved visibility is one aspect of avoiding accident situations) and the dire consequences of serious brain injury, wearing a helmet is a prudent safety measure.  That is, of course, a personal choice which each individual must make for himself.  As a humanist, I feel that it is important that people should have objective and unbiased information on which to base that decision.  In that spirit, I present the following information regarding helmets, primarily to counter some one-sided anti-helmet misinformation which is often repeated.

There is a great deal of misinformation floating around about helmets. Much of it comes from a small group of extremely vocal and dogmatic anti-helmet crusaders. Careful examination using critical thinking skills often reveals that the very "evidence" that the anti-helmet crusaders claim shows no effect of helmet use in reducing injuries, in fact does show a substantial benefit to cyclists. Most of the time, however, it is necessary to carefully analyze the data, as the anti-helmet crusaders often present the data in a misleading and distorted manner. As an example, consider the often repeated claim that hospital data shows no change in the proportion of head injuries when helmets are used by a large proportion of cyclists.  First, there are several problems with the unstated assumptions that go along with that claim:

1. Hospital data, obviously, only includes those cyclists who go to a hospital. A cyclist with minor injuries might not wish to make a trip to a hospital. The vast majority of injuries are of this type, and are therefore not reflected in hospital data. Cyclists who avoid trips to the hospital by escaping serious injury -- and that quite likely includes more helmeted cyclists than unhelmeted ones -- don't show up in the hospital data. The ones who do show up are the ones with the serious injuries.  Therefore, it is quite likely that hospital data for proportion of head injuries won't show the full benefits of helmet use; it might even show increases in that proportion (declining numbers of cyclists treated, on the other hand, may reflect helmet benefits, for the same reason).
2. Head injuries can include bloody noses, dirt, bugs, or other foreign objects in the eyes, chipped teeth, and minor facial scratches and abrasions. Helmets are not intended to protect against these types of head injuries; what should properly be considered are those injuries which helmets are designed to protect against -- cranial fractures and brain injuries, primarily.  Lumping all types of head injuries together dilutes the effect of bicycle helmets, which are not designed to, and therefore cannot, protect against all types of head injuries. Reputable studies take this into account by considering "injury to those areas of the head that a helmet might reasonably be expected to protect -- the forehead, scalp, ears, skull, brain, and brain stem" [quote is from Thompson et al., The New England Journal of Medicine, 1989; vol. 320, issue 21, pages 1361-1367]
3. There are often confounding factors which may affect the number and type of injuries. It is necessary to take such factors into account. One way of doing so is to compare the proportion of head injuries for cyclists to that for pedestrians (and possibly for motorists and motor vehicle passengers as well), since the latter generally don't wear bicycle helmets.  Any common factors can be removed from consideration by looking at either the difference in proportion of injuries or the ratio of the proportions. Such common confounding factors may include changes in road conditions, changes in enforcement of motorist speed limits and/or drunk driving, road safety education, etc.; these could result in either increases or decreases in serious injury rates for all road users over time.  These confounding factors are discussed briefly in a paper by Bruce Robinson, however it is odd that that same paper doesn't take those factors into account by showing the difference in head injury proportions.
4. There are random variations in the number and type of injuries sustained by a population over time. It is necessary to carefully consider the magnitude of that variation, as it may be quite large. Reputable studies present their results as an interval within which the established result lies, to some specified degree of confidence. Usually this is presented as a 95% confidence interval (occasionally a 99% interval).  Few anti-helmet screeds present any kind of range; simply bald assertions.

Much of the "evidence" presented by the anti-helmet zealots lacks any controls for the confounding factors mentioned in #3 above. One notable exception is the data for Western Australia. Note that this data still suffers from the limitations mentioned in #1 and #2 above, even though one author describes such limited data as "almost complete coverage".  Using the Western Australia data to control for the confounding factors of item #3, and bearing in mind that the other limitations of the hospital data discussed above still apply, it is still worth looking at this data in some detail, as it appears to be the best case that the anti-helmet crowd can muster to support their claims.  A graph of the proportions of head injuries to other injuries for cyclists and for pedestrians in Western Australia appears as Figure 8 in a paper by Bruce Robinson. Note that there is another tactic used by the anti-helmet zealots to report the numbers in such a way as to make things look bad for helmeted cyclists. The issue of minor "head injuries" has already been mentioned in #2. Rather than report the ratio of head injuries to other injuries -- which is bad enough as it already includes minor injuries unrelated to helmets -- at least one anti-helmet crusader has instead reported the proportion of individuals having any head injuries. Note that a cyclist with a scratch on his chin (a minor non-helmet-related head injury) and with a compound fracture of his leg would be reported in the latter case as a cyclist with a head injury, while the serious non-head injury would be ignored entirely. For example see this graph. Neither of these graphs shows what is really of interest, namely the difference (or ratio) in the proportion of head injuries for cyclists and for pedestrians. When one plots that difference, it becomes apparent why the anti-helmet zealots don't show it. The following graph uses the data points interpolated from the aforementioned Figure 8:

