Colloquium: Multiple Predictively Equivalent Risk Models for Handling Missing Data at Time of Prediction

Colloquia
September 14, 2020 - 11:15am to 12:15pm
Dr. Sisi Ma
Affiliation: 
UMN BioMed Health Informatics
Location: 
https://umn.zoom.us/j/95990345657?pwd=YVg1RW90ODhQd2lhNU82dzg5SXd1dz09
Host: 
Ju Sun

Abstract: The presence of missing data at the time of prediction limits the application of risk models in clinical and research settings. Common ways of handling missing data at the time of prediction include measuring the missing value and employing statistical methods. Measuring missing value incurs additional cost, whereas previously reported statistical methods results in reduced performance compared to when all variables are measured. To tackle these challenges, we introduce a new strategy, the MMTOP algorithm (Multiple models for Missing values at Time Of Prediction), which does not require measuring additional data elements or data imputation. Specifically, at model construction time, the MMTOP constructs multiple predictively equivalent risk models utilizing different risk factor sets. The collection of models are stored and to be queried at prediction time. To predict an individual’s risk in the presence of incomplete data, the MMTOP selects the risk model based on measurement availability for that individual from the collection of predictively equivalent models and makes the risk prediction with the selected model. We illustrate the MMTOP with severe hypoglycemia (SH) risk prediction based on data from the Action to Control Cardiovascular Risk in Diabetes (ACCORD) study. We identified 77 predictively equivalent models for SH with cross-validated c-index of 0.77 ± 0.03. These models are based on 77 distinct risk factor sets containing 12–17 risk factors. In terms of handling missing data at the time of prediction, the MMTOP outperforms all four tested competitor methods and maintains consistent performance as the number of missing variables increases.

Bio: Dr. Ma is an assistant professor of Medicine in the Division of General Internal Medicine at the University of Minnesota. Dr. Ma's primary research interest is the application of statistical modeling, machine learning, and causal analysis methods in the field of biology and medicine. The questions she seeks answers to include: how to leverage big data and analytical approaches to (1) diagnose and prognose disease and disorders earlier and more accurately. (2) systematically and efficiently identify potential treatment targets for a given disease. (3) identify the best treatment for a particular patient. Dr. Ma also works on theoretical aspects of predictive modeling and causal modeling.