Memorial Health uses chatbots to boost patient experience, streamline workflows

The health system is using AI to provide COVID-19 risk assessments, virtualize waiting rooms for physician appointments and automate the PCP-to-specialist referral process.
By Bill Siwicki
11:59 AM
Memorial Medical Center building in Springfield Illinois

Memorial Medical Center, Springfield, Illinois

Memorial Health System, based in Springfield, Illinois, believes that one of the key challenges the healthcare industry faces today is finding ways to improve communication and engagement with patients. Healthcare may be a business, but it's ultimately about people, and one cannot expect to get anything done unless they have a way to successfully interact and engage with patients over the course of time in meaningful, lasting ways.

THE PROBLEM

Hospital-centric health systems are at risk of being commoditized in a value-based care world, said Jay Roszhart, president of the MHS ambulatory group at Memorial Health System.

"Historically, healthcare has relied on episodic, one-to-one interactions to deliver care, and that is no longer enough," he explained. "There's a lot of talk about consumerism in healthcare, and several big, successful consumer companies are leveraging their customer engagement pedigree in an effort to provide patients with a new healthcare alternative."

While local care from community-based providers may be in patients' best interests, patients aren't going to put up with local and "archaic," he added. Healthcare providers must reimagine how they build lasting, meaningful relationships with their patient populations, and the only way to do that at scale is with technology, he said.

Memorial Health System faces the same challenges as other large providers. It serves hundreds of thousands of patients across a full spectrum of care, but it needed to find a way to connect and communicate in different ways.

"Technology alone is not enough," Roszhart insisted. "Today's consumers want simple, smart and mobile tools. You cannot rely on a patient to download an app and remember yet another password. Success requires multiple ways, working together across platforms, to connect and engage patients. We need 90% technology adoption rates, not 9%."

PROPOSAL

To help with this challenge, Memorial Health System turned to LifeLink, a vendor of AI-based conversational technology to automate intake for telehealth and in-person visits.

"We were determined to find a technology solution that could engage a large patient population at scale, one that was flexible enough to automate hundreds of different patient workflows, and something that was simple," Roszhart said.

LifeLink provides a technology platform that uses conversational messaging to interact with patients on their mobile devices. Interactive chatbots serve as digital assistants for patients and are designed to help guide them through a range of healthcare tasks, ranging from scheduling appointments to checking in for visits, to making sure they follow their referral plans.

"Importantly, patients don't need to download an app or remember a password to interact with the LifeLink chatbots in a secure, HIPAA-compliant environment," Roszhart noted. "The technology is designed to integrate with the EHR, which enables the level of personalization that patients expect. The chatbots know the patient, their history, key dates, their prescriptions, etc."

The flexibility of the LifeLink platform was important to Memorial Health System, Roszhart said.

"Workflow components, such as scheduling, confirming appointments, reminding, completing forms, screening and educating, are built and tested," he explained. "Each solution is also configurable so it can rapidly be deployed to support the unique ways we handle scheduling, or our specific form requirements. For example, when the COVID-19 pandemic hit, we had a conversational, LifeLink-powered screener solution deployed in about a week."

MEETING THE CHALLENGE

Memorial Health System focused the first phases of the conversational technology on what Roszhart described as three important, high-volume, high-value workflow areas.

First, in March 2020, the health system rolled out a chatbot to screen patients and provide COVID-19 risk assessments. In the early days of the pandemic, patient access teams and urgent care centers were overwhelmed with patients, who were worried they may have coronavirus.

The chatbots served as a frontline, self-assessment gateway that captured information and provided patients with answers and guidance 24 hours a day. This helped staff scale and prioritize intake processes by reducing inbound call volume by 25%, and also gave patients some peace of mind.

Second, the health system focused on virtualizing waiting rooms for primary care physician and specialist appointments. The social distancing requirements of COVID-19 prioritized this, but the waiting room experience was actually something that needed change anyway, Roszhart said.

