A Personalised Approach to breathing support Treatments in sick CHildren (PATCH)

Background Paediatric critical care is a high-impact, high-cost specialty with a limited evidence base. This is particularly evident in ventilation, where a severe dearth of scientific evidence to guide practice compromises patient outcomes, and results in unwarranted care variations and NHS waste. To accelerate evidence generation in paediatric critical care, I am currently co-leading a national Bayesian adaptive trial platform (PIVOTAL) in which multiple treatments (sedation, fluids and blood transfusion) are being simultaneously evaluated. Paediatric ventilation research is, however, hampered by heterogeneity in children's disease biology, respiratory physiology, and treatment response. During this Professorship, I will address this challenge by laying the groundwork for future precision medicine ventilation trials within PIVOTAL. Aims & Objectives My aim is to improve children's outcomes by developing personalised ventilation approaches and creating a PIVOTAL ventilation domain to evaluate them. I will tackle biological, physiological and treatment response heterogeneity in PATCH by: Validating emerging paediatric acute respiratory distress syndrome (pARDS) subphenotypes in ventilated children with acute respiratory failure and assessing accuracy of point-of-care tests for rapid subphenotyping Using data from serial point-of-care lung ultrasound scans integrated with clinical parameters to develop decision trees to guide personalised ventilation management Adapting and validating adult sepsis reinforcement learning (RL) models to optimise ventilator weaning and extubation I will also: Establish a PIVOTAL ventilation domain to compare personalised approaches from 1), 2), and 3) with usual care in future clinical trials. Work packages (WP) WP1 (m0-m48): Prospective cohort study nested within PIVOTAL fluid RCT (n=500) to validate plasma protein-based pARDS subphenotypes (WP1a) and assess the accuracy of point-of-care assays (n=250) to rapidly assign subphenotypes (WP1b). WP2 (m0-m42): In the WP1 cohort (n=500), analysis of integrated clinical and serial bedside lung ultrasound scan data using multilevel regression models and machine learning (ML) approaches to identify predictors of intervention response (WP2a) and development of decision trees (WP2b). WP3 (m0-m48): Using an existing granular electronic health record dataset (training set: ~12,000 children from 3 PICUs; test set: ~8000 children from 3 other PICUs), development and validation of extubation failure prediction models using supervised ML methods (WP3a) and extending adult sepsis RL models for paediatric weaning and extubation (WP3b). WP4 (m25-60): Design and setup of PIVOTAL ventilation domain to efficiently evaluate personalised ventilation approaches from WP1-3 in future precision medicine trials. PPI PATCH has been co-designed with parent groups. Two PPI members will sit on the research steering group. Impact and dissemination PATCH will transform ventilation care and improve outcomes for our sickest children. The Professorship, and associated institutional support and research mentorship, offers a unique opportunity to pump-prime an under-researched area and for me to rapidly grow into an international leader. The ventilation domain will be an enduring legacy, nurturing the next generation of research leaders. My clinical role, and leadership of the national PICU research network and NIHR-supported Incubator, provide direct line of sight to research translation and capacity building. I will leverage scientific (conferences, publications) and lay channels (social media, patient groups, public engagement events) to widely disseminate research findings.