HIMSS19 Roundup—New AI tools to address physician burnout; Sharing an Uber in a crowded town

Providence St. Joseph Health using AI tools to reduce physician burnout

Providence St. Joseph Health, the Renton, Washington-based health system operating across seven states, is leveraging artificial intelligence to alleviate physician burnout by reducing the time doctors spend documenting in the EHR.

Speaking at the 2019 Healthcare Information and Management Systems Society (HIMSS) Annual Conference and Exhibition on Monday, Aaron Martin, the health system’s chief digital officer, said the health system is piloting an AI-powered virtual physician assistant platform called Saykara. Physicians who use the tool spend less time charting at the end of the day from home and report having more time to see more patients.

The health system also developed a consumer-facing, AI-powered chatbot, called Grace, designed to help navigate patients to the right site of care and improve efficiency when addressing low-acuity health issues. When clicked, the chatbot has 90% accuracy at routing patients to the right venue of care, Martin, a former Amazon executive, said.

The inspiration for Grace came from outside healthcare: Air New Zealand’s chatbot, Oscar, which answers 75% of travel questions— Heather Landi

Sharing in an Uber in a crowded town

Among the many conference-goers that have traveled to Orlando this week, I had an interesting chance encounter with Charles Herman, M.D., the chief medical officer of a company called Somatix.

We ended up sharing an Uber from our hotel to the conference—you can imagine the logistical nightmare of transporting tens of thousands of people between their hotels and the conference itself—and chatted during the 20-minute trip about the conference and the technology he’s working to sell.

When you hear about smartwatches being used to help monitor elderly patients, Herman is one of the guys trying to make that happen. Specifically, Herman's company says they can offer remote health monitoring through "the flick of a wrist." Hand gestures, as detected through smartwatches or a $50 smart wristband, can actually tell health professionals a lot more than a patient's heart rhythms, he said.

For instance, it can detect whether a person is picking up their medication to take it. It can detect changes in a person's gait to determine whether they are at increased fall risk. The Somatix system can even detect remotely from how a person moves their arm whether they are sipping a hot drink or a glass of cold water, he said.

They're marketing their product to elder care companies that often spend tens of thousands of dollars to set up complex video-based monitoring equipment that is limited to individual patients' rooms. They raised a $6 million series A funding round in 2017 and are in the midst of raising at least $3 million more, an SEC filing shows.

Herman said the tech could help reduce manpower and offer providers insights much earlier. He expects Somatix will eventually market them to patients sent home from the hospital who could be connected with health coaches through the system to help monitor risks.

“Anything that could reduce slightly hospital readmissions could save millions,” he said. 

And eventually, he said—in response to my wondering whether this might someday replace more rudimentary and less "cool" monitoring systems out there—it could even be marketed directly to consumers such as adult children worried about monitoring an elderly parent.

We exchanged cards and went our separate ways, disappearing into the sprawling convention center, Herman looking for his meeting and I looking for representatives of more companies to meet. 

Let HIMSS begin. — Tina Reed

Google Cloud announces partnership for medical imaging research

Medical imaging informatics platform Flywheel is partnering with Google Cloud to integrate Google’s Cloud Healthcare application programming interface (API) with its platform to provide clinical researchers with advanced cloud technology for medical imaging research.

Announced at the HIMSS Conference on Monday, the partnership with Flywheel marks Google Cloud’s expanding footprint in radiology. Flywheel will incorporate Google Cloud’s Big Query, a fully managed enterprise data warehouse for large-scale data analytics, to enable scalable analysis of medical imaging metadata, biomarker data, and tabular data, including genomics. Google’s Cloud AutoML Vision API, a flexible machine learning service, is used to create a comprehensive learning workflow for artificial intelligence models utilizing medical imaging data.

The research platform meets the needs of imaging center directors, principal investigators and clinical research institutions looking to create a scalable infrastructure for machine learning, advanced imaging research, and secure collaboration to advance science and ultimately precision medicine, said Flywheel CEO Travis Richardson. — Heather Landi

A softer side of tech

Among the perks of traveling for events like HIMSS is a chance for this health reporter to check out health systems in person. 

On this trip, I headed out to Nemours Children’s Health System, about a half-hour from where the conference is located, to see how tech is changing their care.

Among the examples: a command center where a group of health professionals remotely monitor vital signs for hundreds of beds in the Orlando hospital as well as at its hospitals in Delaware and Pennsylvania. 

I’ve seen similar setups in hospitals before with rooms full of computer monitors and a wall covered in a large screen. Nemours specifically uses the Epic EHR system and is using data analytics to monitor patients for signs of trouble to help alert their care team earlier as well as help identify patients most at risk of developing an infection before they have a full-blown problem. For example, the system has made a major dent in cases of sepsis, officials said.

But it’s also created other solutions. On the day I visited, one of the health professionals in the command center told us a story about noticing the monitors on an infant were indicating a problem.

She was able to remotely check in on the baby and saw that, while it wasn’t in medical distress, it was fussing and in need of comfort. In the middle of the command center, she decided to start singing a lullaby to soothe the baby while she communicated to the care team on the ground. 

She said it seemed to help a little when a member of the care team wasn’t immediately available. — Tina Reed