“Alexa, what’s my blood sugar level?” Last week, Amazon debuted new skills that allow the Echo Dot to store blood sugar data from internet-connected monitors, schedule appointments, and relay messages from doctors.
As people become comfortable interacting with artificial intelligence as part of their daily routines, healthcare companies are turning to AI as a way to provide higher-quality, more efficient care. The consulting firm Accenture estimates that healthcare AI could be a $6.6 billion industry by 2021.
CEOs discuss AI in healthcare at the 2019 World Medical Innovation Forum
“We have a hundred apps sitting right now [on our] platform, and a lot of that is directed toward improved workflow…for our clinician[s], [simplifying] their lives,” said Kieran Murphy, the CEO of GE Healthcare. “[AI can] ensure that clinicians spend more time with the patient, rather than trying to figure out, ‘What is the data? What does this image mean?'”
Murphy shared his insights during a panel of CEOs and R&D leaders at the World Medical Innovation Forum earlier this month. Held annually in Boston, the event brought together more than 1,700 healthcare professionals to explore the challenges of applying AI in healthcare.
During the discussion, leaders from Bayer, GE Healthcare, Philips, and Siemens Healthineers discussed how their companies approach AI: from setting a strategy through adopting new technology. Read highlights from the panel below.
Balancing Efficiency and Innovative Uses OF AI
AI has the potential to increase productivity and help healthcare companies dramatically decrease their operating cost. Accenture reports that AI applications in the industry could save the US healthcare economy $150 billion dollars annually by 2026.
However, Joerg Moeller, the EVP of Pharmaceutical R&D at the German Pharmaceutical company Bayer AG, cautions against focusing solely on efficiency-driven AI.
“I see two big categories when it comes to the application of AI… One category is what I consider efficiency drivers, and to me that’s kind of homework — [it] has to be done. But it’s typically not what gives you a sustainable competitive advantage,” Moeller says. “The other category is the more disruptive approach. It’s where, with the use of AI, you gain insights that with traditional means wouldn’t be possible or [would be] very difficult to get.”
According to Moeller, certain innovations can combine those two categories. As an example, he points to AI applications at Bayer that help the company develop new drugs.
“[W]e are using AI to [pick] the right chemical reaction…[find] the optimal temperature and optimal solvent. And we’ve made great improvements both in terms of productivity as well as our ability to come out with very innovative [new drugs],” he says. “And that’s something I would [put] in actually both categories.”
Creating a Platform (Not Just Apps)
During the panel, GE Healthcare CEO Kieran Murphy also discussed the importance of building the right platform to host different AI applications. “[Apps] need to have a home [where they] can reside,” Murphy said. “And that’s what we’re trying to invest in.”
While AI applications are designed to collect data and analyze it for a certain task, AI platforms have the ability to synthesize and find patterns across applications.
“We’re [enabling] radiologists and clinicians to develop their own AI,” he says. “It is our responsibility to put together a platform, and that those platforms have to have separate layers, and [that] they…protect the patients’ data, be cyber secure, and we need to work very well with the ecosystem [of hospitals, clinicians, and healthcare providers].”
Making Data More Meaningful
According to Philips CEO Frans Van Houten, Philips focuses on the role data and AI can play in improving the patient’s journey through a healthcare episode. “[W]e look at how patient journeys are designed…[and] how particular care providers and actors play in that ecosystem,” Van Houten says. “[Then, we see] how data can be contextual [and] made available in order to get to the right physicians, get to the most efficient workflows, [and to] avoid mistakes and enhance the experience.”
The imaging and medical device company invests over $2 billion annually in research and development, which includes projects that seek to apply AI. According to Van Houten, the company also employs over 600 data scientists.
However, Van Houten emphasizes that having more data isn’t necessarily the solution to improving the quality of care. He points to one hospital in Canada where care providers said that they now “drown in data” from electronic medical records.
“So apparently just having a repository of data is not good enough,” Van Houten said. “We need to make it specific and actionable, and by having this architecture we can make data more operable.”
Adoption: Using AI to Reimagine Human Roles
Siemens Healthineers, a healthcare technology company best known for digital imaging tools, envisions a future where healthcare AI can assemble different types of medical data into “a digital twin of the patient,” according to CEO Bernd Montag. Then, he says artificial intelligence can scan that digital twin and aid physicians in making diagnoses.
“Imaging is at the core of the digital transformation of healthcare because, it’s generating digital assets,” Montag says. “We have a very clear roadmap to start with automating the products…to make the product more intelligent, and in the next step, help work with companions to read images.”
Montag points to radiology as an example. According to a 2015 study from the journal Academic Radiology, radiologists must look at one image every three to four seconds to meet workload demands. AI could be a pathway to assist radiologists in that work.
However, Montag says that clinicians would need to redefine their roles in order to adopt new technology. For example, Montag says a radiologist would need to view themselves not only as a doctor, but also as a technologist and information specialist.
“[AI] changes how people look at their own roles. It elevates physicians to something different. It changes [Siemens’] role from an imaging and diagnosis company to a decision company and so on,” Montag says. “I think what we need to be aware of is that it’s not just AI adjusting to what we have been used to doing. It is us and the whole ecosystem of players adjusting to the new possibilities and looking [at the field] in new ways…”
