A GP currently has no reliable way to tell whether a feverish toddler has a mild virus or a life-threatening bacterial infection. This project will test whether an existing clinical tool—or a newly developed one—can accurately rule out serious illness in children under five in UK general practice, where serious infections are rare but deadly if missed. Most existing tools were designed for hospital emergency departments, where sick children are more common, and they perform poorly in GP surgeries. The team will validate a decision tree developed in Belgian general practice using a UK dataset of 6,703 acutely ill children, then attempt to build a better tool that relies on objective measures rather than a clinician’s “gut feeling.” If successful, the research could give GPs a simple, evidence-based checklist to confidently send well children home and fast-track the few who need hospital care. That would reduce unnecessary referrals, ease pressure on emergency departments, and—most importantly—prevent missed diagnoses that can lead to sepsis or death. If no existing tool proves accurate enough, the work will identify exactly what data is needed to build one from scratch.
View original technical description
Development work question Can an accurate clinical tool be identified or developed to rule out serious infection in acutely ill children under five years old presenting in UK General Practice? Background Most children with acute illness can be safely managed at home, but a few are seriously ill needing hospital admission. Identifying these children is difficult in General Practice. Most clinical tools have been developed in Emergency Departments (ED) or secondary care. However, these tools perform less well in General Practice, where the prevalence of serious illness is lower. We recently validated the Traffic Light System, published in National guidance (NICE) and National Paediatric Early Warning System (PEWS; recently adopted in hospitals in England) and found that both performed poorly in General Practice. We urgently need to find a clinical tool to aid General Practitioners (GPs). A rapid review of the literature found only one clinical tool developed in General Practice (Van den Bruel s Decision Tree). This has not been validated in children less than five years old, nor in a UK General Practice population. Aim To identify or develop a clinical tool to rule out serious illness in acutely ill children under five years old presenting in UK General Practice. Objectives 1. Validate Van den Bruel s Decision Tree using existing GP dataset 2. Develop an alternative clinical tool using existing dataset 3. Establish a patient advisory group of parent/carers of young children to inform future programme development. Development work plan Work package (WP) 1 will test the accuracy of Van den Bruel s Decision Tree using our existing GP dataset of 6,703 acutely ill children <5 years old, with presenting symptoms and examination findings linked to routine hospital admission data. WP2 will use this existing dataset to develop a new tool; explore whether a tool using only objective measures can be developed (Van den Bruel s includes subjective clinician gut feeling ); and explore whether including urinalysis strengthens tools. WP3 will establish a patient advisory group to gather views on the development work and inform the future programme work and dissemination. Timeline • Project start 1st September 2025 • WP1 completed April 2026; Paper 1 July 2026 • WP2 completed February 2026; Paper 2 May 2027 • WP3 set up by December 2025; meetings throughout project. • Report completed and project end May 2027 • Programme Grant application July 2027 Anticipated impact and dissemination After the development work, we will know how an existing clinical tool performs in this population and will also have developed an alternative clinical tool for use in General Practice. The subsequent programme of research will include external validation of the new tool, assess acceptability of tools, and evaluate the implementation of the most accurate and acceptable tool in clinical practice. If no tool is sufficiently accurate, the programme will collect the data needed to develop an accurate tool. Results will be shared with our patient and clinical advisory groups; two papers will be published; results presented at national and international conferences; NICE informed for updates in guidance.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know