OBJECTIVES: The association between suicidality and antidepressant use has been reported in literature using clinical trial data. This study 1) investigates the relationship between suicidality and antidepressants use in children and adolescents using managed care claims data; 2) examines the effect of an FDA black box warning on antidepressant use in children and adolescents. METHODS: Subjects 5-19 years old, continuously enrolled in a large Midwestern IPA model health plan, diagnosed with depression (using ICD-9-CM codes) or issued at least one antidepressant prescription between January 1, 2004 and June 30, 2005 were included in the study. Descriptive statistics, paired samples t-test, Chi-Square test, and stepwise logistic regression were used to analyze the data. RESULTS: A total of 5104 subjects (7.2% continuously enrolled in 5-19 year olds) met the study criteria; 53% female; mean age was 15 years; 67% had depression diagnosis. Antidepressant use was modestly associated with suicidality using Chi-square test (×2=3.03, p=0.08; contingency odds ratio=1.67). Logistic regression showed that antidepressant use is not a significant predictor to suicidality, while gender, depression diagnosis, other comorbid mental health, drug and alcohol abuse, and specialty provider were significant predictors to suicidality. While average days on antidepressant drug therapy decreased significantly (p=0.009) after the black box warning, the number of inpatient stays for depression after the black box warning increased significantly (p=0.007). CONCLUSION: From this claims database, the association between suicidality and antidepressant use in children and adolescents is not clear. Caution is needed in the use of antidepressants in this population. Increased inpatient stays following the black box warning may be an indication that providers are more aware of the suicidality risk or due to decreased use of antidepressants. It is important to balance appropriate usage and concerns for the safety of medications with the risk of adverse events and subsequent health service use.