Amedisys Turns to Predictive Analytics to Lower Voluntary Turnover

In the home health business, providers rise and fall with their people.

Recruitment and retention difficulties have been the bane of home health operators for years. Even during a once-in-a-generation pandemic filled with health hazards, cash-flow challenges and personal protective equipment (PPE) problems, it was staffing shortages that kept the bulk of organizations up at night.

That’s true because, without the expertise of nurses, therapists, home health aides and other in-home care professionals, providers simply don’t have a business, Amedisys Inc. (Nasdaq: AMED) Chief HR Officer Sharon Brunecz told Home Health Care News.

Advertisement

“At the end of the day, we are a business that revolves around people, right?” Brunecz said. “We don’t make things. The people, they’re our service. As a provider, our ability to attract and retain the talent we need to take care of our patients is job No 1.”

The overall employment of home health and personal care aides alone is projected to grow 34% from 2019 to 2029, much faster than the average for all occupations, according to the U.S. Bureau of Labor Statistics.

Meanwhile, the number of registered nurses leaving the workforce each year has grown steadily from 2010 to 2020, with nursing school enrollment starting to level off. There are similar supply-and-demand imbalances among PTs and OTs as well.

Advertisement

These previously known projections and new trends related to the COVID-19 emergency have home health leaders evaluating their recruitment and retention strategies more than ever.

To maintain a robust workforce, Amedisys has taken several “traditional” steps to lower its turnover rates. In 2018, for example, the Baton Rouge, Louisiana-based home health, hospice and personal care services provider launched a cash-bonus program for high-performing care centers.

But Amedisys is also starting to ramp up its more innovative staffing strategies, specifically by leveraging predictive analytics and a proprietary data platform, Brunecz explained.

“We think we might have found some of the secret sauce,” she said.

‘We challenged ourselves’

Over three years ago, Amedisys executives made the decision to invest in ways to better understand turnover. If the company better understood how and why workers were leaving, it could theoretically jump into action to get them to stay, Brunecz said.

With that goal in mind, Amedisys implemented a “robust employee-listening platform” in the fall of 2018. That initiative revolved around onboarding interviews, quarterly pulse surveys, exit interviews and more.

“We challenged ourselves and said, ‘You know, if we know why people are leaving, then we should be able to take that information and do something more proactive with it,’” Brunecz noted. “If we see when a clinician is at-risk [of leaving], can we intervene to change the outcome, to reduce turnover?”

After kickstarting that data-collection effort, Amedisys began analyzing the facts and figures toward the middle of 2019, running them through its predictive-analytics engine.

That analysis took time, partly because of the company’s sheer size, according to Megan Ambers, the vice president of workforce strategy and HR innovation at Amedisys. As of mid-May, Amedisys employed roughly 21,000 people across 514 care centers in 39 states and the District of Columbia.

“There were some interesting ‘risk drivers’ that were discovered by building out the AI model,” Ambers told HHCN. “There were over 20 different data points for every employee that was used in the model design. In aggregate, there were over 20,000 different kinds of components that came together to build out this model.”

The final product: a model that identified 36 different key risk drivers for all Amedisys employees, tailored to each individual based on role, tenure and similar factors.

“This has been a focus and a journey for us over the last few years,” Ambers said. “It has always been incredibly important to the organization. Really understanding each of our employee populations has been part of that success in [lowering] our turnover.”

Ambers and Brunecz declined to dive into all of the risk drivers, hesitant to give away that “secret sauce.” The HR leaders did give HHCN a few examples, however.

Beyond all of the usual reasons home health workers fall into an “at-risk category,” the Amedisys tool highlighted the importance of schedule stability. Clinicians don’t just want their hours, Ambers explained; they want a reliable routine without too many last-minute changes.

“If my shifts change different days every week, that causes some variability,” Ambers said. “That was something else that I think was interesting that we just really hadn’t quantified in the past.”

Especially during the past year, exposure to death was a risk driver, though one that varied greatly depending on role. Too much exposure to death led to an increased chance of turnover during a 90-day window, but hospice nurses were able to be around death more often compared to a home health nurse.

“If you think about caregiver burnout, a lot of things that we’re hearing right now are around death encounters,” Ambers said. “As caregivers are supporting people who are in transition, the volume of life transitions that they experience has a very different impact on, per se, a home health nurse versus a hospice nurse.”

Understanding the data

It wasn’t just enough to build a predictive-analytics tool for tracking and understanding turnover, Ambers said. Amedisys also wanted to make sure its leaders had streamlined access to the complex data the tool generated — and that they could easily make sense of it.

In turn, the company developed an overlaying application on the model itself that notifies leaders when they have a team member who is nearing the turnover tipping-point. That application then offers resources and guidance for intervention.

“We see that when a leader engages, we have a much higher likelihood of retaining that person,” Ambers said. “When a leader doesn’t engage with a high-risk team member, their turnover is 70% higher than those folks that they engaged with.”

So far, the predictive-analytics investments appear to be paying off.

Amedisys ended 2020 with an overall voluntary turnover of 18.3%, according to supplemental information shared by the company along with its Q4 financial results. At the end of Q1 2021, Amedisys had lowered that figure to 15.9%, a 20% improvement over the prior year’s first quarter for additional context.

The company released the notification application in Q3 of last year. Currently, it has yielded a 100% “leader-intervention rate.”

As for what’s next, Brunecz said Amedisys will continue to explore ways for boosting retention. In the future, it hopes to deploy its data capabilities to recruiting as well.

“We’ve done a phenomenal job here, I’ll say, on the ‘back door,’ focusing on retention and reducing turnover,” she said. “Now, we’re working on, ‘How can we get more predictive in our hiring and onboarding process?’”

Companies featured in this article: