Conversational healthcare: AI with a human face
For an industry that specializes in diagnostics, it’s not surprising that those in the healthcare sector know the challenges they face. But in search of a solution, we can’t afford to let basic standards of care slip. Healthcare can’t become robotic, functional and transactional at the loss of personal care and interaction.
That’s not to say digital humans can perform every role a human currently does. Some roles you might see a conversational AI-powered digital human perform include:
- Digital wayfinding: Triaging patients based on their requirements, helping them to find the level of care they need in a digital world.
- Concierge: Welcoming and helping patients find their way around healthcare facilities, and manage the check-out process.
- Health education: Simply being available to answer the most common healthcare questions, so people have a way to learn about their physical and mental health, and live healthier lives.
- Insurance form specialist: Specifically with medicare in the US, digital humans have been proven to increase accuracy and completion of complex medical forms, including those for medical intake and insurance. In fact, digital humans performed 2X better than other digital means.
- Personal assistant: Tracking and managing appointments by integrating with existing software and back-end systems.
- Follow-up support: 30% of patients say they don’t understand their discharge instructions and have questions about their prescriptions. Digital humans can aid with post-treatment and rehabilitation by answering questions around prescribed medicines or by providing dietary tips.
- Life planner: Helping individuals manage major life changes, such as retirement, by answering questions and helping to plan for the future.
- Companion: Being a friendly face people can turn to for an open-ended conversation, particularly those suffering from loneliness, like some in aged care.
Helping to create the new healthcare digital workforce
A quickly aging population is being compounded by a gap in the number of healthcare workers we have and how many we need to have in the very near future. According to Healthcareers, some 11.6 million workers will be needed in the United States alone by 2026 to meet the demand for new clinicians.
Accenture analysis says AI solutions can plug around 20% of this gap. Conversational AI is likely to make up a majority of this patient-facing workforce, as chatbots, virtual assistants and digital humans step to the fore.
That’s not to say Siri, Alexa or even Sophie – one of our own healthcare digital humans – will be diagnosing patient conditions and prescribing medicine. The value of these virtual technologies is that they can handle more “everyday” tasks, allowing healthcare professionals more time to do these parts of their jobs you wouldn’t want to automate.
Let’s not forget the human impact of this. A 2018 study from BMC Health Services shows that 26% of healthcare workers (even before the pandemic) reported suffering from burn out. Almost one-third said they’re intending to leave their jobs within the next two years.
And lastly, there’s a financial imperative to looking at digital healthcare workforces, too. The WHO estimates that workforce inefficiencies cost health systems roughly US$500 billion every year. Accenture, meanwhile, found that AI applications can potentially save the US healthcare economy around $150 billion annually.
We’re not going to pretend that’s the potential of digital human technology. AI is helping clinicians in things like surgery and diagnostics, not just patient care. But we believe humanizing some of the patient-facing conversational AI technologies will mean these investments won’t strip away too much of the face-to-face experiences patients need to feel cared for and supported.
Improving healthcare accessibility with digital humans
As well as the growing employment issues, there are long stood challenges in the healthcare system that conversational AI is naturally able to help with – or even solve.
The first and potentially most pressing is around access to healthcare. For one, not everyone can use a keyboard or a computer in general, leaving them inherently locked out of most digital healthcare.
Digital humans interact with patients using either speech or text (or a combination of the two. For instance, those who can’t type can use natural face-to-face conversation to find support and help. Digital human user interfaces can also include elements like subtitles and visuals to help people understand better – if they’re more visual learners, for example.
Improving access to healthcare among those who usually struggle to get through the digital front door became the inception for digital humans. And accessibility was the earliest use case digital humans were given to solve. In fact, it was our work on Nadia for Australia’s National Disability Insurance Scheme (NDIS) in 2017 that led to our company effectively being born.
Another contributor to issues of healthcare accessibility is patients receiving support in the language they understand best. Digital humans today can speak over 40 languages and hundreds of dialects. While it’s unfeasible to expect this level of multilingual coverage from real people, one digital human can be taught a new language incredibly simply.
Taking away the fear of judgement
Health illiteracy is an incredibly widespread problem. Realistically, billions of people worldwide – even those in countries with world-class healthcare – may not receive the care they need because they don’t know they need help.
Perhaps the most damaging and compounding issue is that those who have poor health literacy are some of the least likely to seek help from a professional, too – because of a fear of judgement.
We naturally fear what we don’t know, so can be too concerned to seek help. Many people are concerned about wasting a health professional’s time, or decide their issue is too trivial, choosing instead to self-diagnose (and mis-diagnose) using online search engines and the like. The fear of judgement is perhaps most common when it’s in regards to delicate healthcare matters, like mental health and physical ailments patients may find embarrassing.
Again, some of the most vulnerable parts of society are some of the least literate when it comes to their health. Some of the key contributors to health illiteracy include:
- Poor access to education
- Limited proficiency in the English language
- Cultural barriers
- The use of medical jargon,
- And disabilities such as cognitive decline and learning difficulties.
Digital humans in particular have had great, proven success in this field. One of the pioneers of this was Ellie, who was created by researchers at the University of Southern California to help provide judgement-free support to veterans returning home from active duty. Ellie’s very nature as “non-human” (and clearly non-human to those she was treating) made them more likely to disclose information to her that would help clinicians identify signs of post-traumatic stress disorder (PTSD).
Our own health advisor, Sophie, is designed to have a similar impact around COVID-19 during the early days of the 2020 pandemic. With so much information and opinion floating around the web, Sophie was plugged in to trustworthy sources of information, like the CDC and WHO. People asked her questions they may feel too embarrassed to not know, like “what is coronavirus”, and then receive guidance on how to stay safe.
The healthcare burden of loneliness
The last issue we would point out as a particular strength among digital humans is the ability to combat the issue of loneliness.
It’s hard to quantify exactly how detrimental loneliness can be on people’s health. One study shows loneliness among older people has the same negative impact as smoking 15 cigarettes a day. And it’s a cyclical problem: loneliness can cause significant health issues, which a lack of personal care can compound into a greater feeling of loneliness, and therefore greater health issues.
Not only does it affect health outcomes – both physically and mentally – it puts undue strain on the medical industry. Three in four GPs say they see as many as five patients a day who are simply suffering from loneliness. One in every 10 GPs say they see between six and 10 of those suffering from loneliness.
Financially, Harvard research estimates that this lack of social contacts among older adults costs Medicare $6.7 billion a year – mostly on hospitalizations and nursing for those suffering.
While there are deep-seated issues that cause and compound loneliness, and no technology can hope to put an end to it, the sophistication of conversational AI is helping somewhat. With virtual assistants, people have someone to speak to. In chatbots like Woebot, they have a text interface to chat to should they wish. In digital humans, people suffering from loneliness can have a conversation with an AI that can show warmth, empathy and friendliness in its speech and expressions – as real people do.
Australasian mental health app Mentemia uses a UneeQ digital human as a sleep coach to its users. People get a guided sleep plan, administered through conversation. And what’s more, the digital human is a lifelike virtual twin of New Zealand rugby icon Sir John Kirwan, whom people can speak to like they would the real JK about his playing career and more.