Drinking and driving is a major public health problem, but neither the actual incidence nor prevalence of the problem is known. The development of truly effective means for preventing, deterring, and intervening continue to be hampered by this lack of information. Long-Term Objectives: The present proposal is designed to address this problem with a new and more differentiated approach, one which should provide the methodological basis for accurately determining the actual incidence and prevalence of driving after drinking. This approach should also provide the statistical basis for estimating the probabilities of driving under the influence of alcohol (DUI) among the general public, as well as among the motoring public--including drivers convicted of DUI. This approach is based upon four concurrent, interrelated projects which are focused on different, but interdependent segments of the American drinking and driving public: (1) general adult population, (2) nocturnal driving population, (3) convicted first DUI offender population, and (4) convicted second DUI offender population. The basic rationale derives from analyzing the known characteristics in these interrelated populations to determine the similarities among those who """"""""get into trouble with alcohol,"""""""" as well as the differences between them and those who do not.
The specific aims consist of developing, pilot testing, and cross-validating statistical methodologies (based upon the four populations and prior research) for: (1) estimating the probabilities of driving with high, low, or no alcohol for individuals in selected subsets of the general population; (2) estimating the incidence and prevalence of driving at zero, low, medium, and high blood alcohol concentrations; and (3) early identification and diagnosis of high-risk problem-drinker drivers.
The specific aims will be accomplished through multivariate analyses of three domains of data for most subjects: (1) self-reported background data, plus drinking, driving, and drinking-and-driving information and attitudes; (2) official driver record, both pre-and post-contact, and (3) breath alcohol at selected times (e.g., when arrested, appearance at sessions, and probation reviews). Data in these three domains will be used for selected cross-validations, as well as for subpopulation description, prognoses regarding both alcohol program compliance and future drunken driving, and early identification and diagnosis of high-risk drinking drivers.
