Accuracy of drinking behavior recall: Comparing late night BAC by breath analysis to estimated BAC from linked survey responses in a residential college environment

Assessing risk due to alcohol consumption in collegiate populations is often achieved using measures of the frequency of use and the quantity of alcohol typically consumed (number of drinks). Since body weight, gender, and the time spent consuming alcohol all influence the degree of intoxication, a better measure of risk rather than number of drinks is an individual's peak blood alcohol concentration (peak BAC). BAC can be measured in field tests directly using breath test instruments or estimated from survey data responses reporting the quantity of alcohol consumed, time of consumption, personal weight and height, and gender. 2,758 random anonymous double-blind late night breath tests have been conducted along with an anonymous but individual case linked survey of alcohol use just inside the entrance to residence halls in a collegiate environment between spring 2003 and fall 2012. Several methods for computing eBAC values have been evaluated by comparing their predictions to corresponding measured BAC values. An assessment of the accuracy of drinking behavior recall has also been conducted as a function of intoxication level. Finally, mathematical models have been developed to determine the eBAC during the pre-absorptive state and to assess the time after cessation of drinking required on average to reach full absorption and peak BAC for this college population. Results will inform survey researchers on the accuracy of personal alcohol use recall and on the relative accuracy of computational methods for eBAC determination.