Review Article © 2006 Blackwell Publishing Ltd 521 Blackwell Publishing LtdOxford, UKCCHChild: Care, Health and Development0305-1862© 2006 The Authors; Journal compilation © 2006 Blackwell Pub- lishing Ltd2006325521533Review ArticleChildren with autistic spectrum disordersH. Cass et al. Correspondence: Hilary Cass, Great Ormond Street Children for Hospital NHS Trust, London WC1N 3JH, UK E-mail: cassh@gosh.nhs.uk Original Article Medical investigation of children with autistic spectrum disorders H. Cass,* D. Sekaran† and G. Baird‡ *Neurosciences Unit, Great Ormond Street Children’s Hospital NHS Trust †Cheyne Child Development Centre, Chelsea and Westminster Hospital NHS Trust, and ‡Newcomen Centre, Guy’s and St Thomas’ NHS Trust, London, UK Accepted for publication 1 February 2006 Abstract Few well-constructed studies have systematically evaluated medical investigation protocols for children with autistic spectrum disorders. This is in large part due to the heterogeneous nature of the population and changing diagnostic frameworks. This review outlines the studies that have directed investigation strategies to date, and discusses how these might be applied in the clinical situation. The importance of listing the conditions that may be present on the basis of the specific clinical presentation, and using a thorough history and examination to generate a pre-test probability of the target disorders is emphasized if tests are to be useful in directing therapy or broader management approaches. Keywords autistic spectrum, disorders, investigation protocols, learning disability Background Over the last decade, the practice of medicine has come under increasing scrutiny as patients have rightly come to expect more consistent and better- evidenced diagnostic assessment, investigation and treatment. Clinical guidelines have become an important mechanism for trying to improve standards and to ensure equitable usage of limited resources (Shekelle et al. 1999; Woolf et al. 1999), but such guidelines are only as robust as the evidence on which they are based. Woolf and col- leagues draw attention to the fact that only a small subset of what is done in medicine has been tested in appropriate well-designed studies. Frequently the evidence base remains poor and clinical deci- sions continue to be based on heterogeneous case series and expert opinion. Given incomplete and conflicting evidence, the opinion of expert groups will inevitably be biased by experience and value judgements. Applicability to subgroups and to individual patients may well be limited. Set against this background, the increase in the number of children diagnosed as having autistic spectrum disorders (ASDs) (Goodman 2005) and the rising public and media interest in the aetiology of the condition necessitate a rational approach to medical investigation of this group. In response to this need, a number of expert groups have devel- oped practice parameters for diagnosis and assess- ment, which incorporate recommendations for investigation; for example, the Child Neurology Society and American Academy of Neurology (Filipek et al. 1999), and the National Initiative for Autism: Screening and Assessment (LeCouteur & Baird 2003), summarized in Baird and colleagues (2003). Discussion of the evidence base underlying the investigative protocols is necessarily limited within these broader reports. This paper discusses the rationale for the development of these recom- mendations and some of the key considerations in translating guidelines into practice, and should be read in conjunction with the aforementioned prac- tice parameters.