Communications of the ACM
Machine Learning Saves Babies!
Stanford University researchers have developed a revolutionary, non-invasive way of electronically scoring and assessing a baby's well-being. Pictured in the neonatal intensive care unit at Lucile Packard Children's Hospital are (left to right) researcher
Credit: Lucile Packard Children's Hospital
Machine learning could help save the lives of premature babies, who have an increased risk of major health complications, including death within their first year of life.
Hospitals collect an enormous amount of data on the babies for evaluation, monitoring, treatment, and interventions, and the data is often never used again. However, computer scientists are using machine learning to find common patterns in the data, create new measures for predicting declines in health, and form new treatment plans.
For example, the PhysiScore system produces a probability score for each baby that represents the overall illness severity and likelihood of developing major complications. The score is similar to the Appearance, Pulse, Grimace, Activity, and Respiration score administered after a baby is born, but is much more accurate and can be automated on existing monitoring devices for incorporation within the clinical workflow.
Moreover, PhysiScore does not require blood draws, spinal taps, and other invasive procedures to measure data. The information can assist doctors with planning for infant transport and patient management.
From CS Bits & Bytes
View Full Article
Abstracts Copyright © 2012 Information Inc., Bethesda, Maryland, USA
No entries found