To the Editor: We read with interest Rey’s interpretation of the Multimodal Treatment Study of Children with Attention Deficit Hyperactivity Disorder (MTA).1 The MTA was a large randomised study comparing the impact of stimulant medication, behavioural treatment, a combination of the two, and standard community care on attention deficit hyperactivity disorder (ADHD).2 The treatment phase lasted 14 months, during which the children taking medication showed more improvement than the other groups. Participants were then allowed to change their treatment and, at 36-month follow-up, the outcomes in all groups were similar. Rey concluded that, if stimulant medication is not associated with sustained improvement, its place in the treatment of ADHD is limited.

This conclusion overlooks two important points. First, the greater initial improvement in symptoms of ADHD associated with stimulant medication might be important both clinically and socially. The second point is the expected impact of a relatively brief intervention: is it really plausible that an independent effect of 14 months of controlled treatment will be detectable after a further 22 months of self-selected management? The observation that the 14-month treatment phase becomes progressively less relevant as time passes is perhaps not altogether unexpected.

In the unmedicated group, the 23% non-compliance rate during the 14-month treatment phase indicated greater dissatisfaction with treatment than the 10% non-compliance in the medication groups (P < 0.005, χ2 test). This could imply a parental preference for more immediate relief of symptoms, even if it involves their child taking medication. Parental preference can be accommodated if the family and treating physician discuss and agree on a treatment plan, adjusted to optimise functioning. Far from indicating a diminished role for medication, the evidence from the MTA study suggests that the clinical approach would involve most individuals with ADHD being treated with stimulant medication at some stage.

Stimulant medication treats symptoms; it is not curative. It is likely that the role of stimulant medication in the treatment of ADHD decreases as children mature. However, temporary relief of symptoms can be highly valuable for affected children and their families.

In reply: Poulton and Nanan question my statement that the role of psychostimulant medication in attention deficit hyperactivity disorder (ADHD) becomes less prominent when the 3-year results of the Multimodal Treatment Study of Children with Attention Deficit Hyperactivity Disorder (MTA) are taken into account.1 Studies such as the MTA that report dramatic short- to medium-term improvement have, in my experience, increased practitioners’ expectations and reliance on these medications. Their clinical use has gradually widened to preschool-aged children and to the, so far, poorly validated inattentive and impulsive–hyperactive subtypes of ADHD. I observe this increasing the pressure on parents — not necessarily from clinicians — to use stimulants through an emphasis on the consequences of non-treatment, such as underachievement and conduct problems.

The 3-year follow-up of the MTA brings the early findings into perspective: a carefully titrated medication regimen produces no better results 3 years later than behavioural treatment and standard community care.2 In that sense, the role of stimulants versus other interventions has shrunk, and conscientious practitioners will inform parents and children of these findings when examining treatment options. My editorial did not query the many short-term benefits of stimulants but raised questions about when, and for how long, they should be used.

Poulton and Nanan rightly emphasise that stimulants are a “symptomatic” treatment. Further, stimulants increase the ability to concentrate and be on task whether or not individuals meet criteria for ADHD.3 This is further compounded because ADHD, like intellectual disability in the case of intelligence, represents the extreme of a dimension of behaviour.4 The boundary between illness and non-illness depends on where you draw the line, not on qualitative differences. However, there is no good tool to measure ADHD, unlike intelligence, and asessment depends on clinicians’ thoroughness and skill, and on informants, who may or may not be reliable.

The situation with ADHD is also similar to that for nocturnal enuresis, another disorder that lessens with increasing age, although it may persist. While behavioural treatment (the bell and pad alarm) is effective for enuresis, families and clinicians prefer using medication, even though the latter is “symptomatic” treatment and potentially harmful.5 This may be an alternative explanation for the higher non-compliance rate in the non-medicated group: behavioural treatments place more demands on parents, children and schools than a pill.