Last week, Home Health Care News highlighted widespread reports of home health therapy layoffs and visit reductions following the implementation of the Patient-Driven Groupings Model (PDGM).

While therapy-visit reductions are tied to a number of factors, some stakeholders have specifically called attention to what they see as an over-reliance on predictive analytics tools and an under-reliance on clinical judgement at some agencies. Those stakeholders include the American Occupational Therapy Association (AOTA) and the American Physical Therapy Association (APTA), among others.

“We’re really trying to get our arms around and understand the newer … algorithms and predictive tools,” AOTA Vice President of Regulatory Affairs Sharmila Sandhu told HHCN. “That is a great concern. The clinical judgment of the therapist could be overwritten. What if they believe that the patient could benefit from one or two more visits a week and that is not being permitted?”

Backed by minority investments from Amedisys Inc. (Nasdaq: AMED), Encompass Health Corporation (NYSE: EHC) and Coliseum Capital Management, Nashville, Tennessee-based Medalogix is one of the market leaders in home health predictive analytics.

Its CEO and president, Elliott Wood, wants to set the record straight.

Generally speaking, predictive analytics tools like Medalogix are meant to offer home health providers a roadmap to appropriate utilization in a post-PDGM world that is seemingly full of endless twists and turns. They’re meant to complement the clinical experience, education and instincts of physical and occupational therapists — not replace those things altogether. 

“I would say, first of all, that it’s entirely possible some agencies are using predictive analytics tools to dictate visits,” Wood told HHCN. “And I would say that is an irresponsible use of technology, unequivocally.”

Inside Medalogix

Medalogix isn’t the only predictive analytics tool in the home health space, and the concerns raised by therapy professionals last week did not directly target the company.

Since its founding in 2012, Medalogix has grown into a data powerhouse, with several sizable providers using at least one of Medalogix’s flagship tools: Care, Touch, Bridge and Nurture. Besides its provider investors, other home health companies that have touted their use of Medalogix include AccentCare and Elara Caring, both of which previously discussed how the end-of-life data tool Bridge helped refine their operations.

Bridge is designed to help home health providers identify when a patient is nearing the appropriate time for hospice-level care.

“It really just comes down to taking care of patients,” AccentCare Chief Medical Officer Greg Sheff told HHCN in July 2018. “Part of that means making sure they’re in the right level of care. Hospice is a more robust benefit for patients, so if it’s a place that might be better for them, we want to figure that out early and have that conversation.”

Overall, Medalogix creates its actionable insights from information derived from about 30% of all home health patient volume. Some of the records Medalogix’s tools take into account go back 10 years.

“We’re able to look at the characteristics of patients who have come into home health, where they started in terms of clinical and functional impairment, but also where they are at discharge in terms of improvement,” Wood said. “What we’re able to do — using technologies that are available to us today — is find the mathematical relationships between utilization and the trajectory of patients.”

In other words, Medalogix built its entire business around the idea of determining just how much care an individual patient needs to achieve his or her best possible health outcome. By doing so, providers increasingly tasked with doing more with less are able to determine when the clinical value of a visit stops paying off.

Medalogix — launched well before PDGM and the new challenges it brings — has never factored reimbursement or providers’ bottom-lines into its calculations.

“I can unequivocally say that Medalogix has never incorporated a revenue expectation into a predictive model,” he said. “That has never happened.”

‘Driving into a lake’

Internally, Medalogix anticipates that roughly 10% of the entire home health industry will be using Medalogix Care by 2020’s halfway point. Medalogix Care is the particular tool focused on predicting optimal visit utilization and outcomes.

Instead of viewing predictive analytics tools as the captain steering the ship, agencies and their clinicians should view them as key resources for guiding decisionmaking and care planning, Wood said, emphasizing that input from nurses and therapists is still critical.

In some ways, he said, using predictive analytics to shape utilization strategies is like using Google Maps to inform commuting decisions.

“We have a customer whose chief clinical officer does a fantastic job in what we call the integration discussion — outlining how you begin to integrate AI and machine learning as clinical-decision support,” Wood said. “He’ll start by asking [clinicians] if they’ve used Google Maps or Apple Maps getting from one visit to the next, and you’ll see all these hands raised. Those tools let you see the most efficient route and alert you to any roadblocks that might get in our way … but you can still open your eyes and see if there’s a better path in front of you to get there.”

The danger in using those mapping tools is ignoring “the eye test” and common sense.

When that happens, users might run the risk of driving into a clear obstruction because of faulty directions, similar to what happened when The Office’s Michael Scott drove into a lake after taking GPS directions too literally in the TV show’s fourth season. 

“The agency that’s using a tool like [Medalogix Care] can’t force the nurse or the PT to drive into a lake,” Wood said.

As part of its PDGM strategy and overall evolution, Encompass Health has been using Medalogix, Homecare Homebase and other technology systems to generate objective, evidence-based care plans and drive incremental efficiency in administrative support functions.

As of the end of January, the Birmingham, Alabama-based Encompass Health had Medalogix deployed at 74% of its home health operations, April Anthony, who leads the company’s home health and hospice segment as CEO, said Friday during a fourth-quarter earnings call.

Encompass Health is on track to reach 80% or higher deployment by the end of the first quarter.

“One of the things I want to say with the Medalogix Care tool is it is not a tool that was designed under the PDGM model,” Anthony said. “It’s a tool that is designed to say, ‘What do patients need to achieve their highest possible outcome?’ And what we’re finding in the deployment of the tool is that it’s really creating that much greater specificity of care planning tied to specific patient issues and needs.”

In some cases, that means fewer visits. In other cases, it means more visits, she said.

Mark Tarr, the CEO of the overall Encompass Health organization, agreed. He also reaffirmed the idea that — when used responsibly — predictive analytics tools are meant to guide what PTs, OTs, nurses and other home health clinicians call for.

“Using our historical patient data, this module tells us how many visits a patient needs and when they need them to optimize the outcome for the patient,” Tarr said during the Q4 call. “It’s a scientific approach that helps guide, not replace, clinical judgment in developing the right care plan for each patient.”

Why now?

There are several reasons why the topic of predictive analytics has risen into a hot issue within the home health industry, according to Wood.

PDGM — a payment overhaul that shifts reimbursement away from volume and towar patient characteristics — is obviously a driver. On top of that, home health agencies are facing a phasing out of Requests for Anticipated Payment (RAPs), tighter claims review processes and a long list of other challenges.

The changing home health reimbursement climate puts margin pressure on agency operators, forcing them to become smarter and more efficient — or risk financial ruin.

“What we see with PDGM is all this financial pressure is happening. You’re having your rates changed as an agency, you’re having margins compressed, so you’re having all of this institutional pressure financially,” Wood said. “And the question is, ‘How can you continue to do what’s best for the patient?’”

Medicare Advantage (MA) growth is likewise a driver for predictive analytics.

Historically, MA has been a lousy payer for home health providers, with many plans gravitating toward rock-bottom rates and strict visit utilization. But with federal policymakers striving for parity between fee-for-service Medicare and MA, providers have begun to figure out innovative ways of working within the Medicare Advantage construct.

“You have a lot of organizations that see a future with Medicare Advantage partners,” Wood said. “They see a huge opportunity for home health to play a bigger role. But the way that you win within MA is taking risk. And only way you can take risk is if you have — and use — data.”

Supporting Anthony’s point about increasing visits in response to data insights, Wood has actually seen Medalogix Care implementations that showcased how more visits led to clear reductions in re-hospitalization rates.

“MA is an entirely different animal,” Wood said.