8-Year Follow-up of the MTA Sample

Abstract

A Dance to the Music of Time is the collective title given to a 12-volume series of novels by English author Anthony Powell that spans the period 1914Y1971 and involves more than 300 characters. The work is notable for the way Powell advances the narrative while, at the same time, developing further background to the story through reminiscences of the narrator and conversations between the characters. Reporting of the NIMH Collaborative Multisite Multimodal Treatment Study of Children With Attention-Deficit/Hyperactivity Disorder (MTA) is unfolding in a similar fashion. The MTA started with 579 children aged between 7 and 9.9 years who were randomized to receive one of 4 treatment conditions for 14 months. At the completion of the randomized phase of the trial, participating children and their families were free to resume or initiate treatment under the supervision of community clinicians. Their progress has been followed by the MTA study team. Eight primary articles have advanced the narrative by reporting the main findings of the study at successive waves of follow-up, whereas supplementary articles have developed the background to these main findings. Overall, more than 70 articles have been published on the MTA data set to date, creating a challenge for those wishing to keep the myriad of findings from this study in focus. Readers are directed to two recent articles that seek to review and integrate the MTA literature. This issue of the Journal includes a primary article reporting key outcomes at 6 and 8 years, and a supplementary article on the level of agreement between parental accounts of their child’s adherence to medication and objective assessment of adherence through salivary assay for methylphenidate. Less than 2 years have passed since the publication of the MTA 3-year follow-up data. This was the first in the sequence of primary articles to show that differences in key outcomes attributable to treatment group assignment during the first 14 months of the trial had vanished. The authors were careful to point out that the study had not been designed to demonstrate benefits of the randomized treatments beyond 14 months. After all, in the intervening 22 months, under the supervision of community clinicians, the treatment received by the participants available to follow-up had grown to look not identical but similar. Gradual extinction of the effect of assigned treatment as evaluated by intent-to-treat analyses was predicted after the 24-month follow-up. Nevertheless, the authors of the article in this issue of the Journal tested the hypothesis that there may be a sleeper effect, with the benefits of assigned treatment emerging again later in development. A betting person would give short odds against this being the case, and they are correct. There were no differences between the four assigned treatment groups after 6 and 8 years on repeated measures of psychiatric symptoms, academic function, and social functioning. Nor were there differences on new measures salient to adolescence such as grade point average, arrest by the police, or psychiatric hospitalization. The minority of participants who continued with medication at 8 years was at no clear advantage over those who did not, but as the study was no longer controlled, the finding should be interpreted with caution. Fresh attempts to elucidate the mechanism underlying the convergence in outcome of the four assigned treatment groups are unwarranted, given the effort already directed to the interpretation of the convergence at 3-year follow-up. We accept that the absence of a sleeper effect is a reflection of the reality and not the consequence of bias in the study. The authors are correct in their statement that it is purely speculative whether persistence of intensive treatment beyond 14 months would have led to sustained differences between the assigned treatments. The article by Molina et al. reports two secondary analyses. The first involved the grouping of participants not by assigned treatment but according to trends in attention-deficit/hyperactivity disorder (ADHD) symptoms from baseline through 14, 24, and 36 months, as described by Swanson et al. Evaluable data were available for 485 participants. On first pass, it may seem that an initial large improvement in ADHD symptoms that then plateaus over Accepted January 27, 2009. Dr. Hazell is with the Discipline of Psychological Medicine, Concord Clinical School, University of Sydney. Correspondence to Philip Hazell, Ph.D., F.R.A.N.Z.C.P., Thomas Walker Hospital (Rivendell), Hospital Rd, Concord West, NSW 2138, Australia; e-mail: philip.hazell@sswahs.nsw.gov.au. 0890-8567/09/4805-0461 2009 by the American Academy of Child and Adolescent Psychiatry DOI: 10.1097.CHI.0b013e31819c242f