Lift Off! How Health Actuaries Are Using Non-Traditional Features and Advanced Methods to Increase Predictability

While algorithms like XGBoost and recurrent neural networks are all the rave, the backbone of predictive analytics and machine learning comes down to the right data and a strong foundation of engineering that data. In this presentation, actuaries and data scientists from the Cleveland Clinic and Milliman discuss novel data points and features that have provided increased predictability in models used across health care. This data ranges from publicly available information to patient-entered questionnaires, and includes social determinants of health, third-party consumer information, measures of functional and emotional status, and new twists on familiar concepts. The group will also discuss methods in feature engineering and their applications to real-world problem solving. Data and feature engineering are what really provide the lift in predicting outcomes. At the conclusion of this session, attendees will be able to: • Identify non-traditional data points to improve predictability • Understand various concepts regarding feature engineering • Explain how these two items help create better risk adjustment models

Joseph Dorocak, ASA,MAAA

Senior Financial Analyst

Joe Dorocak is a new Associate in the Society of Actuaries, having completed all requirements in April, 2018. He has worked the Cleveland Clinic since 2012, where his focus has been the analysis of numerous shared savings contracts including Cleveland Clinic’s Medicare ACO and the Clinic’s Employee Health Plan, and was recently promoted to manage the organization’s Population Health Analytics team. Joe is a 2012 graduate of the University of Dayton where he majored in Applied Mathematical Economics.

Donald McLellan, ASA, MAAA

Director Risk Analytics

Don is the Director of Risk Analytics at the Cleveland Clinic. He is responsible for Population Health Analytics, Shared Savings Contract Analysis, and Payer Reimbursement. Don is a member of the American Academy of Actuaries and the Society of Actuaries. He has nearly 25 years of healthcare experience including both the payer and provider perspectives.

Joseph Long

Joe Long is an assistant actuary and data scientist with the Minneapolis office of Milliman. He joined the firm in 2013. Joe specializes in the application of data science and machine learning within the actuarial setting. In addition to his consulting work for insurers and other risk taking entities, this work includes assisting health, life and long-term care actuaries developing products and tools that utilize predictive modeling. Joe has assisted in developing models for predicting relative healthcare costs included in the Milliman Advanced Risk Adjusters (MARA) and Milliman IntelliScript software products. He is also a key team member pioneering Milliman's internal research that applies predictive analytics to long-term care industry experience to develop first principle assumptions and claim cost guidelines. Prior to joining the firm, Joe taught statistics while pursuing his Master's in applied statistics.

Hans K. Leida, PhD, FSA, MAAA

Hans works in health insurance. He has consulted to insurance companies, Blue Cross Blue Shield plans, HMOs, healthcare providers, government health programs, and employers. Recently, he has been working on provider payment analysis and benchmarking, as well as advanced predictive modeling and risk adjustment. Hans is also frequently quoted in the media on healthcare reform in publications such as The Wall Street Journal, Reuters, Bloomberg Businessweek, and Modern Healthcare. Hans has completed many projects involving individual and group health strategy, pricing, and rate filings. In 2007, he co-authored a paper for America's Health Insurance Plans (AHIP) on the impact of guaranteed issue and community rating laws adopted by certain states in the 1990s. That paper (which was updated in 2012) has been widely cited with the advent of federal healthcare reform, most notably by the Chief Justice of the U.S. Supreme Court in his majority opinion for the case upholding premium subsidies in federal exchange states (King v. Burwell). More recently, Hans co-authored the second edition of Individual Health Insurance. This textbook has been on the Society of Actuaries' exam syllabus for many years. Hans also has significant experience with risk adjustment and predictive modeling of healthcare costs. He was the lead developer of the prescription-drug-based risk adjuster included in the Milliman Advanced Risk Adjusters (MARA) software product.

Key:

Complete
Failed
Available
Locked
Total Watching
1 Question
1 Question To view the session, please answer the "Total Watching" survey to let us know how many people are watching with you and then click Enter Live Session (sessions will open 5 minutes before each session begins).
Webcast Session
08/25/2020 at 11:00 AM (EDT)  |  75 minutes
08/25/2020 at 11:00 AM (EDT)  |  75 minutes
Evaluation
Select the "Complete Evaluation" button to begin.
Select the "Complete Evaluation" button to begin. Please complete the evaluation for this session.