American Hospital Cleveland Clinic and SAS Share COVID-19 Predictive Models to Help Hospitals Around the World Plan for Current and Future Requirements

To fight the novel coronavirus pandemic, U.S. hospital Cleveland Clinic and analytics company SAS have created innovative models that help hospitals forecast patient volume, bed capacity, ventilator availability and more.

The models, which are freely available via GitHub, provide timely, reliable information for hospitals and health departments to optimize health care delivery for COVID-19 and other patients and to predict impacts on supply chain, finance and other critical areas.

Unlike some forecasts that focus on a projection based on a single set of assumptions, these analytic models were used to create worst-case, best-case and most-likely scenarios, and can adjust in real time as the situation and data change. For example, the models can factor in social distancing’s dampening effect on disease spread.

“These predictive models were developed jointly by two organizations that understand patient populations, data and modeling,” said Chris Donovan, executive director of Enterprise Information Management & Analytics at Cleveland Clinic. “We are sharing the models publicly so health systems and government agencies globally can use them in their own communities. Our hope is that others contribute their ideas and improvements to the models as well.”

Cleveland Clinic is using the models to support its decision making. With this information, Cleveland Clinic can predict and plan for future demands on the health system, such as ICU beds, personal protective equipment and ventilators. After reviewing possible COVID-19 surge scenarios in the U.S. generated by the models, Cleveland Clinic elected to activate a plan that prepared it for the worst-case scenario and has built a 1,000-bed surge hospital on its education campus for COVID-19 patients who don’t need ICU care. The hospital system also used the models to inform decisions about organizing and activating new labor pools.

The GitHub link where the models are available has been visited more than 1,700 times in the past two weeks, resulting in more than 50 downloads.

At the heart of the work is an epidemiological SEIR model in which people move through the stages of Susceptible, Exposed, Infected and Recovered over time. The SEIR model developed by SAS and Cleveland Clinic is based on a University of Pennsylvania open source model that has been recoded and expanded on the SAS® analytics platform and continuously improved with real-time feedback from Cleveland Clinic epidemiologists and data scientists. The resulting models include flexible control of model parameters and different model approaches that consider regional health and demographic variations and state-level assumptions.

“These models can help hospitals, health care facilities, state departments of health and government agencies forecast the impact of COVID-19 and prepare for the future,” said Steve Bennett, Ph.D., Director of SAS’ Global Government Practice. “The models can also assist more vulnerable, less developed health systems in the fight against COVID-19.”