Abstract
OBJECTIVES: (i) To estimate the rate ratio (RR) of use of alcohol alone, cocaine alone, and both substances simultaneously on acute injury or an aggressive incident, (ii) to compare the RRs for simultaneous use within 3 or 6h of the event; and (iii) to compare the RRs of two measures of exposure, "hours of feeling effects" vs estimates based on self-reported quantity and frequency of use. METHODS: The study employed a case-crossover design with the frequency approach. Clients (N=616) in substance abuse treatment for alcohol or cocaine issues from 2009 to 2012 completed a self-administered questionnaire on their substance use within 3 and 6h before a recent injury or physically aggressive incident. Clients also reported detailed quantity and frequency information in relation to their typical substance use, as well as information on "feeling effects". The RR of acute harms due to substance use was estimated using the Mantel-Haenszel estimator. RESULTS: In the 6-h window before the event, use of cocaine alone, alcohol alone and simultaneous alcohol and cocaine use were each significantly (P<0.05) related to a recent injury and aggressive incident. Simultaneous use was not significantly greater than use of either drug alone. Estimates of RR based on simultaneous use for a 3-h window before the event were consistently larger than those based on a 6-h window, and comparisons were significant (P<0.05) for an aggressive incident but not an injury. With reference to the two measures of exposure, three of eight comparisons of RRs were significantly larger for feeling the effects of the substance in comparison to quantity and frequency of substance use. CONCLUSION: These findings are consistent with increased likelihood of harms related to the acute effects of alcohol alone, cocaine alone or simultaneous use. The results are suggestive that the acute effects of these drugs may be better measured within a 3-h time window than a 6-h window. Finally, we found that "hours of feeling effects" yielded higher estimates of RR than the quantity-frequency approach; however both measures support the overall findings.