The Effect of Motorcycle Helmet Use on the Probability
of Fatality and the Severity of Head And Neck Injuries

Footnotes

¹The before-after methodology is employed by Dare et al. (1979) and McSwain and Lummis (1980), while helmeted- nonhelmeted comparisons are found in Chang (1981), Dare et al. (1979), Heilman (1982), Hurt et al. (1981a, 1981b), Kraus et al. (1975) Luna et al. (1981) and Scott (1983).

²The systematic overrepresentation of nonhelmeted riders in accident samples is a manifestation of the relation between helmet use and risk-averse driving behavior. Dare et se. (1979; p. 14), Hart et al. (1975; p. 544), Heilman et al. (1982; p. 663), Hurt (1981a; p. 6), Mueller (1980; p. 590), and NHTSA (1980; p. IV-21) either document this occurrence and/or discuss this relation. Scott (1983; p. 33) establishes the relation between alcohol use and helmet use.

³Data supplied by the Motorcycle Industry Council Inc. reveals that between 1976 and 1980 the percentage of total motorcycles 450cc and over increased from 21.9% to 37.8% and that the percentage of vehicles 750cc and over increased from 11.0% to 22.4%. Between 1976 and 1982 the average annual miles traveled per motorcycle increased from 1525 to 2955. Between 1975 and 1980 the percent of total motorcycle owners under the age of 18 increased from 16.2% to 24.6%, while the under 24 group increased from 38.1% to 48.9%.

⁴See Snively (1983).

⁵Federal Standard No. 218 requires that motorcycle helmets pass two distinct impact attenuation tests. The impacts are generated by a guided free fall that results in impact velocities of 11.66 and 13.40 mph.

⁶Relative velocity is defined as 4(v cos 0 + V)2 + (v sin 0)2 where v is the crash speed of the motorcycle, V is the crash speed of the other vehicle and 0 is the angle of impact, where 0 < 0 < 180.

⁷It is assumed that the most severe injury is associated with the largest use of energy. Thus if another vehicle is involved in that injury, the rider's velocity must be calculated relative to the other vehicle. In all other cases, it is assumed that the rider does not impact another vehicle but rather a fixed object. Qualitatively and quantitatively similar results, to those reported below, are obtained for a third variant of kinetic energy- -one which uses the relative velocity in all instances.

⁸Given that multiple injuries and thus multiple injuries mechanisms were reported compressibility was based on the nature of the injury mechanism associated with the operator's most severe injury. A qualitative measure was used to distinguish between compressible and less compressible objects. The latter group included environmental factors composed of asphalt, concrete, metal, and wood along with the "hard points" of vehicles as defined by Hurt et al. (1981a; coding appendix E). The former group included glass, water, soil, dirt, sand, and gravel and the "soft points" of other vehicles. The statistical results can be explained by the small variation in the compressibility of the typical objectives impacted in road accidents and the minimal amount of deformation (energy absorption) incurred by such objectives.

⁹A continuous relation exists between age and reduced pulmonary functions, reduced cardiovascular reserves, particularly under stressful situations, brittle bones (osteoporosis), rigid ligaments, and coexisting diseases which may complicate the process of homeostasis.

¹⁰See Baker and Fisher (1977) and Champion et al. (1975) and references therein.

¹¹Vasodilatation and the blocking of antidiuretic hormones are two such problems. See Champion et al. (1975) and the references therein for further discussion.

¹²The AIS developed by the American Association for Automotive Medicine (1976), classifies injuries using the following scores: zero, no injuries; 1, minor injuries: 2, moderate injuries; 3, severe injuries--no threat to life; 4, serious injury--life-threatening, survival probable; 5, critical injury--survival uncertain; and 6, fatal injury. Under this classification system, the cumulative effect of multiple injuries is measured by the sum of squared AIS.

Head injuries are defined as those occurring in the following regions: Basal, Frontal, Face, Mandible' Maxilla, Nasal, Occipital, Orbit, Parietal, Brain, Sphenoid, Temporal, and Zygoma. Alternative specifications of the HS variable which exclude different combinations of regions considered to constitute the face were tested and the results did not deviate qualitatively from those reported below.

¹³The lower truncation in this case is zero. Out of a sample of 644, 248 were nonlimit observations.

¹⁴Neck injuries are defined as those occurring in the following regions: the general cervical area, cervical vertebrae 1-7, and the foremen magnum. Alternative specifications of NS which include different combinations of the above regions and in some cases the throat region produce the same qualitative results.

¹⁵Out of a sample of 644, 68 were nonlimit observations.

¹⁶The average weight of the human head is 8-12 pounds while the average weight of the helmet used in our sample is 2.7 pounds. m us the weight of the helmeted head increases by 23-34%. The helmet literature has paid little attention to the relationship between helmet use and neck injury. For example an analysis of the this relation has never been an objective of NHTSA research, see NHTSA (1980; p. II-5). The overall quality of the statistical analysis of this issue is significantly below that of the fatality and head injuries studies criticized above and empirical findings have supported both sides of the issue. Studies that suggest a positive relation are found in Bowman and Schneider (1980), N.Y.S. DMV (1969), Dare et al. (1979) and the references cited in Beier et al. (1983; p. 596) and Voge and Borowsky (1983; p. 606). Studies that support a negative or no relation include: McSwain et al. (1980), Hurt et al. (1981a, 1981b), N.Y.S. DMV (1979), Scott (1983), Bowman and Schneider (1980) and the references cited in Mueller (1980) and NHTSA (1980).

¹⁷For a detailed discussion of this methodology and its relative merits, see Hurt et al. (1981a, pp. 1-35).

¹⁸Variables for which missing values were deleted include H, EA, A, BA rider height, weight, motorcycle crash speed, other vehicle crash speed, coefficient of breaking friction, traffic density, marital status, drug impairment, and precrash separation of rider from vehicle. Means were assigned in the following cases: EX, training, operator education, number of children' income, number of prior tickets and accidents, and the normal component of helmet impact velocity.

¹⁹Fatality rates per 100 accidents, reported in Dare et al. (1979), McSwain and Lummins (1980), and Scott (1983), range from .0109-.0292 and are consistent with our estimates of .0228 and .0262. Alternatively, partial derivatives, ~~ , evaluated at sample means are reported in Appendix B.

²⁰Using the point estimates in equation 3 and 4, the critical helmet impact speed beyond which helmets no longer reduce head injuries are 38.31 and 41.29 respectively. While such impact speeds are possible, experience shows that they are outside of the normal range (0-25 mph) of impact speeds, see Hurt (1981a' VI, Sec. 9).

²¹Different variants of HD and ND, where these variables are assigned a value of 1 either if AIS > 3 or AIS > 4 are tested. The results are qualitatively the same as those reported below.

²²Exclusion of all insignificant variables with the exception of H, and HI in equations 9 and 10 produce the same qualitative results.

²³See Mueller (1980), Hartunian et al. (1983), and Scott (1983).

²⁴Deviations between individual costs and societal costs may result from the structure of insurance rates which tend to redistribute the high costs associated with high risk policy holders to all policy holders.

²⁵See Peltzman (1975) and Wilde (1982). For the case of helmet laws' Adams (1983) offers empirical support for this hypothesis.

²⁶Horsepower restrictions have been considered on the European continent, see Russo (1978) and the references therein.

© Copyright Jonathan P. Goldstein Ph.D. 1986. All Rights Reserved.