Alcohol and highway crashes. Closing the gap between epidemiology and experimentation

It is frequently observed that alcohol-impaired drivers involved in certain types of crashes apparently 'couldn't stop in time'. However accurate this observation, it is simply a description and is not an adequate explanation. Therein lies the basis for the gap between epidemiology and experimentation. If alcohol actually does degrade a motorist's performance and increase the probability of his being responsible for a fatal crash, then alcohol-induced changes in driving behavior should be manifest and should be measurable. However, no controlled study has previously been conducted to obtain systematic but unobtrusive data on the actual influences of alcohol upon real-world driving behavior in its natural environment. One recently completed field study is reported which was designed to provide such data by means of unobtrusive electronic measures of nocturnal driving performance. In the present paper, to illustrate one approach to closing the gap between epidemiology and experimentation, one combination variable was selected which lends itself readily to comparison across the full spectrum of alcohol investigation: reaction time and braking performance. A review of the literature concerning this interrelated variable examined alcohol influences upon reaction time as investigated in laboratory, simulator, and instrumented car experiments, as well as alcohol influences upon braking performance in instrumented car experiments and in our recent field study involving unobtrusive electronic measures. The reviewed experiments were interpreted in terms of a conceptualization of the information-processing sequence which leads up to brake use. On the basis of the reviewed results, it was concluded that alcohol increases reaction time (both simple and choice) appreciably more in driving situations than in laboratory experiments. It was also concluded that a consistent alcohol impairment of the qualitative aspects of braking performance is manifest in driving situations, for example, as reflected by changes in brake-pressure modulation. More specifically, the braking performance of motorists and subjects at high BACs is abrupt, unsmooth, and less controlled than that of sober motorists or the same subjects with no alcohol. It was concluded that at least in terms of quality, alcohol impairs response implementation of the final stage of the information-processing sequence. It was hypothesized that the abrupt, lower quality braking performance observed at high BACs results--at least in part--from less time remaining available for stopping because the information-processing time has been increased by alcohol at some previous stage--apparently the response-selection stage. At the most general level, it was suggested that high BACs both increase the time necessary to begin applying the brakes, as well as reduce the degree of control in the actual use of the brakes during the course of stopping. These two factors in combination probably account for a large part of the alcohol contribution to highway crashes.