The 131st Annual Meeting (November 15-19, 2003) of APHA

Abstract #68935

J. Quinn Schroeder, MS¹, Hsiao-ye Yi, PhD², Chiung M. Chen, MA³, Mary C. Dufour, MD, MPH⁴, and J. Quinn Schroeder, MS¹. (1) Alcohol Epidemiologic Data System, CSR Incorporated, 2107 Wilson Ave. Suite 1000, Arlington, VA 22201, (2) Alcohol Epidemiologic Data System, CSR, Incorporated, 2107 Wilson Blvd., Suite 1000, Arlington, VA 22201, 703-312-5220, hyi@csrincorporated.com, (3) Alcohol Epidemiologic Data System, CSR Incorporated, 2107 Wilson Blvd., Suite 1000, Arlington, VA 22201, (4) National Institute on Alcohol Abuse and Alcoholism, NIH, Willco Building, Suite 514, 6000 Executive Boulevard, Bethesda, MD 20892-7003

This study compared drinking measures derived from Quantity-Frequency (QF) and Graduated Frequency (GF) survey questions on alcohol consumption. Comparisons were made in terms their overall consumption volume estimates and predictive power for drinking and driving behavior. Data were drawn from the National Survey of Drinking and Driving Attitudes and Behaviors, 1999, administered by the National Highway Traffic Safety Administration. The GF drinking measures were constructed based on a stochastic model of drinking patterns developed by Gruenewald and colleagues. Consistent with findings from previous research, QF measures provided higher overall volume estimates at low levels of consumption and GF measures provided higher volume estimates at high levels of consumption.

In multivariate models predicting driving after drinking or driving while intoxicated, both QF and GF measures improved predictions made from demographic information (i.e., age, sex, race/ethnicity, education, income, employment, and marital status). GF measures provided slightly better predictions than QF measures in whether driving after drinking occurred, but almost no difference was found between the two measures in predicting whether driving while intoxicated occurred. Additional analyses were performed for those who reported drinking and driving and driving while intoxicated. In these analyses the number of times either behavior was reported was regressed on demographics and alcohol consumption measures. Both QF and GF measures provided similar predictions, with GF measures being slightly better in terms of predicting the frequency of driving after drinking and QF measures being slightly better in terms of predicting the frequency of driving while intoxicated.

Learning Objectives:

• Identify drinking patterns with increased risks for alcohol-impaired driving
• Assess differences between demographic and drinking pattern information in terms of their ability to predict alcohol-impaired driving
• Evaluate the relative predictive capacity of different measures of drinking patterns (i.e. graduated frequencies versus quantity-frequency)

Presenting author's disclosure statement:
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.