Session – Digital Health and Healthcare

Deep Learning and its Application in Digital Health

Speaker: Dr. Yugang Jia (贾宇岗)

Data Science Manager

Verily Life Science, Boston, MA 02110

Mailing Address:

10 Wickham Road, Winchester, MA, 01890



Dr. Yugang Jia is a data science manager in Verily Life Sciences (formerly known as Google Life Sciences), where he joined in 2020 as Head of Data Science for Verily Health Platform business. He obtained his Ph.D. degree in Electrical and Electronic Engineering from University of Bristol, UK, in 2007. He was VP of healthcare data science at Fidelity Investments from 2017-2019 and senior scientist/innovation lead at Philips Research from 2007 to 2017. His research interests are mainly in data science/artificial intelligence and their application in digital health and healthcare informatics. Dr. Jia has worked with leading physicians, Accountable Care Organizations (ACO), health systems, and payers to accelerate the adoption of data science and machine learning in various clinician and patient decision support applications and have more than two dozens of pending/granted patents in the field of healthcare analytics. He was on the technical program committee of the International Conference on Healthcare Informatics (ICHI) in 2015, 2016, and 2017.


Total joint replacement (TJR) is one of the most commonly performed, fast-growing elective surgical procedures in the United States. Given its huge volume and cost variation, it has been regarded as one of the top opportunities to reduce health care costs by the industry. Identifying patients with a high chance of undergoing TJR surgery and engaging them for shopping is the key to success for plan sponsors. In this talk, we experimented with different machine learning algorithms and developed a novel deep learning approach to predict TJR surgery based on a large commercial claims dataset. Our results demonstrated that the performance of the gated recurrent neural network is better than other methods regardless of data representation methods (multi-hot encoding or embedding). Additional pooling mechanisms can further improve the performance of deep learning models for our case.