JSM 2011 Online Program

The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.

Abstract Details

Activity Number: 572
Type: Contributed
Date/Time: Wednesday, August 3, 2011 : 2:00 PM to 3:50 PM
Sponsor: Section on Health Policy Statistics
Abstract - #302675
Title: Model Selection Stability for Population Health Risk Models
Author(s): Diane M. Richardson*+ and Daniel Louis and Mary Robeson
Companies: Center for Health Equity Research and Promotion and Thomas Jefferson University and Jefferson Medical College
Address: Philadelphia VA Medical Center, Philadelphia, PA, 19104, USA
Keywords: risk model ; selection ; stability ; population health ; hospitalization
Abstract:

Clinical diagnostic data available in regional health databases is a rich source of information for modeling population health outcomes. Models that predict patient risk are used for case-mix adjustment, profiling, and resource allocation. However, selection of a meaningful subset of predictors is plagued by instabilities that limit the practical utility of the models. To assess the potential of newer methods not widely used in population health applications to improve model stability, we developed models for hospitalization risk in a general population, comparing stepwise, lasso, and stability-based selection. We used health system data to create over 200 clinical and disease stage indicators as predictors, and modeled future risk of potentially avoidable hospitalization. We report measures of model selection and prediction stability, and illustrate issues relevant to the selection of highest risk patients for enrollment in enhanced care management programs. Our results demonstrate that model selection stability can be improved markedly, and suggest that model selection stability be reported as a performance measure for population health risk models.


The address information is for the authors that have a + after their name.
Authors who are presenting talks have a * after their name.

Back to the full JSM 2011 program




2011 JSM Online Program Home

For information, contact jsm@amstat.org or phone (888) 231-3473.

If you have questions about the Continuing Education program, please contact the Education Department.