‘Machine learning’ a boon for insurers, but can’t replace human touch in healthcare

The concept of “machine learning” has tremendous potential to help health insurers leverage data and improve care, though one prominent insurance CEO argues that such disruptive technologies will never be able to replace the valuable role of clinicians. At UPMC, the Pittsburgh-based integrated health system’s investment in big data analytics gave it a “$1.6 billion advantage,” Pamela Peele, chief analytics officer for the company’s Insurance Services Division, tells Healthcare Finance. Peele’s team, she says, invented its own models that marry predictive analytics with claims and local demographic data. Then machine learning–a process in which software roots out trends that the system can act on–analyzes the data. For example, UMPC conducted “pure text mining” in about half a million clinical notes in members’ electronic medical records to look for word signals…


Link to Full Article: ‘Machine learning’ a boon for insurers, but can’t replace human touch in healthcare