Next-gen telehealth: AI, chatbots, genomics and sensors that advance population health


Original post, click here

While the use of telemedicine systems has been expanding in recent years, especially as more payers, including the government, have begun reimbursing for telehealth services in some instances, the industry is on the verge of more widespread virtual care.

But what will that ultimately look like? The next generation of tools will feature enhancements ranging from chatbots to machine learning to genomics to remote diagnostic tools and better sensors, among others.

Here’s a look at what to expect in the near future. 

Machine learning and chatbots effective, but only in specific cases

Both machine learning and automation are trying to solve an inherent issue in virtual healthcare: scalability, said Roeen Roashan, senior analyst of digital health at consulting firm IHS.

“Virtual health technology is not scalable by nature, and this has in fact been a major limitation for more than two decades,” Roashan said. “In remote patient monitoring, you still need a rather significant workforce to support the value chain. Machine learning is making its way to remote patient monitoring. Companies like PhysIQ and Sentrian offer solutions that make better use of clinical workforce when monitoring patients remotely by automatically tracking whether a baseline for the individual patient is improving or not. This improves scalability by a factor of 7x or 8x.”

“Machine learning is making its way to remote patient monitoring.”

Roeen Roashan, IHS

To truly reach scalability, certain functions of virtual healthcare must be decomposed and outsourced back to the patient, he added. Therefore, one needs automation, and systems that are led by the patient, but where the value is driven by, for example, chatbots, he said.

“In certain cases, these chatbots create higher relative value than human personnel,” Roashan said. “For instance, in one study focusing on building mental resilience among youth, a chatbot delivered stunning results. Users who chatted with Wysa had a 45 percent reduction in depression, and medication adherence rose by 10x.”

This is not to say that mental health is better delivered through chatbots but, rather, that certain aspects of care should be automated. These mental health patients actually preferred to speak with a bot, and most conversations took place between 10 p.m. and 3 a.m., often including taboo topics, Roashan explained.

“Chatbots can deliver value, but only in some parts of the healthcare value chain,” he said. “The next big step for these automated services is to include voice, and there are plenty of companies working with Alexa to do exactly this.”

AI will be transformative for healthcare

As Roashan believes machine learning has a role to play in the next generation of telemedicine technology, Joel Barthelemy, founder and CEO of GlobalMed, a telemedicine technology vendor, added that artificial intelligence can dramatically improve telehealth.

“AI will transform primary care via telemedicine,” he contended. “The rise of machine learning will enable ‘AI engines’ to take over much of the care now provided by primary care doctors. Represented by a human-like avatar, an AI engine will remotely lead patients through a sequence of questions similar to – but more comprehensive than – the ones that primary care physicians and nurses now ask about symptoms and health histories in office visits.”

“With the advent of genomic medicine, telemedicine will be used to provide much more personalized care to people based on their genetic characteristics. “

Joel Barthelemy, GlobalMed

The AI engine will tailor these questions to what it already knows about the patients and what their answers reveal, he added. It will not ask questions that other providers have already asked because the answers will be online in the patient’s personal health record, he explained.

“The AI engine will also take advantage of video conferencing to inspect patients’ skin and observe their general health condition,” Barthelemy said. “It also will consider the results of home and mobile vital signs monitoring on inexpensive devices and wearable sensors. A ‘lab on a chip’ might even be used for instant blood tests, the results of which could be transmitted to the AI engine.”

Currently, the U.S. Food and Drug Administration does not allow a device to diagnose a patient without the input of a physician. But an AI engine could say there’s an 84 percent chance that a patient’s stomach symptoms are related to the flu and that there’s a 4 percent chance they are due to indigestion. A doctor could receive this report and decide whether or not he or she agrees, Barthelemy said.

“An AI engine also is not allowed to prescribe drugs, and that is unlikely to change in the foreseeable future,” he said. “But, based on the experience of many thousands of similar patients with the same symptoms and signs, the AI engine could advise the doctor on which medication to prescribe or which treatment to try. As these learning machines become more and more accurate, doctors will agree with them most of the time unless they know something unique about a patient.”

In that case, they might contact the patient and ask some more questions; but in most cases, patients will no longer have to see primary care doctors in person, he contended.

New sensors will show up in more places

On another front, sensor technologies built into households and consumer electronics that can trigger consumption of as-needed medical services may be about to change population health forever, said Roy Schoenberg, MD, CEO and co-founder of American Well, a telemedicine technology vendor.

“This real-time loop that combines sensors, analytics and telemedicine marks the transition from passive to proactive healthcare.”

Roy Schoenberg, MD, American Well

“We see this with Apple’s new heart study with Stanford Medicine, where Apple Watch detects cardiac issues and magically prompts a telehealth doctor to show up on the patient’s phone to initiate a medical intervention within minutes,” he said. “This kind of real-time loop – that combines sensors, analytics and telemedicine – in my mind marks the transition from passive healthcare, which engages with the patient’s ‘chief complaint,’ to proactive healthcare, which dispatches live care as problems are detected.”

The implications will extend not only to how people experience care, but also to the practice of medicine, healthcare cost, healthcare policy, and quite likely how people age and longevity, he added.

Genomics, diagnostics and remote monitoring

Another next-generation feature of telemedicine will be genomic therapy, Barthelemy said.

“Today, telemedicine is used to consult with specialists who may be thousands of miles away from the patient,” he said. “With the advent of genomic medicine, telemedicine will be used to provide much more personalized care to people, based on their genetic characteristics. This will require patients to go to nearby medical centers that are equipped with low-cost genomic sequencing equipment.”

Based on a patient’s genetic makeup, and what is known about how patients like them respond to various therapies, a group of geographically scattered doctors who are being advised by AI engines will use telemedicine to confer and quickly decide how best to diagnose and treat that patient, he explained.

And a new breed of remote telemonitoring and diagnostic tools that allow patients and physicians to interact in a way that closely mimics an in-person experience will highlight the next generation of telehealth, said Cory Costley, chief product officer and co-founder at Avizia, a telemedicine technology vendor.

“Today, there are a handful of telehealth platforms equipped with state-of-the-art functionalities, such as enhanced far-end camera control and real-time data from connected peripherals to empower a remote physician to examine a patient’s physical appearance and capture images or sounds of the body,” Costley said.

Five years ago, some of these diagnostic examinations weren’t possible at such a high level, he said.

“Tools like these truly bridge the gap between what you can do remotely and what you can’t do remotely,” Costley said. “The technology is very timely, too, as growing physician shortages in many parts of the country have exacerbated healthcare access problems.”

Twitter: @SiwickiHealthIT
Email the writer: bill.siwicki@himssmedia.com