Jessica Gronsbell, a Data Scientist at Alphabet’s Verily Life Sciences, will present:

“Leveraging Digital Data for Clinical Research”

Abstract: The widespread adoption of electronic health records (EHR) and their subsequent linkage to specimen biorepositories has generated massive amounts of routinely collected medical data for use in translational research. These integrated data sets enable real-world predictive modeling of disease risk and progression. However, data heterogeneity and quality issues impose unique analytical challenges on the development of EHR-based prediction models. For example, ascertainment of validated outcome information, such as the presence of a disease or treatment response, is particularly challenging because it requires manual chart review. Outcome information is therefore only available for a small subset of patients in the cohort of interest, unlike the traditional setting where this information is available for all patients. In this talk, I will discuss semi-supervised and weakly-supervised learning methods for predictive modeling in such constrained settings where the proportion of labeled data is very small. I demonstrate that leveraging unlabeled examples can improve the efficiency of model estimation and evaluation procedures, which in turn substantially reduces the amount of labeled data required for developing prediction models.

Jessica Gronsbell is a data scientist at Alphabet’s Verily Life Sciences. Prior to joining Verily, she earned her PhD in Biostatistics at Harvard University under the direction of Prof. Tianxi Cai, and was a postdoctoral researcher in the Department of Biomedical Data Science at the Stanford School of Medicine with Prof. Lu Tian. Jessica’s primary statistical interest is in risk prediction using modern digital data sources such as electronic health records and mobile health data.

A social tea will be held at 11:00 a.m. in A434 Mayo. All are Welcome.