Objective: To identify factors associated with treatment initiation of atomoxetine (ATX), stimulants (STIM), or long-acting stimulants (LA-STIM) in children with ADHD using Medicaid. Methods: Data were from the IMS Health LRx Database. Patients covered by Medicaid age <18 years old were selected if they initiated treatment with an ADHD medication categorized as ATX, any STIM, or LA-STIM between January 2005 and December 2005. Initiation was defined as the first use of a medication preceded by 120 days without a prescription in the same category. Stepwise logistic regression was used to identify the factors associated with initiations of ATX vs. STIM or ATX vs. LA-STIM adjusting for age (0-5 vs. 6-12, 13-17 vs. 6-12), gender, prior ADHD medications, other concomitant psychiatric medications, provider specialty, and a number of unique of previous ADHD medications used. Results: A total of 24,141 patients (68.62% male) most recently initiated treatment with ATX, 144,451(68.84% male) with STIM, and 129,323 (68.82% male) with LA-STIM. Increasing age was associated with increased likelihood of ATX initiation (p<.05) relative to STIM. Other significant factors (all p-value<0.001) were initiation concomitant with use of antidepressants (OR=1.29), antimanics (OR=1.34), antipsychotics (OR=1.49), anxiolytics (OR=1.69), or anticonvulsants (OR=1.33). ATX initiation also became more likely with increasing number of historical ADHD treatment episodes. Pediatricians were more likely to prescribe STIM over ATX relative to general psychiatrists (OR=0.76). Factors for initiation of ATX vs. LA-STIM were similar to those for the STIM comparisons. Conclusion: The factors associated with initiation of ATX vs. STIM or ATX vs. LA-STIM suggest that despite common indications, atomoxetine and stimulants may be addressing different needs. The findings suggest that ATX is preferentially prescribed for patients with psychiatric comorbidities or potentially complicated treatment profiles.