Abigail Lewis, PhD

My research intersts broadly encompass geospatial methodologies, healthcare utilization, bias in health data, and integration of electronic health record (EHR) data with external data sources.

My dissertation covered selection bias reflecting healthcare utilization in EHR data. By incorporating external data sources, including US Census data and the area deprivation index, into research pipelines leveraging EHR data, we were able to better understand target population representation in the EHR, measure geographic dispariteis in EHR data quality, and identify geographic trends in performance of predictive models trained on EHR data. This work informs transparent dialogue around underrepresented groups and minoritized patient populations in health informatics research and establishes novel approaches for mitigating resulting biases in EHR data-based research.

As a postdoc in the Clinical AI and Informatics Lab at WashU School of Medicine, my work has transitioned to focusing on mobile health intervensions that can support equitable representation in healthcare research and predictive model development with a focus on applications in Neurology and Alzheimer disease.