PediSEP1 · Risk stratified care to reduce antibiotic use and AMR transmission in African hospitals

In suv-Saharan Africa (sSA), most government-run hospitals aim to follow WHO or national syndrome-based guidelines. Clinical syndromes, such as ‘severe pneumonia’ are designed to be sensitive for pathologies like invasive bacterial disease that require specific treatments to prevent death. However, low specificity means that a far wider group of children is captured by a syndrome definition including many with a self-limiting infection and very low risk of death. Currently, decisions regarding admission, discharge, and escalating or de-escalating antimicrobials are typically made by staff with limited paediatric training, very often interns, and current guidelines have very limited or no advice on de-escalation and discharge. These factors contribute to antimicrobial over-use, overly broad-spectrum prescribing, and unnecessarily prolonged treatment ‘to be on the safe side’. Excessive duration of antibiotic treatment means a greater risk of acquiring AMR from the hospital environment, significant costs to health providers and families, overcrowding, a low nurse:patient ratio, and reduced care for higher risk children.We propose that major reductions in antibiotic use, exposure to the hospital environment and transmission of AMR can be achieved by risk stratified care enabling very low-risk children admitted to hospital to be cohorted away from longer-stay patients, stop antibiotics and go home earlier with inexpensive phone follow up. We will engage policymakers, undertake mathematical modelling of antimicrobial usage, AMR transmission and costs across sSA within risk strataand conduct a trial of risk stratified care, and develop a tool 'PPS-plus' for monitoring of AMR transmission through simple cross sectional surveys.Relevance to the work programme includes combatting the major global threat from AMR through identification of personalised treatment options, a better evidence-base for policymaking, and digital tools to optimise clinical workflows.