Using Decision Science to Tailor Patient Care at Mercy Virtual
1904labs presented at Prepare.ai 2019
Mark Bennetts - Product Management, Mercy Virtual
Brandon Fischer - Practice Lead, 1904labs
Matt Pitlyk - Technical Lead, 1904labs
Prepare.ai Session: This is an exciting time in health care because of Machine Learning and Artificial Intelligence (AI). Cloud services and modern processing power have helped technology catch up to the research and theory behind AI. Health care organizations can now turn to the most advanced Machine Learning tools and techniques to streamline patient care workflow, identify new treatment plans, and automate diagnoses. Health care organizations and their data scientists are often well-positioned to leverage their data to create insights, ultimately, to inform and automate health care decisions. The revolutionary potential of Machine Learning and AI makes many organizations eager to put them into practice. However, the haste to apply Machine Learning can be challenged by a number of factors. These challenges can arise, not because of limitations of the technology, but from issues surrounding the project itself. To help illustrate these issues, we walked attendees through a case study from work 1904labs is doing with Mercy Virtual. This case study involves building a solution to predict one of the questions every health care company wants to know: the likelihood of hospitalization. Drawing on real world experience applying Machine Learning with leading global enterprises, we will explore some of the factors that challenge Machine Learning, discuss how to recognize when you "out-science" your problem, and what to do when you are faced with this realization. Join us as we explain the challenges we faced in training a machine to recognize high, at-risk patients, the hurdles encountered along the way, and how we moved forward.
Download / View the Presentation
