Perspective on the paper by Himmelmann et al (see page 921)

Descriptions of the epidemiology of a disease require a specific definition of what comprises the disease. This assumes particular relevance in relation to cerebral palsy (CP) because it is not a single disease entity and it has a multifactorial aetiology, with considerable variation in presenting symptoms and signs. Early descriptions of CP by Little, Freud and Ostler describe it as a collection of motor disorders dating from the time of birth or early childhood and caused by lesions of the brain.1 Since then there have been several attempts to define the syndrome.2 3 4 5 Not only has a universally acceptable definition of what comprises the complexity of CP proved elusive, but its classification into a several subtypes has also presented problems. Classifications have been made based on the anatomical site of the brain impairment; the clinical symptoms and signs, eg, spastic conditions, dyskinesias or ataxias; the topographical involvement of the extremities, eg, hemiplegia, diplegia, quadriplegia; the degree of muscle tone, eg, isotonic, hypotonic or hypertonic; and the timing of the presumed insult, eg, prenatal, perinatal or postnatal.6 To these problems of definition and classification are added the variation over time of symptoms and signs in a child and the variation between assessing clinicians and interpreting these symptoms and signs. An algorithim applying reliable standardised neurological …