Hospitals across the U.S. face a challenge when forecasting the number of COVID-19 patients they'll treat and Cedars-Sinai Medical Center uses dozens of questions to help determine the number they should expect.
The hospital asks numerous questions such as: Who will need intensive care treatment or a ventilator? How many coronavirus patients will need to be hospitalized? And how much PPP will we need?
These questions are then used in a machine learning platform by a data science team, according to a press release from Cedars-Sinai. This platform will predict staffing needs, local hospitalization volumes and the rate of confirmed cases among other things. The team runs multiple rounds of forecasting to ensure the data is 85% to 95% accurate.
"Our goal is to have the capacity and the right care available every day to treat the patients who need us, which fluctuates on a daily basis," Michael Thompson, executive director of Enterprise Data Intelligence at Cedars-Sinai, said in the press release. "We need to match that daily demand with the necessary resources: Beds, staff, PPE and other supplies."
Estimates produced by Thompson's team help managers schedule the proper amount of employees, determine the amount of medical supplies needed and how many hospitalizations to expect.
The program was first designed to optimize how care is provided to patients, but when the coronavirus pandemic came about, the program was altered for current needs.
Before the pandemic, the program would analyze patient vital signs to predict treatments needed, give the likelihood a patient will be readmitted and tell which patients were satisfied with their hospital care.
The system helps Cedars-Sinai learn from past mistakes because the program gets "smarter" from mistakes it makes, Thompson said.
"If it predicts that tomorrow we’ll have 100 COVID-19 patients, but that actual number turns out to be 90, then the platform automatically goes back and tries to relearn what changed to cause the outcome to be different," Thompson said in the press release. "The platform hones its ability to recognize patterns and becomes smarter every day."
