2283_2018_AI_in_Healthcare_Research_blog_image_1200x750-830157-editedPratap Khedkar co-wrote this blog post, the second in a series on ZS’s 2018 AI in Healthcare study, with Paul Darling.

We trust artificial intelligence to navigate commuters around real-time traffic and pitch in when it’s time to cook dinner, and it’s even being called on to tackle the ever-looming laundry pile (for those who can afford the hefty price tag). AI also is well on its way to autonomously steering cars and landing airplanesalthough most consumers don’t feel comfortable completely foregoing human judgment when people’s lives are at stake.

We’ve found that that pretty much sums up how patients feel about AI’s potential role in healthcare delivery. They’re OK with AI taking on certain tasks, but they want the diagnoses and treatment decisions left in their physicians’ hands.

In an earlier post, we discussed how physicians have reservations about handing over patient-related decisions to a machine. ZS’s 2018 AI in Healthcare study found that 66% of doctors approve of using AI to help with workflow management, administrative assistant tasks and patient experience analysis. However, most physicians aren’t entirely comfortable with letting AI play a role in actual care delivery. 

Just as physicians aren’t prepared to have an AI “colleague,” patients aren’t so sure about handing over care delivery to a robot. But there’s evidence that they’re optimistic about the interpersonal gains that could come from AI-enabled healthcare: More than half of the patients we surveyed think that AI will improve the physician-patient relationship by giving doctors more time to devote to discussing and addressing patients’ health concerns and questions.

Contrary to what you might suspect, our survey found that older consumers are the potential “early adopters” of these AI-enabled physician visits. Nearly three-quarters of respondents who are 65 years old and older selected “undivided attention” as the main driver for patients’ approval of AI playing a role in the healthcare setting, compared with just 41% of those in the 18-to-24 age bracket. This likely is due to the time that older patients tend to spend in doctors’ offices compared with younger patients. And older patients’ experience in the pre-AI healthcare realm could be another influencing factor here: They’ve witnessed the rise of EHRs and other distractions that occupy a growing proportion of physicians’ one-on-one time with patients. Could AI bring back the “Norman Rockwell-esque” bedside manner that these patients remember?

Serving as further proof of a generational divide, our study found that respondents in the 18-to-24 group are more interested in AI’s ability to increase overall efficiency (70%) and decrease out-of-pocket costs (50%). Despite the variance in where AI can best contribute to healthcare, patients clearly are open to the technology, but the question of exactly how AI will enter healthcare remains.

When considering where AI fits into the picture, patients agree with physicians that, for now, the technology is best equipped to tackle administrative tasks. In fact, 73% of patients think that AI can best contribute by assisting the doctor with hands-free note taking and looking up information during healthcare encounters. But they’ve also left the door open for AI to improve upon some of the aspects of healthcare that aren’t up to snuff.

Patients Prefer ‘Assist and Advise’ Over Autonomous Diagnosis

As far as the technology goes, AI is positioned to play a prominent role at the point of care. There’s AI-powered software that can analyze a patient’s medical history and demographics, images, EHRs (yours and other patients) and tests to determine whether illness or disease is present. There’s AI-enabled software that analyzes a patient’s genetic information to determine the likelihood of getting a certain disease, and to pinpoint the gene mutations behind certain types of cancer. There are AI-enabled solutions that help doctors determine the best therapies and care pathways for any given patient. There are even AI-enabled patient advocacy tools such as software that helps care managers guide patients through the healthcare system by connecting patients to financial and social support, and coordinating between physicians, insurance companies and employers.

AI is equipped to process data at a speed and accuracy that humans can’t compete with, but patients and physicians have drawn a hard line between using tools like AI to collect useful health information and relying on a machine to make actual healthcare decisions. Our research shows that both patients and physicians want AI to be relegated to a support role in the exam room, assisting or advising on diagnoses without being entrusted to make the diagnoses on its own.

And once again, older patients appear to be more willing to trust AI’s abilities: Just 20% of respondents in the over-65 bracket are concerned about AI’s ability to diagnose accurately, compared with 40% of the 18- to 24-year-old patients we surveyed. This is an interesting twist of results considering the propensity of younger consumers to entrust so many of their daily living needs to AI-powered technology, from splitting a dinner check with a friend through an app to ordering a pizza via a chatbot.

The Opportunities and Challenges Presented by Digital Health and Data Privacy

While patients might be hesitant to invite AI into the exam room, they’re more than willing to leverage AI in their overall health and wellness efforts, which could be a solid stepping stone to a more automated healthcare future. Our survey found that 45% of patients use wellness apps and 34% have adopted wearable technologies. And, in a nod to the consumer convenience craze, 45% of patients use telehealth services in their regular care.

But beyond the need for human understanding and judgment, patients also list privacy as a significant hurdle for AI to overcome in healthcare. In our survey, 53% of patients attributed their hesitation about AI to data and privacy concerns. And those under 50 are two times more concerned than older patients (the 50-plus crowd) about the implications of data privacy in using smart EHR platforms.

Typically, research has shown that younger consumers are freer with their data than older consumers, so the difference here could be because older patients are more accepting because they’ve had more frequent healthcare issues and therefore are more willing to share data that could improve their situation, but we didn’t explore that theory in our study.

Given consumers’ acceptance of AI in many other aspects of their daily lives, the tech’s indoctrination into the medical mainstream is moving at a slow pace. But if we’re reading the tea leaves, AI will get there. Healthcare delivery is becoming more technologically advanced and more data-driven every year, making the analytical firepower provided by artificial intelligence a natural next step—at some point. But as in other industries, AI isn’t likely to completely supplant humans. It will augment and enhance people’s abilities, automating administrative tasks to allow physicians to spend more time with patients, and driving more accurate diagnoses and treatment decisions by synthesizing and serving up the necessary data and insights. In a sense, we’re setting a benchmark with this study of where healthcare stakeholders’ appetite for AI currently stands. More to come on that front in our next post.   

Stay tuned for more insights from our 2018 AI in Healthcare study, including the ways that institutions can drive physician adoption of AI.


BLOG POST: How Do Doctors Feel About Robots in the Exam Room?

BLOG POST: Is Pharma Ready to Move From Hype to Hope?

BLOG POST: The State of AI Adoption in Pharma


Topics: Patient Engagement, Paul Darling, artificial intelligence, data privacy, care delivery, healthcare administrative tasks, AI in Healthcare study, automated healthcare