US Patent 9633171 Predicting neonatal hyperbilirubinemia

A multi-variable statistical predictive leading-indicator approach is employed for identifying newborns at risk of clinically significant hyperbilirubinemia and for determining to administer interventions to at-risk newborns. In embodiments, a multi-variable logistic regression statistical model capable of calculating a probability of clinically significant hyperbilirubinemia is generated. Using an input data set for a newborn and the multi-variable logistic regression statistical model, a probability of clinically significant hyperbilirubinemia is determined for the newborn and presented to a clinician.