The study sample consisted of 127 patients 18 years old and older who visited the 3 main hospital emergency rooms (ER's) in Pachuca during the 4 months of October-November 1996 and June-July 1997. Also, from November 1, 1996, until December 12, 1996, a multiple-stage, stratified probability sample was drawn of all residents 18-65 years old living in Pachuca. The final sample consisted of 920 persons. This general population sample was weighted to take into account the selection process and the differential nonresponse rates across gender. Both ER patients and controls from the general population were interviewed with a questionnaire that asked about demographic characteristics and alcohol consumption within the 6 hours prior to an event that occasioned their going to the ER. Other questions on alcohol consumption included the quantity and frequency of usual drinking, the frequency of drunkenness, and alcohol dependence. Questions were also asked regarding any drug use and conduct problems during the last year, as well as depressed mood during the previous week. The study found that ER patients who reported drinking an alcoholic beverage within 6 hours of the event that brought them to the ER were more likely to have suffered a violence-related injury, and alcohol dependent patients were also more likely to have been treated for a violence-related injury compared to noncurrent drinkers. When both recent alcohol use and alcohol dependence were considered simultaneously in multiple models among current drinkers, patients with violence-related injuries were more likely to report recent alcohol use but not alcohol dependence. Depressive symptoms, but no conduct problems, were also associated with violence-related injuries in simultaneous regressions that included alcohol variables. An association between violence-related injury and depressed mood and conduct problems was found only at high levels of these symptoms. Among all ER cases and controls, those who consumed alcohol within a 6-hour period were 34 times more likely to have a violence-related injury during that period. This relationship held, albeit diminished, when alcohol dependence, age, sex, depressed mood, and conduct problems were simultaneously considered. 4 tables and 30 references