Athenahealth Adding Agentic AI Patient Engagement Features

Boston-based health IT vendor athenahealth has begun rolling out several features as part of a new agentic AI-based patient engagement suite within its athenaOne platform for ambulatory practices. 

The company said the expanded suite in its cloud-based SaaS EHR includes new text and voice capabilities that let patients connect with a virtual assistant at their doctor's office to ask basic questions or schedule appointments.

Paul Brient, chief product and operations officer at athenahealth, recently described some of the new features in a conversation with Healthcare Innovation. 

One new feature that is being embedded directly in existing athenaOne workflows is waitlist scheduling, which will automatically identify open time slots due to cancelations and send patients text messages to fill them. 

“This is totally automatic,” Brient said. “If someone calls in and cancels an 11 o'clock appointment, text messages go out and say, 'Hey, we have 11 o'clock open today. Do you want it?’ You have five minutes to respond. If you say yes, you're in. If you say no, we go the next person. It's a win-win for everybody. It's great for the patients. It's great for the practices. It takes a bunch of load off the practices. Today a lot of those cancellations go unfilled,” Brient said. “This is too new for me to give you stats, but probably in three or four months we should be able to have some good stats around how many slots they were able to fill that otherwise would go unfilled. But I say that is the ultimate no-brainer.”

Another new capability, Patient Conversations, is set to deliver both text and voice functionality to give patients a faster, more convenient way to connect with their doctor’s office. 

The texting feature, currently in alpha and scheduled to be available more broadly later in the first half of this year, combines real-time two-way texting and secure web chat, handling questions about office hours and insurance. 

When needed, the patient can be handed off directly to staff for additional follow-up, without losing context. The voice feature, available in the second half of this year, is a virtual front desk assistant that delivers multilingual support for inbound calls, handling the same routine requests without wait times. Following the initial text and voice functionality, advanced capabilities such as real-time language translation and intelligent task automation will let staff prioritize the interactions that matter most, athena said. 

“Multilingual used to be a huge problem, but with LLMs now it's not a problem at all,” Brient said. “It has been a huge unlock for us, because we used to have to translate everything by hand and do all this kind of work. Now we can put it into an LLM, and it does a really fine job basically in real time.”

One of the alpha users of the texting features has been Tracy Witz, practice manager at Waisbren Medicine, a small primary care practice in Cape May, N.J. The texting feature is starting to reduce the volume of calls, she said, adding that previously they would have to call a patient and give them instructions if the doctor was making a change to their medication. “Even though you get a patient to repeat things back to you, you never know if they've really got it,” she said. “What I love about the texting is now if we need to have that conversation, we're still texting them that information so that they have a record to go back to.” The practice also then has a record of that text-based conversation. 

Witz said the practice is seeing a reduction in the volume of calls around prescriptions. “We don't need to speak to a patient. We can just shoot them a text and let them know the information. Being able to text information about refills and respond to their questions via text is so helpful,” she said. “Sometimes athena is ahead of what we think we need or know we need. We've really been happy with the progression of the services within athena. It's been really good for us.”

Athenahealth is working to offer enhanced patient self-scheduling. Brient noted that in the past there have been both technology and cultural issues to overcome to enable patient self-scheduling.  

“Providers do want control over their schedule,” he said. “For instance, they don't want to have five new patients in the morning. They want to have a rhythm to their business. It has been  really hard to express that in technology, historically, in a way that you could put it out there for patients to self-schedule. With AI, though, we can observe the past schedule and the doctor can describe their preferences to the system the same way they would to a human scheduler. AI can then express that in self-schedule. So it is a technological uplift that will enable self-scheduling to work.”

An intelligent patient assistant feature will allow people to reschedule appointments and it will respond to patient questions about their payment balance. “The idea is it's kind of chatGPT meets your patient portal, where you can ask questions,” Brient explained. “We're starting with the more administrative side of things.”

He added that there are other improvements involving the patient portal, including self check-in for appointments before you go. “The next iteration of that will allow you to just have a conversation with AI about why you're coming, and have it ask you some questions, and then it will fill in all the forms.”