Staffing a skilled nursing facility can be a daily chore, requiring administrators to strike a delicate balance of the right number of nurses.

That’s been the case for Greystone Healthcare Management Corp., based in Tampa, FL, with a network of 27 facilities. It provides an array of services that include skilled nursing, assisted living and home health. In January, the company had to use about 5,000 nurse agency hours to meet patient needs. That ballooned to about 18,000 in April, said Chris Masterson, senior vice president of clinical innovation.

A robust economy means companies pay more than a certified nursing assistant’s typical starting wages, and schools are unable to churn out licensed practical and registered nurses fast enough to meet demand.

“We just can’t compete with the labor-intensive work that a CNA has to accomplish for $10 or $11 an hour, compared to Walmart or Amazon for $15 an hour,” Masterson said.

That’s led to Greystone collaborating with a team of doctoral students from the University of South Florida and Purdue, trying to come up with a better way than the typical one-size-fits-all approach to staffing, that has relied on past experience, government mandates and nurse-to-resident ratios.

Greystone shared one year’s worth of de-identified patient data — along with other info, such as “anecdotal” electronic health records, and past payroll reports — from one of its SNFs in Pinellas County, FL, which the students have used to a create a predictive model that forecasts upcoming service demands. It then helps managers determine the ideal mix of RNs, CNAs and LPNs to help meet the complicated demands of each nursing home resident.

“The nursing home is a complex system and we want to improve performance using engineering technology,” said Mingyang Li, Ph.D., an assistant professor at USF.

The team is currently testing a trial version of the model at the same SNF. Masterson and Li presented at the International Society for Gerontechnology World Conference in May, and have submitted a grant proposal to the National Science Foundation.