How to Improve Patient Outcomes with AI without Perpetuating Historical Biases
9AM PST / 10AM MST
Clinicians, care managers, administrators, and executives all want to harness AI to improve clinical outcomes and drive healthcare efficiencies. Unfortunately, historical data, which is the fuel of modern AI systems, is often tainted with unintentional biases. How can we use this data without cementing these biases into new AI systems?
Join us for a live webinar with Duncan Renfrow-Symon and Cory Kind, Customer-Facing Data Scientists at DataRobot, as they discuss how unintentional biases in historical healthcare data have affected provider/payer services and outcomes for some patients.
They’ll review examples of biases in the data supporting major studies, and in the algorithms used to determine risk. You will also see how the functionality within the DataRobot platform can be used for the critical task of detecting and mitigating these biases when time, resources and, most importantly, lives are on the line.
During this webinar, you will learn:
- The sources of AI bias encountered in healthcare data
- How to automatically identify potential biases in your own data
- How to apply a metric that captures and mitigates the biases you are trying to avoid