ABSTRACT / PROJECT SUMMARY The long-term goal of this K23 Career Development Award is to prepare Dr. Son H. McLaren, MD, MS for an independent research career focused on improving the health and outcomes of infants and children presenting to acute care settings with bronchiolitis and other potentially severe infectious illnesses through the use of predictive modeling, integrating continuously measured (time series) physiologic or vital sign data and machine learning techniques. Bronchiolitis, a viral lower respiratory tract infection, is the most common reason for emergency department (ED) visits and hospitalizations in infants ≤90 days old, hereafter referred to as young infants. In an ongoing prospective cohort study of young infants with bronchiolitis, Dr. McLaren found that young infants are more likely to have potentially avoidable hospitalizations, when compared with broader age of infants <2 years old. Over-hospitalization contributes to preventable exposure to nosocomial infections and medical errors, undue caregiver psychosocial distress, and avoidable health care costs. As no clinical model has been developed to predict the risk of respiratory progression in young infants, many ED clinicians struggle with the decision to discharge infants to home even when they are well appearing. An accurate, reliable prediction model that identifies infants at low risk of respiratory progression using clinically practical predictors would facilitate safe ED discharge. Prior literature suggests that viral etiology and physiologic time series data may augment prediction model accuracy, but little work has been done specifically in infants with bronchiolitis. To address these gaps in knowledge, Dr. McLaren will pursue the following specific aims: to derive and internally validate clinical prediction models using a) clinical history, physical examination findings, and viral etiology (Aim 1) and b) physiologic time series data from bedside monitors (Aim 2), to identify young infants with bronchiolitis at low risk of developing respiratory progression. Completion of these aims will create novel prediction models incorporating comprehensive clinical, viral, and physiologic data that can inform clinical decisions at the bedside and enable safe and objective disposition decisions. Dr. McLaren has formed a strong team of mentors and advisors with complementary expertise in bronchiolitis epidemiology, clinical prediction modelling, time series data analysis, and machine learning techniques to successfully conduct the proposed study. With their dedicated mentorship, didactics, and rich training environment of Columbia University, Dr. McLaren will achieve the following training goals: a) learn advanced research methods in clinical prediction modeling, b) become proficient in analysis and model building using physiologic time series data, and c) gain mentored experience in federally funded, multicenter emergency medicine research networks. The completion of this K2...