PreoP-SSI: Prediction and prevention of pediatric surgical site infections - PROJECT SUMMARY Despite evidence that half of all surgical site infections (SSIs) may be preventable, SSIs continue to increase in the United States and are a substantial cause of morbidity, mortality, and healthcare costs. There is a lack of evidence-based guidelines for pediatric SSI prevention. Previous efforts to identify pediatric risk factors to inform actionable recommendations have been limited by small sample sizes and data availability. There is an urgent need to provide clinicians with evidence-based, individualized SSI risk and prevention recommendations to optimize patient care, reduce infection risk, and improve shared decision making and informed consent for children and families undergoing surgery. The objective of this proposal is to harness the power of machine learning to generate SSI risk prediction models using electronic health record (EHR) data to inform pediatric SSI care and the design of an EHR-based clinical decision support tool. This study will leverage a large national pediatric surgical dataset to train, validate, and test statistical and machine learning algorithms that will then be applied to an external test set from Stanford Medicine Children’s Health to evaluate performance and applicability for real-world clinical care (Aim 1). The investigators will then apply human-centered design to create and test the usability of an EHR-embedded clinical decision support tool prototype that provides clinicians with real-time, evidence-based SSI risk estimations and prevention guidance (Aim 2). The long-term goal of this project is to produce a clinical decision support tool that will be ready for prospective testing to augment real-time SSI prevention decision making to help clinicians care for surgical patients with higher reliability using evidence-based, patient-specific actions. This research will support NICHD’s focus on disease prevention and health promotion efforts through improving early detection of children at risk for infection, optimizing timing of prevention efforts, and ultimately preventing adverse health outcomes from SSI. The methods employed in this study will also advance NICHD’s aspirational goals to leverage machine learning and artificial intelligence for precision medicine. The proposed training, guided by an expert mentorship team, will enrich the applicant’s skills in machine learning and prediction, translational data science for precision health, and clinical informatics and technology implementation. The applicant will benefit from interdisciplinary expertise, directed mentorship, and coursework from both the University of California, San Francisco and Stanford Medicine Children’s Health, two world-class research and clinical environments. This research and training will prepare the applicant for a future career as an independent researcher focused on optimizing pediatric health through evidence-based EHR tools with real-word impact on patient care and outcomes.
