“All Models are Wrong, Some are Useful”:
Building Useful Models for Clinically-Effective, Resource-Efficient Healthcare Delivery
Prasanna Desikan, PhD, MS, MHA
Director, Health Risk Analytics
Over the last two decades there has been tremendous technological progress in the computing infrastructures and methodologies, and the enabling technologies to collect large amounts of variety of data. This has resulted the field of data science emerging and exploding in its capabilities to be applicable to a variety of problems. As the field expands, it is become increasingly significant to navigate through its vastness to build the right analytic solutions that are useful. As George Box, one of the most accomplished statisticians of the 20th century, famously wrote: “All models are wrong, some are useful”, it is very important to choose the right model that can be used and operationalized. Dr. Desikan starts the seminar with a description of the landscape of Data Science as a field, and its applicability to healthcare delivery. He helps us understand the area of healthcare analytics, what questions can be answered with available data sources and digs deeper into case studies that are examples of how to build useful models to improve quality healthcare delivery while managing resources efficiently. He addresses the challenges and approach in general to build such models. Finally, Dr. Desikan points out the future of healthcare analytics as it relates to the advancements in the enabling technologies, how they impact and shape the analytics.
Building access instructions for non-University guests
Display this event page on your smart phone at the Mayo Building public entrance on Church Street, or the Moos Tower public entrance on Washington Avenue.