Citation

Young DJ, Stockwell TR, Cherpitel CJ, Ye Y, MacDonald S, Borges GLG, Giesbrecht N. J. Stud. Alcohol 2004; 65(5): 605-612.

Affiliation

National Drug Research Institute, Curtin University of Technology, Perth, Western Australia.

Copyright

(Copyright © 2004, Rutgers Center of Alcohol Studies)

DOI

unavailable

PMID

15536770

Abstract

OBJECTIVE: This study describes and examines the development of surrogate measures of acute alcohol-related injury for use in the evaluation of community-based prevention initiatives. METHOD: An international collaborative study of alcohol and injury, the Emergency Room Collaborative Alcohol Analysis Project (ERCAAP), provided a subset of data on 8580 emergency room (ER) presentations from five countries and 28 ER facilities. RESULTS: Presentations most likely to be alcohol related were those made between 12:00 AM and 4:59 AM (56%), on Fridays, Saturdays or Sundays (26%); and those among injured persons who were male (28%), who were aged between 18 and 45 years (24%) or who were unmarried (24%). Multilevel logistic regression models confirmed the significance of the above variables as predictors of alcohol involvement prior to the injury event. The strongest predictor variable was presentation between 12 midnight and 4:59 AM with an odds ratio of 4.92 (Wald Test chi2 = 397.6, p < .001). Being male had an odds ratio of 3.01 (Wald Test chi2 = 247.25, p < .001), and presenting on a Friday, Saturday or Sunday night had an odds ratio of 1.50 (Wald Test chi2 = 49.6, p < .001), whereas being under 45 (odds ratio [OR] = 1.20, p < .05) and being unmarried (OR = 1.2, p < .01) were less strong predictors. Combining all these values for variables raised the probability of prior alcohol involvement in such injury presentations to 0.65, although only 3.37% of all cases met these criteria, limiting applicability of this combined variable as a surrogate measure for intervention studies. Probabilities of prior alcohol involvement are presented with other combinations of values for the predictor variables. CONCLUSIONS: Frequency of nighttime injury presentations to ER facilities, particularly by men, can be used as a reliable surrogate measure of alcohol-related injuries for various epidemiological and evaluation purposes.