Juniper Saves Money, Retains Staff with Voice Recognition Tech

The 21st Century Senior Living Community is a series brought to you by CDW, a provider of technology solutions and services focused exclusively on serving the healthcare marketplace. The series takes a clear-eyed look at how leading providers and their partners are creating the next generation of senior living communities by raising the bar on services, design, and technology.

Products like the Amazon Echo have already made their way into senior living settings, allowing residents to hear the dinner menu or get medication reminders simply by asking for them. But voice recognition also holds promise for making clinicians’ lives easier, while enabling senior living companies to cut costs and improve retention.

At least, Juniper Communities believes this to be true, based on initial results from a technology pilot with a startup called LexiconAI.


Bloomfield, New Jersey-based Juniper operates 22 senior living communities, offering services from independent living to long-term skilled care and post-acute rehab. In mid-September, the organization began testing LexiconAI at its Juniper Village at Brookline location in State College, Pennsylvania. The community is home to about 240 residents and offers the full continuum, from independent living to rehab and skilled care.

After being approached earlier this year by LexiconAI CEO Matt Rubashkin, Juniper CEO Lynne Katzmann (pictured above) decided to pilot the product. She believed that the software might help reduce time spent on documentation by letting caregivers dictate their notes rather than type them, she told Senior Housing News.

Juniper saw this play out in October, the very first full month that LexiconAI was implemented. Though the trial has been brief and involved just nine clinicians, the numbers have been striking.


On average, users have been saving 25 minutes a day on documentation time, according to data gathered by Juniper and LexiconAI. However, this varies quite a bit depending on how much time users were initially spending on documentation—one person was spending more than four hours a day on documentation and now has shaved off more than 30 minutes, Lexicon CEO Matt Rubashkin told SHN.

The time savings, coupled with improved documentation, translates to a healthier bottom line.

LexiconMD-Juniper Communities 2017 User Study, provided Courtesy Juniper Communities

The time savings equates to a value of $3,956 per year per registered nurse using the technology, Juniper and LexiconAI calculated. Rubashkin also worked with Juniper’s national director of wellness to calculate how much additional revenue Juniper would bring in on an annual basis due to improved documentation of care, based on what they’ve seen so far. That came out to $1,200 a year per registered nurse using the technology.

Reducing the burden of documentation also has helped Juniper keep clinicians from turning over, according to Katzmann.

“Where we were going to lose people, they’re now happy,” she told Senior Housing News.

How it works

Rubashkin studied biomedical engineering at Johns Hopkins University in Baltimore before making his way to California for a Ph.D. program in the Bay Area, at Berkeley and the University of California, San Francisco. After that he was with a consulting firm called Silicon Valley Data Science (SVDS). It was there that he began to see how much time physicians and other clinicians spent doing documentation, and set out to create a better technology solution to alleviate the issue.

It was only about a year ago, in October 2016, that he pitched the idea of a better speech recognition solution to SVDS. The company provided pre-seed funding and helped get the project off the ground; Rubashkin now has amicably parted ways with SVDS, which maintains partial ownership in the startup.

There are other voice recognition products out there serving the medical field, but these rely on the last generation of technology, Rubashkin said. He believes LexiconAI rises above the pack by offering an improved speech engine leading to fewer than one error per paragraph, which is nearly as accurate as a human typing. The product also has features that clinicians want and need, such as the ability to navigate by voice through different fields of an electronic health record, such as the PointClickCare system that Juniper uses.

However, LexiconAI can integrate with any EHR though an iOS application or web browser, Rubashkin said. The Juniper clinicians do their note-taking on iPads.

“They basically do progress notes after visiting with the individual resident,” Rubashkin explained. “They’ll open their PCC application, select their patient, create a new progress note, and instead of typing it, they’ll dictate it.”

Rubashkin also sees senior care as a part of the market that can especially benefit from voice recognition technology, given the increasing documentation demands in these environments as the population ages and regulatory requirements grow. Though LexiconAI has been tested in radiology, emergency departments and other settings, the senior living and care market is now the primary focus for the company, which is also in the process of raising capital to grow.

Though Juniper did not pay to do the pilot, LexiconAI now is available on the market and is looking for customers. Pricing is competitive with other speech recognition products and can be done either on the basis of individual subscription licenses or for a whole organization, according to Rubashkin.

Juniper, for one, is on board. The company is now rolling out LexiconAI to another 50 clinicians, including some in memory care and skilled care.

Katzmann sums up her enthusiasm by saying, simply, “It’s very cool.”

Written by Tim Mullaney

Companies featured in this article:

, , ,