Bicyclists, Helmets and Head Injuries: A Rider-Based Study of.Helmet Use and EffectivenessRichard C. Wasserman, MD, MPH, Julean A. WALLER, MD, MPH, Melanie J. Monty, BS,Alice B. Emery, AB, and Danielle R. RobinsonTABLE 1.—Characterlstlcs of Overall Study Group and of Helmet UsersAbstract: We interviewed 516 bicyclists over age 10regarding helmet use and head injunes. Although 19 per centowned helmets, only 8 percent were weanng them wheninterviewed. Riders wearing helmets were more highlyeducated and reported higher car seat belt use. Nearly 4 percent of the bicyclists reported striking their heads in a cyclingmishap during the previous 18 months; those wearinghelmets at the time of the mishap were less likely to havesustained head injuries. (Am J Public Health 1988; 78:1220-1221.)IntroductionBicycling injuries account for at least 1,000 deaths1 and500,000 emergency room visits2 each year in the UnitedStates. Head injuries account for approximately 85 per centof bicycling deaths3 and two-thirds of bicycle related hospitaladmissions.4 In light of these statistics, several authoritieshave suggested that bicyclists wear helmets. 5-7Few data are available to either support or refute thisrecommendation. In laboratory studies, many helmet brandshave been shown to absorb the necessary 300-400 G ofacceleration to minimize brain injury, but a literature searchreveals only one published study on the effectiveness ofhelmets in actual bicycle mishaps. Dorsch, et at,demonstrated an association between helmet use andreduced severity of injury in a mail survey of Australianbicycling enthusiasts.11Weiss recently reported the prevalence of helmet useamong adult recreational cyclists (85 per cent), universitystudents (210 per cent), and school children (2 per cent) inArizona,’ but no published studies address why bicyclistswear helmets.We therefore designed a study to address the followingquestions:- What is the prevalence of helmet use amongbicyclists in traffic?.. What factors are associated with bicycle helmet use?- Are helmets effective in preventing bicycling headinjuries?MethodsBicyclists were interviewed at roadside during July andAugust 1984 in and around Burlington, Vermont, a semi-urban university community with a predominantly Whitepopulation of 120,000. Interviewers were stationed at 16roadside locations during daylight hours on all days of theweek. Using hand-held signs. they attempted to stop andinterview all bicyclists who appeared to be over the age of’10.Characteristics% of Sample(N—516)%% Wearing Helmet(N (N=40)Age11—19 41.9 3.7 20-2939.0 8.0 30+ 19.1 16.2Sex Male64.4 8.5 Female35.6 6.6Education (years completed) <=12 43.0 3.6 13 -1643.0 7.7 > 16 14.0 20.8Martial Status(if >age 16)SingleMarried76.823.2 8.0 17.1Cigarette Use Nonsmoker90.3 64 Smoker 9.72.0Seat Belt Use (self estlmate of % of time belt worn) <25% 25.2 0.0 25-75%19.8 3.9 >75%55.0 12.7Belief in Head injury susceptibility Unlikely to Hit head Not unlikely to Hit Head 54.645.45.710.3Belief in Head lnjury Seriousness Not serIous44.1 4.4 SerIous55.9 102Belief In Bicycle Helmet Effectiveness Not effectIve18.1 5.4 Effective 81.98.3Five hundred sixteen (516) of 683 cyclists (76%) stopped andconsented to a five minute interview. There were nodifferences in helmet use or sex between those who stoppedand were interviewed and those who did not.The interview solicited information about demographics,bicycling practices, attitude toward head injury and helmetuse, health-related practices (e.g., cigarette smoking). andbicycling injury experience over the previous 18 months.ResultsThe subjects interviewed had a mean age of 23.4 years(SD ± 10.1) and mean education of 13 years (SD ± 3.6); 58per cent were students.Although 18.8 per cent of the respondents reportedowning a helmet. only 7.8 percent were wearing helmets atthe time of’ interview. The most common reasons given fornot wearing a helmet were that the rider was on a short trip(28 per cent). that helmets were uncomfortable (24 percent)and simple negligence (13 per cent).Table 1 shows selected characteristics of the studygroup and the per cent of helmet wearers in each categoryAmerican Journal of Public Health (AJPH) September 1988, Vol 78, No, 9PUBLIC HEALTh BRIEFSPUBLIC HEALTh BRIEFSTABLE 2—Multlvariate Dlscriminant Analysis to Predict Helmet UseVariablesStandardized Canonical Discri. Function Coefficienta CanonicalCorrelationbAge*-.12Male Sex*.23Education*.85Married*.46.231Age*.21Male Sex*.22Education*.46Seat Belt use**.54Belief in susceptibility**.41Belief in seriousness**.30.338a-Reflects relatIve contribution of variable to the discriminantfunction.b.When squared reflects the proportion of variance in groupmembership explained by the discriminant function* Entered directly.** Entered stepwiserider’s belief in personal susceptibility to and seriousness ofhead injury appeared associated with helmet use.Multivariate discriminant analyses were performed toexamine the relative contributions of these variables inpredicting helmet use. Results are seen in Table 2.in the analysis of demographic factors alone, thediscriminant function coefficients suggest that education andmarital status contribute most strongly in predicting helmet use.In an analysis controlling for age, sex, and education, seat beltuse and belief in susceptibility to and seriousness of headinjury assume importance in predicting helmet use. Theproportion of variance in helmet use explained by this seconddiscriminant function, however, is a modest 11.4 per cent.Head injury experience for the previous 18 months was asfollows: 21 bicyclists (4 per cent) had struck their heads in amishap, and seven (1 per cent) reported sustaining head injury,three with concussions and four with lacerations requirIngsutures.Of the 21 riders who reported falling and striking theirheads, eight were wearing