October 23, 2017 | Written by: Jason Gilder, PhD
Categorized: Blog Post | Value-Based Care
Share this post:
Jason Gilder, PhD is the Senior Director of Informatics and Analytics at IBM Watson Health.
Predictive analytics models have long been an asset in the healthcare industry thanks to the ability to forecast future events, risk and utilization. Their predictive power will continue to grow as they incorporate advanced technologies such as cognitive computing.
Applications that use cognitive computing can absorb and analyze both structured and unstructured data, from multiple sources, expanding the information that organizations can use in their predictive modeling. This can include data from wearable devices, a patient’s genetic makeup and volumes of text from medical literature.
Consider that a typical regression model only uses about 10 to 20 input variables across a fixed regression equation. Cognitive models, on the other hand, can use thousands of inputs in a flexible, dynamic environment. As this base of knowledge evolves over time, the application is able to understand, reason and learn in a way that resembles the human thought process.
The result: deeper insights and broader connections than what’s possible with traditional regression models.
Cognitive technology can bring a new level of context and logic to predictive modelling, helping providers better understand a patient’s or population’s risk profile, including expected utilization and the likelihood of adverse events (readmission, complications and more).
Equipped with this intelligence, hospitals and health systems will be able to make proactive decisions to improve quality of care and operational efficiency, preparing those organizations for value-based success.
Learn more about how cognitive computing can advance value-based health in the IBM Watson Health white paper: Predictive Analytics in Value-Based Healthcare: Forecasting Risk, Utilization, and Outcomes.