What Patients Really Want from AI in Healthcare

Lately, when people ask me about artificial intelligence and healthcare, I tend to talk to them about cars.

This isn’t just because both self-driving cars and robot doctors tick the “Jetsons future” box beloved of media coverage around AI, it’s also because both of these still seem a very long way removed from reality, and for a very similar set of reasons.

In both cases the technology, ethics, regulation, and legality are problematic, and lives—whether of patients, drivers, passengers, or pedestrians—are obviously too valuable to risk. Neither of these are industries in which Silicon Valley should be encouraged to “move fast and break things.”

What we have seen with driving, however, is technology delivering a series of incremental benefits over time. Cruise control, for example. Or automatic parking. And obviously GPS and route-finding. Tools that don’t replace drivers, but rather which let them keep more focus on the road ahead.

Asking People About Healthcare and AI

To try and get a sense of how a similar path might play out for healthcare, Syneos HealthTM surveyed approximately 800 patients and 200 carers across the U.S. and Europe to understand their hopes, fears, concerns, and ideas around artificial intelligence.

Importantly, these were older people—not a young and relatively healthy “generation now” audience, but rather older people with chronic and costly conditions such as type 2 diabetes and atrial fibrillation: Healthcare’s “Generation Right Now.”

We wanted to understand what “cruise control” might look like for the person with co-morbidities and a polypharmacy regimen. Or what “route-finding” could look like for someone adjusting to a new course of treatment.

The first thing we found was that the future of cars and the future of health have one more thing in common—that people seem happy with what they’ve already got. Despite the enthusiasm of Silicon Valley and the investor community, fewer than one in five U.S. patients sees any benefit to their healthcare in receiving a diagnosis or a treatment recommendation via AI.

This matches the 80% of U.S. drivers who told a Kelley Blue Book poll in 2016 that when it comes to the future of self-driving cars they’d always like to have the option to drive themselves.

Switch the conversation, however, from patients’ interactions with their doctors to their interactions with wider healthcare teams and services and a different picture begins to emerge.

AI Services Patients Would Actually Use

For example, 71% of U.S. patients would be comfortable interacting with an AI receptionist for booking an appointment. And beyond administrative tasks, 64% would be equally comfortable with a “virtual nursing assistant” providing ongoing support and monitoring around either general health and wellness or a specific prescription product.

Our survey also points to what form these interactions might take. When asked which of a series of “human” characteristics they would find most appealing in a virtual nursing assistant, the strongest preference by far was for a realistic voice. This was judged to be more important than “robotic” characteristics such as a physical presence or a face.

This preference for voice is significant for two reasons.

Firstly, an increasing body of evidence shows that conversation/dialogue between people and virtual digital coaches is capable of building empathetic relationships and healthy outcomes.

(For anyone interested in diving in further to this topic, Google can quickly point you to the data on the Woebot Facebook Messenger service, and the Lark app for weight-loss and pre-diabetes).

Secondly, there’s the phenomenal rise of Amazon’s Alexa, which experienced record sales over last year’s Black Friday weekend and holiday periods. And while Alexa’s primary usage seems to be playing music or reading the news, the opportunity it presents for patient support is highlighted by new data from DRG Digital – Manhattan Research suggesting that more than half of U.S. online adults are interested in using such AI-enabled voice assistants for health.

It also seems more than likely that Alexa will play a starring role in Amazon’s recently announced healthcare venture with Berkshire Hathaway and JPMorgan Chase, which could be a watershed moment for voice and health.

But this doesn’t necessarily mean it’s all good news for Amazon.

Overcoming Patients’ Lack of Trust

In announcing their partnership, Warren Buffett of Berkshire Hathaway described healthcare costs as “a hungry tapeworm on the American economy.” But much of that cost comes from “Generation Right Now” who our survey suggests are unlikely to trust big corporations with their health data.

In our survey, fewer than one in five older U.S. patients would trust either a technology company or a pharmaceutical company to provide AI-powered virtual health assistants. They would rather these were provided by established healthcare stakeholders—including doctors, hospitals, and pharmacists.

The swiftest route, then, to realizing the benefits of AI for patient support and engagement looks likely to be one of smart partnership and collaboration between all these stakeholders.

Technology companies have the necessary smarts, but they don’t necessarily have patients’ trust when it comes to health. Doctors have this trust but aren’t as close to the technology; and life sciences companies are the experts in their products and have strong relationships with groups of doctors, but our data suggests they’re the least trusted of all.

To best support patients on their products, life sciences brands will need to create new advisory boards and working groups that engage physicians at the highest levels. Without the confidence, comfort, endorsement, and recommendation of physicians it seems unlikely that an older audience with chronic conditions will take up new digital tools—however good the technology. Those brands will also forge new partnerships with proven technology leaders that own the platforms and devices consumers and patients already trust.

More than anything else, these partnerships should also involve nurses themselves. After all, we all know what the health equivalents of cruise control and GPS look like for patients with chronic conditions. They look like nurses.

And while Alexa could never replace any nurse, if any discussions with this wider group of stakeholders can help extend and scale the type of support that nurses provide then these are conversations well worth having.

  • Duncan Arbour

    Duncan Arbour is SVP Innovation at Syneos Health Communications in Europe. He’s worked with digital for the last 20 years, in healthcare for the last 10, and for Syneos Health since 2012.


You May Also Like

businessman holding a virtual AI chatbot in hand

5 Human-Generated Guidelines for Using Generative AI

Generative AI’s potential for tsunami-like disruption has sounded the alarm for knowledge workers worldwide. ...

What Do Facebook Ad Changes Mean for Medical Advertising?

Facebook is currently weathering crises on multiple fronts. Investor confidence is declining, user patience ...