3-year follow-up of the NIMH MTA Study.


The 1999 Multimodal Treatment Study of Children With ADHD (MTA) was a randomized clinical trial that compared four ADHD treatments for children ages 7-9 (n=579): medication management (MedMgt), behavior therapy (Beh), their combination (Comb), and regular community care (CC). After 14 months of treatment, MedMgt and Comb showed significantly greater improvements in ADHD and ODD symptoms than Beh or CC. An initial follow-up at 24 months (10 months after treatment) continued to show superiority of MedMgt and Comb, but with reduced effect size. This is possibly because at 14 months, the study became observational, and families could choose their own treatment. Those initially assigned to take medication were more likely to stop after 14 months, and those assigned behavior therapy were more likely to start taking medication after 14 months. This paper reports 36-month outcomes (22 months after treatment ended), and shows no significant difference in outcomes between the four groups. While not significantly different from each other, all four groups showed significant improvement from baseline. Because there was no untreated control group, this study does not address how much of this improvement was due to treatment.


  • This is an observational follow-up study to the MTA Study. MTA assigned randomized treatment for 14 months. From 14 to 36 months the study was observational. Families were free to choose their own treatment, but barriers to care differed by community.
  • Of the original 579 subjects, 485 (83.8%) participated in the 36-month follow-up.
  • Five clinically relevant measures used. Three were rated by parents and teachers: ADHD symptoms, ODD symptoms, and social skills. Remaining measures were the reading score (Wechsler Individual Achievement Test), and overall functional impairment (Columbia Impairment Scale).


  • At 36 months (22 months after discontinuing randomized treatment), all treatment groups showed significant improvement in outcome measures from baseline (ADHD/ODD symptoms, reading, social skills, functional impairment), however groups did not show significant differences from each other. This is in contrast to the clear superiority of groups receiving medication (MedMgt and Comb) at the 14-month mark. This advantage was diminished at 24 months, and gone by 36 months.
  • After 14 months, families were free to choose their own treatment options, and medication use between groups converged to a degree. Medication use in MedMgt and Comb dropped from an average of 91.5% of subjects at 14 months to 71.1% at 36 months. In contrast, medication use in the Beh group increased from 13.7% to 45.2%, and the CC group stayed about the same (around 60%).
  • Whereas medication use in the 14 months of randomized treatment was associated with significantly greater improvement in symptoms, medication use in the 24-36 month (non-randomized) time period was actually associated with increased symptoms relative to those not receiving medication. This is possibly due to selection bias, where children with worse symptoms (especially those in the unmedicated Beh group) started taking medication, and children with fewer symptoms stopped taking medication.


Jensen PS, Arnold LE, Swanson JM, Vitiello B, Abikoff HB, et al. (2007) 3-year follow-up of the NIMH MTA study. J Am Acad Child Adolesc Psychiatry 46: 989–1002

View on PubMed