SEMINAR that have resulted in different interpretations of a core deficit, but in the context of a general theory of attention these differences maybe primarily semantic. 11 The ratio of boys to girls with ADHD/HKD is between 3: 1 and 9: 1 but this may decrease with age. Part of the difference between sexes may be referral bias related to symptoms of disruptive behaviour since boys have more hyperactive/impulsive symptoms and more conduct and oppositional symptoms than girls. Referral biases in clinical studies are also related to the professional expertise of the therapist and this bias may account for the wide variation in the percentage of ADHD and HKD cases with other common disorders of childhood deemed to have separate aetiologies (eg, learning disabilities and tics). 12 Follow-up studies of children diagnosed with ADHD or HKD have shown that although hyperactive/impulsive symptoms decrease with age the symptoms of inattention do not, whereas symptoms of other disorders, such as conduct and anxiety disorders, increase with age. 13 About one-third of individuals with ADHD diagnosed in childhood still meet the criteria in adulthood, and a diagnosis of antisocial personality disorder emerges in about 20% of those diagnosed as children. After decades of different operational definitions, DSM and ICD manuals in their most recent versions2, 3 now recognise the same problem behaviours as the basis of the diagnosis, in almost identical sets of 18 symptoms (panel 1). However, there are still three major differences in the decision rules.

In the symptom domain groups (inattention, hyperactivity, and impulsivity) an ICD-10 diagnosis of HKD needs some symptoms in all three groups whereas DSM-IV (ADHD) does not (panel 2), but instead specifies partial subtypes if symptoms are from only one domain. ICD-10 makes special provision for a combined diagnosis category if a conduct disorder is present and, because of the high frequency of this combination, uses the presence or absence of a conduct disorder as the basis for the main subdivision of HKD. DSM-IV does not make any special provision for conduct disorder as a comorbid condition but allows its diagnosis as it does other psychiatric disorders. Another difference between the classifications is the use of other comorbid conditions as exclusion criteria (panel 3). ICD-10 aims at a single diagnosis and does not recommend the HKD diagnosis in the presence of internalising disorders such as anxiety and depression. DSM-IV aims to recognise as many diagnoses as there are symptom patterns. There are other minor differences, such as in the criteria for crosssituational pervasiveness, which are more rigorous in ICD-10 than in DSM-IV. The overall result of these differences in decision rules is that HKD is a subset of ADHD in ICD-10 and can be used to identify a refined phenotype. There are sensible recommendations about making the diagnosis in clinical practice. 4, 5, 8 Clinicians should elicit the history of specific symptoms from those who know the child best—usually the parents and teachers. Rating scales, with the specific ADHD/HKD symptoms, have been developed and provide a systematic approach for