Decentralized clinical trial design

Google and Microsoft, rivals in cloud computing, have turned their attention to healthcare as they look to win over hospital systems as customers. Google has struck long-running partnerships with insurer Highmark, where it plans to build tools to help patients to share health information between visits, and Mayo Clinic, where it is tasked with developing a suite of AI solutions. Microsoft, in the meantime, made one of its largest acquisitions to date, with its $19.7 billion purchase of clinical documentation company Nuance.

At MedCity INVEST Digital Health, leaders from both companies talked about their interoperability efforts and the future of AI in healthcare. Both rivals agreed on one thing: that these solutions should serve as a support, not a replacement, for clinicians.

Before he started practicing as a neuroradiologist, Microsoft Health Corporate Vice President Dr. Greg Moore was an engineer. In the early days, like many entrepreneurs, he wanted to build something that could make a diagnosis better than a doctor. But now, in his work at Microsoft, he’s more focused on using machine learning to help alleviate burnout by helping document encounters and augment physicians’ work — an area that Microsoft surely will focus on with its acquisition of Nuance.

“Healthcare is an intensely human endeavor. It’s an endeavor between the provider, a doctor or a nurse, and the patient,” he said. “I think it should remain that way.”

Where he could potentially see AI entering the clinic is to help pick up patterns in test results, for example, or in his field of radiology, where several cleared tools are already being used for clinical support.

Similarly, Google Cloud Global Director of Healthcare Solutions, Aashima Gupta, said technology should be used to aid doctors and nurses, many of whom were already overloaded with work before the pandemic. But she also cautioned against simply throwing more data at clinicians.

“Physicians don’t want more data,” she said, referring to her previous work leading digital health at Kaiser Permanente. “Making sense of that data variety, data volume, crunching it together and giving that insight, that’s what is more meaningful and where I believe AI/ML can help with pattern recognition.”

Interoperability must be patient-centric
Looking to the future, both she and Moore spoke with excitement about the next wave of innovation as health systems have invested in their digital infrastructure and interoperability regulations are expected to free up more health data. But record-sharing challenges persist today, and have been a major obstacle for public health agencies and patients during the pandemic.

Gupta spoke of her own experiences with her mother, who previously had Covid-19.

“It was pretty devastating,” she said. “She’s recovering from it, she’s on the mend.”

When her mom was discharged home from the hospital, she saw the interoperability challenges firsthand.

“Yes we had telehealth, virtual care, I had her pulmonologist and physical therapist arranged at home, but when her oxygen started dropping, there’s no link between the digital and the physical world,” she said. “I saw that firsthand as a very worried daughter. How do we make this care transition, where care is tied to people, not to places?”

Both now and after the pandemic, she hopes that interoperability will be treated as more than just as a regulatory mandate.

“It’s really for the patient, for the caregivers, and to help empower them to take care of their own health,” she said.

Photo credit: AnuStudio, Getty Images