Cleveland Clinic shares predictive model to help hospitals plan for COVID-19: 4 things to know

Cleveland Clinic created a model to help hospitals forecast patient volume as well as supply availability and needs in partnership with SAS, and is now sharing the predictive model with other health systems via GitHub.

Four things to know:

1. The analytic models are designed for healthcare organizations to create best-case, worst-case and most-likely scenarios, and can be adjusted in real time as the data changes. As a result, hospitals can make decisions about their COVID-19 response preparations based on initial numbers and then as the situation changes based on social distancing efforts, they can update their models.

2. Cleveland Clinic used the model to make decisions about ICU beds, PPE and ventilators. The health system decided to prepare for the worst-case scenario projections and built a 1,000-bed surge hospital at its education campus that could support COVID-19 patients who didn't need the ICU.

3. The health system also used the predictive model to make decisions about activating new labor pools.

4. The model is heavily informed by the SEIR model, describing the susceptible, exposed, infected and recovered over time. Cleveland Clinic and SAS developed the model based on a University of Pennsylvania open source model, and it updates continuously based on data from epidemiologists and data scientists. The model considers regional health and demographic variations as well as state-level assumptions about the situation.

More articles on digital transformation:
15 digital health, telehealth startups that raised millions in the past 3 months
Bayhealth, EMS teams implement new communication tech: 5 details
OSF Ventures joins $16M funding round for digital health company: 5 details