Making Voice Assistants Allies In Health

Orbita CEO Bill Rogers has a relative with dementia. As dementia patients often do, she repeats a lot of the same questions, and it drives her aging husband crazy. But there’s someone who’s happy to answer those questions as many times as she asks, and she never gets impatient. That person’s name? Alexa.

Voice-activated technology has a lot of potential in healthcare. There are cost and time savings to be reaped, vitals to be tracked and medication reminders to be issued. Proponents believe that voice tech can help the same number of professionals give better care to a greater number of people. But some of the benefits, such as what Amazon’s virtual assistant has done for Rogers’ relative, can’t be quantified.

So, just what does the healthcare world stand to gain, and what will it take to see greater adoption of voice-activated tech in the space? Industry professionals weigh in.

The Story So Far

To date, voice technology in healthcare is mostly being used to automate call routing and self-service tasks. It can connect callers with the right department and save them from long hold times or repeating their query to multiple agents. It can help them find the nearest pharmacy to fill a prescription or get medical testing done the same day. This is the low-hanging fruit of voice tech in healthcare.

While the patient experience benefits are clear, and perhaps greater than the financial savings, there are some small cost savings taking place in this scenario. Next Caller Director of Customer Success Tim Prugar said it costs an average of $0.30 per call to manually identify the caller. If a healthcare agent can bypass that part of the conversation and get straight to the meat of it, those savings do add up, Prugar said.

Orbita takes it a step further by using voice tech for remote patient monitoring, medication adherence, chronic care management and patient education. Combined with other IoT devices like blood pressure sensors, scales or respiratory trainers, this tech can really be a powerful tool, said Rogers.

It can also take some of the pressure off family members and caregivers, he explained. When a smart device is tracking things like pain level and mobility, medication adherence, blood pressure movements and other stats, it frees up family members to connect with their loved one instead of spending the whole visit asking about the illness. So, in a roundabout way, it’s a supporter not just of the patient’s physical health, but of his/her mental and emotional wellbeing (and those of the people around him/her) as well.

Then, there are those who are really having fun with voice tech – like Alegeus, which taught its voice assistant Emma to tell wry jokes between explaining how consumers’ health plans work, and like Catalia Health’s Mabu, a physical bot that blinks and builds relationships with the patients it serves. Mabu uses artificial intelligence blended with psychology and medical expertise to help patients with ongoing illnesses navigate their conditions.

Home healthcare workers have to visit their patients monthly, weekly or even daily, but with help from a voice-activated assistant like Mabu, they could (and do) provide more focused care to a greater number of people, with the assistant handling some of the routine functions like talking the patient through taking their medicine or completing physical therapy exercises.

Doctors can’t stand over patients at all hours to ensure these things are done, and done correctly, said founder and CEO Dr. Cory Kidd. Augmenting professional care with an AI-powered robot can fuel better treatment compliance while also saving healthcare systems from having to hire more employees, or increase the hours and salaries of existing ones.

Where Voice Tech Falls Short

“Voice assistants will never fully replace human interface,” Rogers said. Flat out. At the same time, though, “voice is making rapid progress as the third wave of the internet, with websites coming first, then mobile apps. I fully expect the application and opportunities with voice to be bigger than those.”

When it comes to handling phone calls, there’s no denying that automation can make the process faster and more efficient in some cases.

However, if the customer wants to know what to do after the diagnosis, why his copay is so high, or anything else of a complex and sensitive nature — Cogito Chief of Customer Success Douglas Kim calls them “moment of truth interactions” — then a human, not a robot, still needs to address that topic. That’s not a chatbot exchange; that’s a phone call. And that goes for all generations, said Kim — millennials as much as boomers and retirees.

In terms of an at-home voice assistant like Mabu, Dr. Kidd said, “These technologies are not replacements for human care. They are not designed to do everything patient care requires; they are simply focused on keeping patients on the therapies recommended by doctors. Voice-activated technologies cannot have open-ended conversations.”

Finally, across the board, the Achilles heel of voice tech is accuracy. Quality is important, said Kim, but if a patient asks the voice-activated assistant to find the closest MRI center and the top three results are closed, then the patient loses trust in the platform — and loses it fast. It doesn’t take much, Kim said, for people to give up on a bot that can’t seem to do its job.

What Needs to Happen Next

Dr. Kidd notes that voice technology has come a long way but still has a long way to go. Better voice-to-text comprehension is possible — and, indeed, necessary if this tool is going to be of use in the healthcare sector. Getting the accuracy where it needs to be (in terms of both comprehension and data) will be no small feat, but according to Prugar, the greatest threat in this brave new world is going to be security.

“Fraudsters love the phone channel,” said Prugar.

Fraudsters can spoof numbers and social engineer the call to access sensitive medical records and other personal identifying data that they can use for account takeover, identity theft or synthetic identity construction. To combat this, Prugar said organizations will need to implement real-time authentication like anti-spoofing technology, dynamic blacklists and voice biometrics.

The final component of voice tech acceptance is going to be hardest to control, and that’s the people.

“Across the industry, many are putting out chatbots and Alexa interfaces, but call volume is spiking 15 to 20 percent across the board,” said Kim. “You cannot, as a healthcare provider or payer, be in a position of dictating to customers how they should interact with you. You need to be ready to receive your customer service requests across all the different channels that they demand.”

In other words, if people don’t want it, they don’t want it. The tech has to be good enough that consumers want to use it.

Rogers agreed. “Consumer acceptance is significant,” he said. “Think about a time when people said, ‘I’ll never need a website,’ or ‘My flip phone is fine. Why would I need a smart phone?’ Adoption will come, and with it, the need for creating voice skills.”

The final hurdle? “We don’t know how it’s going to be regulated by the FDA,” noted Dr. Kidd. “The FDA has recently created a new unit focused on digital health, but it will be a few years before we know how it shapes up for voice activation specifically, since all FDA regulations pass through Congress.”

Even if people love it, there’s no guarantee that Alexa will narrate our future medical experiences.