• diarrhoea
• evidence based guideline

We would like to respond to the commentary¹ on our paper on the above.² The following points require clarification.

Dr Baumer states that “Dr Armon and colleagues have used a formal consensus process to provide guidance ...” The recommendations provided were evidence based wherever evidence was available, with a formal consensus process used additionally. A systematic review of the literature was undertaken. We complied with the essential elements of a systematic review in accordance with the Scottish Intercollegiate Guidelines network. Thus the literature must be identified according to an explicit search strategy; selected according to defined inclusion and exclusion criteria; and evaluated against consistent methodological standards.³ As is stated in the text, mesh heading and text word searching was performed.

Formal consensus development was used:

• To develop recommendations where evidence was not available. We would argue that in most guideline development processes this is the prime area of difficulty because in paediatrics, research is sparse and consensus is required for a guideline to be complete. The transparency of the development process is essential and we were explicit about which recommendations were based on consensus rather than on evidence. The Delphi method imparts greater rigour to a process that is often performed by a small selected group of individuals.

• To act as an internal peer review. The guideline is available to a large number of participants during the development process so that any inaccuracies can be identified early on. The participants have an opportunity to check the literature grading and to inform the development group of any papers not yet identified.

• To improve the implementation of the guideline, since the Delphi process facilitates the translation of recommendations into an algorithm which can be readily followed by clinicians. This is an important consideration in all guideline development.⁴

We stated that we included scientific reviews of the literature and guidelines written by national bodies in our inclusion criteria. The two systematic reviews references in Dr Baumer's commentary were appraised and graded as level 1 evidence since they both included sufficient detail of the methods to suggest that they were based on thorough systematic reviews of the literature.^5,6 Despite sound methodology in performing a systematic review, the evidence from the articles found still has to be translated into recommendations for putting the evidence into practice. Formal consensus was therefore used to determine how a large body of clinicians commonly looking after children with this presenting complaint would apply the evidence found.

It is important to differentiate guideline development from a Cochrane review. Established organisations who have been developing guidelines such as SIGN do not advocate hand searching and contacting of experts for unpublished studies in every circumstance.³ This decision is left to the guideline developer and depends on the subject area. In view of the scope of this review (to produce a guideline for the child presenting to hospital with acute diarrhoea) it was impractical to perform hand searches and contact experts in the field for all the clinical questions that were addressed. At the time of the development of the guideline the RCPCH standards on guideline production did not stipulate hand searching or contacting of experts.⁷

We accept that with only two nurses on the panel, nursing staff views would not be fully represented. Unfortunately few of the nurses approached were prepared to take part.

We would agree with Dr Baumer that primary care input is valuable and would be a useful extension of the work. At the outset, however, we limited the scope of the guideline to children presenting to hospital, in the light of previous research showing that 16% of medical presentations to hospital were with diarrhoea.⁸

Parents views on admission were not sought during the Delphi process. However, we have incorporated the importance of the parents views in the guideline itself (boxes 28, 38, 43). Again parental views could be formally sought as an extension to the work.

Finally, the implementation of any national guideline at local level requires modification by those who will use it and this we fully endorse.