ommunity acquired pneumonia (CAP) can be deﬁned clinically as the presence of signs and symptoms of pneumonia in a previously healthy child due to an infection which has been acquired outside hospital. In developed countries this can be veriﬁed by the radiological ﬁnding of consolidation. In the developing world a more practical term—acute lower respiratoryinfection—is preferred, reﬂecting the difﬁculties in obtaining a chest radiograph. Ideally, the deﬁnition would include the isolation of a responsible organism. However, it is apparent from many studies that a pathogen is not identiﬁed in a signiﬁcant proportion of cases that otherwise meet the clinical deﬁnition (see Section 3 on Aetiology). As it is assumed that CAP is caused by infection, the presumptionis that current techniques have insufﬁcient sensitivity to detect all relevant pathogens. Treatment guidelines therefore have to assume that, where pathogens are isolated, they represent all likely pathogens. There is a clear need for better diagnostic methods. In creating guidelines it is necessary to assess all available evidence with consideration of the quality of that evidence. This we haveendeavoured to do. We have then produced key points and management guidelines based on the available evidence supplemented by consensus clinical opinion where no relevant evidence was found. A summary of the key points and a further summary prepared speciﬁcally for use in primary care are also available on the Thorax website (www.thoraxjnl.com) and the British Thoracic Society website(www.brit-thoracic.org.uk).
in paediatric infectious diseases, a specialist registrar in paediatrics, a paediatric nurse, a general practitioner, and a guidelines methodologist. No external funding was obtained to support the development of the guidelines. Because of the breadth of scope of the topic, the guideline development was divided up into 12 sections and members were allocated to each. Eight of the12 sections had at least two members allocated. Identification of evidence Search strategies were developed for each of the 12 sections (excluding guideline methodology) with the assistance of an information scientist. These search strategies (see Appendix 1) included MeSH and free text terms and had no language restrictions. They were run on Medline (Winspirs, Silverplatter) and the CochraneLibrary (Issue 3, 1999). Where searches yielded more than 1000 citations, these were limited to English. Assessing the literature The sets of references generated by researchers were sifted for relevance to the clinical topic of the guidelines. Where two or more members were working on a section this was done independently. Initial sifting was on the basis of the title and abstract (as obtained fromthe specialist resource). Where there was doubt about whether a reference was relevant, the full publication was obtained. Studies from countries where the populations or clinical practices were very different from the UK were excluded unless they addressed questions which could be generalised to the UK (such as clinical assessment). The methodological quality of the publication was assessed using achecklist adapted from one previously developed for this purpose.1 Synthesising the evidence Once individual papers were checked for methodological rigour and clinical relevance they were categorised according to study design.1 The evidence was synthesised by qualitative methods. The content of the identiﬁed papers was summarised into brief statements that were thought accurately to reﬂect therelevant evidence and the category of that evidence was indicated after each citation. The recommendations resulting from the evidence were graded according to the strength of that evidence (table 1). The strength of each recommendation ([A], [B], [C], or [D]) was indicated after each recommendation. Where
METHODS OF GUIDELINE DEVELOPMENT
Scope of guidelines These guidelines address the...