Background: Although nonfatal injury prevalence is higher among substance abusers than in the general population, few studies have estimated the injury risk for clinically recognized substance abusers. The extant studies, moreover, analyze rates of visits for injury treatment rather than rates of injury events. This study estimates the excess risk of medically treated and hospitalized nonoccupational injury for people under age 65 with medically identified substance abuse problems and private health care coverage.
Method: We conducted a retrospective cohort study by using medical claims data from Medstat Systems, Inc., with a longitudinal database of health care claims for 1.5 million people with health care coverage from 70 large corporations. Claims histories for anyone who had an alcohol-related or drug-related primary or secondary diagnosis during 1987 to 1989 were analyzed. A random sample was selected from the remaining people without a substance abuse diagnosis in their medical records. Injury rates were compared among groups. We used logistic regression to estimate odds of medically treated and hospitalized injury, controlling for demographics.
Results: Medically identified substance abusers had an elevated risk of injury in a 3-year period; alcohol and drug abusers had the highest risk (58%), followed by drug-only abusers (49%), alcohol-only abusers (46%), and controls (38%). Alcohol and drug abusers were almost four times as likely to be hospitalized for an injury in a 3-year period when compared with controls. Injury risks were elevated substantially more for female then male substance abusers.
Conclusions: This study greatly improves on available information about the risk of injury for drug and multiple-substance abusers. Medically identified substance abusers, especially adult women, have a higher probability of injury, more hospitalized injuries, and more injury episodes per person injured than nonabusers. More aggressive identification and subsequent treatment of female substance abusers appear warranted.