"The whole concept of a waiting room is not a patient-friendly experience," he said. "LifeLink chatbots digitize the intake forms and pre-appointment check-in processes through a mobile chatbot. When patients arrive for their appointment, they remain outside the facility until the exam room is ready and then are alerted via text message when it is time to come inside."

This new capability has been reducing no-show rates, improving backend processes and making patients happier by eliminating the need to sit in a waiting room full of other potentially sick people, he added.

Third, Memorial Health System is automating the primary care physician-to-specialist referral processes.

"Instead of handing a patient a stack of paper instructions, a digital assistant will reach out to a patient and assist them in getting an appointment scheduled with the right specialist," Roszhart explained. "This is the beginning of a broader adherence program that we will eventually fully automate with conversational technology."

It's important that patients follow their plans when it comes to follow-up appointments, wellness programs, medications and behavior. A conversational digital assistant is an ideal way to reach out, connect and remind people to take care of themselves. It's a big part of the future of healthcare, Roszhart contended.

RESULTS

"When you talk about results for patient engagement, the big one is how well you are doing [with] reaching and interacting with the patients," Roszhart said. "We track engagement as a top-level metric. If the chatbot reaches out to 100 patients to remind them about an appointment, how many acknowledge they will arrive and what percent of them complete all the forms digitally before they arrive for their appointment?"

Once patients are using the technology to complete tasks that previously required staff members to finish, the health system begins to get into a number of other, potentially significant process efficiencies that are just now beginning to show, he added.

"From there, we also track ROI," he said. "Simply put, chatbots allow for a level of scale that would otherwise require human capital. Our savings on FTE expense vastly outweigh our expenses related to operating the chatbot."

Further, the health system focuses a lot of its results efforts around how these conversations with patients help to build loyalty to the health system. The previous model of simply focusing on a patient experience during a once or twice a year visit with a doctor was not enough to drive loyalty, Roszhart said.

"We need many more connection points and engagement opportunities, augmented with easy-to-use technology, in order to successfully build loyalty," he said.

ADVICE FOR OTHERS

"You need to consider what it will take to get the technology adopted by patients," Roszhart advised. "Just because you build it does not mean they will come – especially when you're talking about engaging a massive, diverse patient population. One of the things that drove us to conversational technology was the ease of use. If you can talk, you can use it."

The good news is that most patients now carry a smartphone, so, on the one hand, there's a minicomputer in their pocket they spend hours using every day, he said.

"On the other hand, the strategy is not asking them to download another mobile app," he added. "There are more than 300,000 healthcare apps out there, and the vast majority are not used. If you can find a way to engage the mobile user with the embedded smartphone capabilities like web browsers and text messaging, the odds of patients actually using the technology increase."

Finally, look for flexibility in the solution, he suggested. Conversational technology may look very simple and basic on the surface, but below the waterline is a fairly robust technology framework that's required to pull it off, he pointed out.

"If the solution asks a patient to type 'Y' to accept an appointment or 'N' to decline, that's not conversational and will likely not meet expectations," he said. "Apple and Amazon are setting the conversational bar with Siri and Alexa. And while that kind of open-ended, language-based conversational engagement is years off in healthcare, we can do better than 'Y' and 'N.'"

Well-crafted conversational messaging is ready for prime time in healthcare, but it's important to look closely at how exactly the user experience unfolds, Roszhart said.

"By simply eliminating telephone hold queues, clipboards, waiting rooms, passwords, complexity and administrative hurdles, healthcare can take a big leap forward with patients," he concluded. "And, once patients begin to embrace the technology, we will see multimillion-dollar savings in operational waste and overhead."

Twitter: @SiwickiHealthIT
Email the writer: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.

More regional news

(Photo courtesy ChristianaCare)

By
Pain management therapy demonstration, Airrosti low-code EHR

Pain management therapy demonstration at Airrosti. (Credit: Airrosti)

By
Want to get more stories like this one? Get daily news updates from Healthcare IT News.
Your subscription has been saved.
Something went wrong. Please try again.