Helmet use was already significant prior to the mandatory helmet law in Western Australia, which went into effect on January 1, 1992. The chart above shows the difference in the proportion of head injuries for cyclists and for pedestrians, normalized to the proportion for pedestrians. The pedestrian data serves as a control for common confounding factors as mentioned in #3 above and in the text of Bruce Robinson's paper. Bars above the horizontal axis (in red) indicate relatively higher incidence of head injuries (including minor head injuries) for cyclists than for pedestrians; bars below the axis (in green) indicate that cyclists fare better than pedestrians. Clearly a substantial benefit accrued to cyclists as helmet use increased in the very late '80s and early '90s (1995, the last year for which data was reported by Robinson, is somewhat of an anomaly).  The minor up-and-down variations from the mid-'70s through the mid-'80s are due to the random variations mentioned in #4 above, and are not significant.  The display of the data above also eliminates some of the problems inherent in the presentation of the data in Robinson's Figure 8: the bars shown above for each calendar year extend the full width of the calendar year and are centered at the middle of each year, whereas Robinson's graph uses connected points where the calendar year data points are shown at the beginning of the calendar year even though the point represents data collected through the end of that calendar year.

Further insight can be gained by looking at the distributions of the proportions of head injuries for pedestrians and for cyclists. The following histograms show the distribution of the percentages from B. Robinson's Figure 8. The green lines are smoothed fits to the histogram data (technically, a bisquare kernel density estimate with a triangular kernel), while the red lines are the closest fits to the histograms for a normal (Gaussian) distribution.  The histogram (and its smoothed fit) for cyclists clearly shows two distinct peaks in the distribution; one in the 40's corresponding to no helmet use, and the other in the mid-20's corresponding to the proportion of head injuries with widespread helmet use -- this two-peaked distribution cannot be modeled as a single Gaussian distribution. The corresponding histogram for pedestrians shows only a single peak around 40%, and the close fit of the histogram bars, the smoothed fit, and the normal distribution speaks for itself.

Clearly the Western Australia data -- the only substantial data presented by the anti-helmet crusaders which has non-cyclist data that can be used as a control -- shows a substantial reduction in the proportion of head injuries for cyclists with increased helmet use.  However, the anti-helmet zealots will no doubt continue to make the same tired claim that the data shows no effect.  Many of the deceptive practices mentioned in the references on my critical thinking page are used in the materials presented by the anti-helmet crusaders; yet others, such as red herrings and ad-hominem attacks, are used by those crusaders against anybody who points out the flaws in their arguments.

It should be noted that the suggestion that helmet use should show up in reduced head injuries for cyclists compared to pedestrians, as reported by hospital data, originated from the anti-helmet crusaders. The bar chart above, based on the same data used by those crusaders, and showing precisely that difference that they suggested should be used, does show exactly the substantial helmet benefit which the anti-helmet zealots claim doesn't exist.  In other words, their arguments used with their data clearly shows that their conclusions are flatly wrong.  Since the argument and data are theirs, it is reasonable to presume that they knew that their statements regarding the supposed lack of effectiveness of helmets were wrong, and that they were therefore deliberately trying to mislead others.  They certainly know so by now, and continued claims that the data shows no effect can only be seen as deliberate attempts to mislead. It is wise to bear that in mind when considering any statements made by those anti-helmet crusaders.

Many other claims made by the anti-helmet crusaders can similarly be shown to be misleading. Another often-repeated claim is that helmeted riders take more risks, thereby offsetting any benefits in injury reduction.  Most of the time, that claim is presented without any supporting evidence whatsoever.  Nor do the anti-helmet zealots bother to point out that those cyclists who choose to use helmets are quite likely more safety conscious than those who choose to ride without the protection afforded by bicycle helmets. In fact, there is a study which shows that helmeted riders do in fact ride more safely than unhelmeted riders.

For the record, my position on mandatory helmet laws (N.B. helmets and helmet laws are two very different things!) is that I think that such laws are unnecessary and a waste of resources.  I feel the same way about laws requiring thirsty people to drink water.

Back to Cycling