We are glad our article provoked some discussion and we agree with Dr Shah about the need to provide impartial advice and to determine an individual's preferences. Although the outcomes of the adult attention-deficit hyperactivity disorder (ADHD) service he describes are impressive, we do not know that these are attributable to medication alone, rather than other aspects of the care received in a specialist service. Only randomised controlled trials can establish whether medication has specific efficacy, after which effectiveness in real clinical practice and cost:benefit ratios have to be considered. Since we published our paper, the Medicines and Healthcare products Regulatory Agency (MHRA) has withheld approval for methylphenidate hydrochloride for adult ADHD on the basis that differences from placebo are small and do not outweigh documented adverse effects (http://news.wooeb.com/959215/adhd-drug-concerta-disapproved-for-adults-in-europe).

Dr Bhattacharya and Dr Lepping point out that ADHD is conceived as a dimensional rather than a categorical condition, but this does not change the arguments against it. The proposed trait is still defined by ‘symptoms’ that are universal experiences and diagnosis involves subjective judgements about impairment and what the impairment is caused by. The idea that the symptoms represent a unitary underlying condition that represents an evolution of a childhood disorder is simply an assumption, which is not currently supported by evidence.

Dr Bhattacharya accuses us of being one-sided and not being objective, but we would point out that no one is truly objective and everyone has their own perspective. We would suggest that we are being more objective than others by not starting from the presumption that adult ADHD is a useful and valid category. There are numerous articles that provide an opposing point of view. With regard to the randomised trials of drug treatment, since National Institute for Health and Clinical Excellence guidelines are so influential and have endorsed the validity and drug treatment of adult ADHD, it is important to point out the evidence on which these conclusions were reached. We also refer to a meta-analysis of trials of methylphenidate, which found no significant difference in parallel group randomised trials, and the Cochrane review of amphetamines quoted by Dr Bhattacharya also found a lack of evidence of long-term benefit and a high risk of bias.

We do not see that our presentation of the genetic data differs substantially from the way it is presented by Kooij et al, also quoted in Dr Bhattacharya's letter. In any case, we know that most molecular genetic findings are not replicated. The references we used to support the idea that there have been challenges to the concept of childhood ADHD include a book by one of us that was referred to because it reviews the literature in this area, and an article challenging the consensus statement on ADHD that was authored by 32 authors, as well as ourselves.

Dr Bhattacharya and Dr Lepping highlight the problem of comorbidity. The idea that the frequent comorbid conditions are distinct problems, or secondary to ADHD symptoms, rather than competing ways of conceptualising the same problems, is simply an assumption that follows from accepting the diagnosis of adult ADHD. Inattentiveness is too vague a concept to be used to clarify the diagnosis and, given the inclusive nature of inattentiveness ‘symptoms’, is likely to be identified by most people with mental health problems, as well as many without.

In response to Dr Lepping, studies on levels of the dopamine transporter in ADHD are contradictory, despite the consensus. ^{Reference Varrone and Halldin1} Stimulants have well-documented psychoactive effects, and so it is not surprising that they change behaviour in the short term, producing large effects sizes. What is at stake is whether or not they help people in the long term. Evidence in children is not convincing. The Mutimodal Treatment Study of Children with ADHD (the MTA study), which has been criticised on many grounds, found only marginal benefits of a ‘medication management’ package over behavioural therapy alone or routine community treatment (often including stimulants) at 14 months. ² At 3 years there was no difference between the groups, and there was no effect of compliance. ^{Reference Jensen, Arnold, Swanson, Vitiello, Abikoff and Greenhill3} At 8-year follow-up, analysis according to randomised group and actual medication used failed to show any advantage for medication. ^{Reference Molina, Hinshaw, Swanson, Arnold, Vitiello and Jensen4} Other naturalistic follow-up studies have also failed to demonstrate any advantage for long-term medication in children ⁵ and, as we describe in our paper, the evidence in adults is even weaker. Without evidence of long-term benefits, we suggest there is no justification for prescribing medication.