Ten years after: Examination of adolescent screening questions that predict future violence-related injury

Purpose: To determine which screening questions used in routine adolescent health care maintenance visits correlate with subsequent violence-related i njury. Methods: A prospective cohort study was undertaken of adolescents initially seen at the East Boston Neighborhood Health Center (EBNHC) in 1986. Risk f actor data were collected based on the adolescent health intake form in the medical records. The primary outcome measure, time until first violence-re lated injury was determined through identification on chart review of the t reatment of any such injuries at the urgent care center at EBNHC in the sub sequent 10 years, Kaplan-Meier survival statistics and Cox proportional haz ards models were used to account for loss of patients to follow-up. Results: Median follow-up for this sample was >5 five years. Male gender, c igarette smoking, alcohol use, other drug use, poor relationships with pare nts, not being in school or failing school, and history of fighting in the past year, predicted violence-related injury within the follow-up period. T he number of fights in the past year appeared to have a dose-response effec t on risk of subsequent violence-related injury. A simple screening instrum ent consisting of items concerning school status, drug use, and fighting hi story was used to stratify youth into low, moderate, and high risk of viole nce-related injury during the follow-up period. Conclusions: These results suggest that a simple three-item screening instr ument may be used to stratify the risk of future injury at the time of adol escent health maintenance visits. Further research is indicated to validate this finding in other populations. Interventions designed to assist adoles cents who are not in school or who have drug use problems should also incor porate violence prevention strategies. (C) Society for Adolescent Medicine, 1999.