Biostatistics Weekly Seminar

Estimating risk in large public health and comparative effectiveness studies

Amber Hackstadt, PhD
Vanderbilt University Medical Center

There is an abundance of secondary data sources, including claims data and electronic health records, that provides a valuable resource. These data can inform medical researchers about patient care outside of the clinical trial setting, and in “real world settings” with the potential to greatly improve comparative effectiveness and health services research. Traditional statistical analyses that use observational data may produce biased estimates due to residual confounding. Several studies have supported the use of propensity score weighting to help reduce such bias but few studies have focused on using propensity score weighting for time-to-event outcomes in the presence of competing risks. In this seminar, I will explore the performance of approaches to estimate the cumulative incidence functions for time-to-event outcomes in the presence of competing risks when using propensity score weighting to address the lack of randomization. This project is motivated by a study comparing the effects of the anti-diabetic medications metformin and sulfonylureas after reaching a reduced kidney function threshold. The data sources includes electronic health records for nearly 100,000 US veterans linked to other large administrative sources. Large clinical trials that investigated diabetes treatment effects on cardiovascular outcomes have excluded patients with reduced kidney function, rendering this population understudied.

Zoom (Link to Follow)
27 January 2021

Topic revision: r1 - 18 Jan 2021, AndrewSpieker

This site is powered by FoswikiCopyright © 2013-2022 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Vanderbilt Biostatistics Wiki? Send feedback