New AI tools target providers looking to succeed in TEAM demo

This article was originally featured in Modern Healthcare.

Technology companies and post-acute providers say new artificial intelligence tools could be secret weapons for those in the Transforming Episode Accountability Model to keep patients from bouncing back to the hospital.

Some tools flag at-risk patients, while others consolidate and streamline information on patients moving from acute care to post-acute care in TEAM. Software companies and post-acute providers offering the technology say AI could give health systems, nursing homes and home health companies an edge in securing collaborators for TEAM.

But the model launches in January and some tools are just hitting the market, making it hard to gauge if they will resonate with providers still struggling to figure out how to navigate the new payment model.

Related: Health systems look beyond the obvious for AI products

“The people that are ahead of the game, are starting to [invest in technologies] now. The ones that are trying to figure out how many cases they will have and where patients will go are still in the early evaluation stage,” said Phyllis Wojtusik, executive vice president of value-based care at healthcare technology company Real Time Medical Systems.

Under the Centers for Medicare and Medicaid Services’ TEAM model, hospitals will receive a bundled payment for care around five procedures. CMS encourages them to partner with companies providing post-acute care, as it will account for approximately 60% of spending tied to the episodes of care.

New AI tools for post-acute care providers aimed at preventing rehospitalizations have hit the market over the last month ahead of the new pay model. Some products aren’t targeted directly at TEAM, but could benefit from the program.

For example, Software company Homecare Homebase rolled out Predict last week. The AI technology sits within Homecare Homebase’s electronic medical records system for home health providers. It uses machine learning to assess which patients are at risk of going back to the hospital and alerts clinicians if additional care is needed to keep them at home, said Haley Woods, director of product management, AI and innovation at Homecare Homebase.

The company is also adding an enhancement that will flag patients participating in TEAM, Woods said.

“It will be some kind of visual indication in addition to the existing product where we are saying the patient is high, medium or low risk and here is why,” she said.

In a pilot for Predict earlier this year, Woods said two home health companies saw about a 20% reduction in rehospitalizations.

A similar AI tool — Real Time Medical Systems’ TEAM Insights — scours data within nursing homes’ electronic records for TEAM patients at risk of rehospitalization. The tool alerts both skilled nursing facilities and health systems about patients that may require additional care.

Real Time Medical Systems also introduced its new AI technology last week and is already in discussions with health systems interested in buying the tool, Wojtusik said.

“Some of these hospitals haven’t done value-based care and really don’t have tools or processes in place to complete that circle,” she said. “We are giving them the infrastructure to do that.”

Technology company PointClickCare is offering a different AI option. The company rolled out its Reason for Transfer tool in early September to help track skilled nursing patients who end up in emergency departments. The AI technology connects hospitals’ electronic medical records systems to nursing homes’ EMR systems. The tool synthesizes patient data to help ED clinicians make smarter care decisions and helps avoid unnecessary readmissions, said Brian Drozdowicz, chief revenue officer at PointClickCare.

He said the company did not design the AI tool specifically with TEAM in mind, but participating providers have expressed interest in it.

“Historically, there hasn’t been a ton of communication between skilled nursing facilities and emergency departments when something happens,” Drozdowicz said. “[The technology] gives clinicians the full picture of what has happened to that patient over the past 24 hours to make the right thing happen.”

Some post-acute care providers that want to collaborate on TEAM are still weighing their options. Bayada Home Health Care is looking at new AI tools on the market, but may develop its own tool instead, a spokesperson said. The company provides home-based care across 22 states.

AccentCare is developing its own AI tool, rather than buying technology from a vendor.

“We think we can build something that is more accurate for the specific need of the patient on the right modality,” said Chris Mayne, AccentCare president of home health.

He said the Dallas-based company is developing AI technology that will help the home care company determine if a patient might be better suited for non-medical home care instead of home healthcare, which could help cut costs. AccentCare operates across 22 states. Mayne said the company is in talks with 10 to 20 hospitals participating in TEAM.

Some hospitals taking part in TEAM are still sizing up post-acute providers for partnerships, as well as their own plans for AI technology.

Three Baptist Health hospitals will be participating in TEAM. The Louisville, Kentucky-based healthcare system is using AI to support patient care and reduce readmissions, a company spokesperson said. But he did not say if the company is making additional AI investments or considering partnerships with post-acute providers that are.

Southcoast Health has been using AI technology from PointClickCare that flags at-risk patients in nursing homes for two years at its three Massachusetts hospitals, said Jennifer Schachter, the health system’s program manager for accountable care organizations. All three hospitals will participate in TEAM and Schachter said the New Bedford, Massachusetts-based health system is still evaluating emergency department transfer technology for the model.

“We are just trying to figure out which tools we do need, how can we utilize them and make sure we can get the best outcomes for these patients,” Schachter said.