U.S. flag An official website of the United States government
  1. Home
  2. Science & Research
  3. Science and Research Special Topics
  4. Advancing Regulatory Science
  5. A Rapid Query Model to Address Prioritized COVID-19 Questions Using EHR Data
  1. Advancing Regulatory Science

A Rapid Query Model to Address Prioritized COVID-19 Questions Using EHR Data

CERSI Collaborators: University of California at San Francisco (UCSF): Atul Butte, MD, PhD; Rohit Vashisht, PhD

FDA Collaborators: Formerly of Office of the Commissioner: Amy Abernethy, MD, PhD; Tamar Lasky, PhD; Laura Roe; Joseph Franklin, PhD, JD; Aloka Chakravarty, PhD

Project Start Date: September 2020
Project End Date: February 2022

Regulatory Science Challenge

Rapid collection, analyses, and reporting of COVID-19 data is needed to support COVID-19 pandemic response efforts. Public health professionals, healthcare providers, and others, use multiple data sources to improve our understanding of how the pandemic is impacting communities, and guide public health interventions and services. Example data include but are not limited to: COVID-19 case identification, hospitalization, death rates, and distribution/effectiveness of available treatments.

To help address this need, researchers developed a rapid query model in collaboration with University of California Health (UC Health) and U.S. FDA. The model enables FDA scientists to query, or ask questions of, UC Health data collected from over 7 million patients. Answers to these questions can be provided within days of receipt.

Project Description and Goals

UC Health, comprised of 19 health professional schools, 5 academic medical centers, and 12 hospitals, has built a secure central data warehouse of 7 million patient records. Patients documented within this central data warehouse receive care both within and outside the state of California. UC Health created a standardized Health Insurance Portability and Accountability Act (HIPAA) Limited Data Set called the UC COVID Research Data Set (UC CORDS) soon after the COVID-19 pandemic began. The UC CORDS database captures medical information related to patients infected with COVID-19 and various aspects of their medical care. UC Health uses UC CORDS to rapidly answer questions posed by the FDA. The rapid-query model is a multi-step framework comprised of a computational pipeline used to rapidly query UC CORDS and refine the quality of the questions being asked by conducting follow-up statistical analysis. UC Health proposes to deliver preliminary answers to COVID-19 questions within 2-4 days of the question transmittal, and more detailed answers within two weeks of the finally refined question. Answers are delivered in regular reports accompanied by tables of counts and data visualizations.

Research Outcomes/Results

Examples of research outcomes of this project are as follows:

  • The UC Health team rapidly generated answers using real-world data to various questions posed by the team FDA on a daily or weekly basis. Example reports generated include but are not limited to: rates of COVID-19 treatment administration of Remdesivir, corticosteroids, interferons, cyclosporin A, tacrolimus, FK506, and Metformin. Daily medication counts were visualized using calendar plots that provided insights on the utilization of these medications throughout the pandemic.
  • The team participated, in collaboration with five other healthcare systems, in assessing the real-world performance of SARS-CoV-2 antibody tests in the United States.
  • A poster related to certain aspects of the rapid query model was presented in 2021 FDA science forum.
 
Back to Top