3 minute read

CommonSpirit Uses Predictive Modeling Tools to Forecast COVID-19 Cases

CommonSpirit is using de-identified cell phone data, public health data, and its own system's data to forecast demand and prepare for upticks of COVID-19 patients.
CommonSpirit is using de-identified cell phone data, public health data, and its own system's data to forecast demand and prepare for upticks of COVID-19 patients.

Chicago, Ill. — June 22, 2020 — CommonSpirit Health, which has 137 hospitals and more than 1,000 care sites serving patients across 21 states, is using de-identified cell phone data, public health data, and data from the system’s own care sites to forecast changes in health care demand and prepare for upticks of COVID-19 patients as communities nationwide relax social distancing orders.

The forecasts made possible by these predictive modeling tools provide valuable insight into the peaks and valleys of COVID-19 infection rates, enabling better planning for potential surges of new patients while hospitals continue working toward resuming their full spectrum of needed services. These models take into account fixed data, such as population and availability of health care providers, as well as variable data such as social distancing relaxation and normalized county new cases. While no single platform can make a perfect prediction, these models are one tool in CommonSpirit’s toolkit to help care sites prepare for a patient surge. The data are combined with careful operational planning and the agility to respond to any differences from the forecast.

“The balance of modeling and real-time validation prepares us to be informed yet also able to safely care for our patients and employees as we begin to resume needed health care services,” said Dr. Joseph Colorafi, System VP of Clinical Data Science for CommonSpirit Health. “We need to be prepared for asynchronous waves of COVID-19 patients in different regions until a higher level of herd immunity is present.”

CommonSpirit is able to generate a predictive outlook for roughly 75 percent of its markets across the country, including communities in Texas, California, Arizona, the Pacific Northwest, and parts of the Midwest and Southeast. In many cases, the data show that a surge of new COVID-19 cases can be expected about 2-3 weeks after social distancing requirements are relaxed.

CommonSpirit’s predictive modeling platform is built on a forecasting model that leverages several data sources, including local, state, and federal public health information, and de-identified data from cell phones and CommonSpirit care sites. The cell phone data show how much people are traveling outside their home communities, providing a direct measure of physical distancing. All data sources are de-identified and HIPAA-compliant, so that no information can be associated with a specific individual.

As one of the largest health systems in the country, CommonSpirit is able to use this platform to inform the way care sites meet the needs of local communities and help patients feel confident that they can safely receive the care they need.


About CommonSpirit Health

CommonSpirit Health is a nonprofit, Catholic health system dedicated to advancing health for all people. It was created in February 2019 through the alignment of Catholic Health Initiatives and Dignity Health. CommonSpirit Health is committed to creating healthier communities, delivering exceptional patient care, and ensuring every person has access to quality health care. With its national office in Chicago and a team of approximately 150,000 employees and 25,000 physicians and advanced practice clinicians, CommonSpirit Health operates 137 hospitals and more than 1000 care sites across 21 states. In FY 2019, Catholic Health Initiatives and Dignity Health had combined revenues of nearly $29 billion and provided $4.45 billion in charity care, community benefit, and unreimbursed government programs. Learn more at