There's a lot of excitement about ambient artificial intelligence in healthcare, but smart hospital platform Artisight has its eyes on developing fully hands-free autonomous monitoring and documentation in operating rooms, patient rooms and other areas throughout the hospital.
The company offers AI-driven hospital infrastructure and optimization solutions and envisions its technology operating as an "air traffic control" hub for operating room coordination.
Artisight developed voice-activated sensors, AI and computer vision technology, powered by NVIDIA GPUs, to enable real-time observation in hospital operating rooms. The company's technology also automates documentation directly in the electronic health record (EHR) system.
This eliminates the need for clinicians and other healthcare staff to manually document key milestones in an operating room procedure.
Andrew Gostine, M.D., chief executive officer and co-founder of Artisight, claims the company is the first smart hospital platform to document operating room activity directly into the EHR with computer vision.
"We are the world's only system that can passively capture and passively write into the EMR with zero clicks. We're really now unlocking the full promise of a smart hospital by being the first ambient intelligence company to deploy computer vision and autonomously write into the EMR," Gostine told Fierce Healthcare. Other smart hospital platforms still require manual input by clinicians, he noted.
"We're the only vendor in the country that has the capabilities to process video streams in real time and then push that data into the major medical record systems, so that staff don't have to do that. We can trigger automations and on top of these really accurate data sets. We're seeing hospitals optimize their staffing schedules and their OR procedure times to do more surgeries in less time, which reduces overtime surges and improves revenue," he said in an interview.
In the past two years, Artisight has grown from 40 hospitals to more than 425 hospitals across 30 health systems, he said.
In the operating room, Artisight's smart hospital platform can now autonomously detect and document patient entry and exit, procedure start and end time, and other key milestones in a patient procedure. Each action is recorded in real-time within the EHR, providing a timestamped account of events without manual input, according to the company.
Artisight also autonomously prompts clinicians to complete the next steps in their workflow based on what is happening in the room. For example, when a patient enters the room, the surgeon is alerted to scrub in, or if an operation exceeds its expected duration, the next surgical team is notified, without human intervention or oversight.
The company's use of voice-activated sensors, AI and computer vision marks an evolution of smart hospitals, Gostine said. "We now bring awareness and sensory capabilities that directly complement and support clinicians and staff. In a smart hospital, clinicians are empowered to focus on patients and know that their time and the patient's time is valued."
Artisight's platform was designed as a central hub for a smart hospital, integrating computer vision, a multi-sensor network, indoor positioning and real-time location systems, voice-activated services and video conferencing along with a hospital's EHR and other third-party systems.
The company, founded in 2015, is advancing the use of deep learning and IoT sensors to facilitate healthcare automation and optimization.
Gostine, a practicing anesthesiologist and ICU physician at an Advocate Health Care hospital in Chicago, saw first-hand the problems with hospital workflow in ORs and ICUs.
"Through medical school residency, fellowship and then working as an attending physician in the operating rooms, I started to realize that the ORs are essentially like an airport that lack any type of air traffic control," he said.
"They're very inefficient, and like an airport, it costs billions of dollars to build them. And if you build something that's billions of dollars, and then don't use it efficiently, that adds a lot of cost with not much revenue, which ultimately drives up the cost of care for all Americans," he said. "Also, if you start having limited access to the ORs, then you see hospitals rationing care, and that's where you see huge disparities start to explode, where, essentially you get marginalized patient populations that don't get access to procedures that they might need."
The fundamental premise of Artisight is that operating rooms run inefficiently because there's no coordination system with a lack of real-time data capture.
"Artisight started building smart hospitals back in 2013 even before deep learning came out because I realized all of the data that we use to run the ORs and the hospitals in general are generated by humans," Gostine said.
On average, clinicians spend 3.5 to 6 hours per day on data entry and documentation in the EHR. Automated, real-time documentation ensures greater procedural accuracy, streamlines workflows and reduces time wasted on post-procedure data entry, according to the company.
Gostine cited data from Ochsner Health, one of Artisight's health system customers, that, on average, there is a time gap of 127 minutes between when a nurse provides care and then documents that care in the EHR.
In an operating room, that time gap leads to significant inefficiencies and lost revenue. Each minute of OR time costs hospitals hundreds of dollars, and delays cascade through the schedule, preventing additional surgeries from occurring and reducing overall OR utilization.
"I ultimately realized we need to get away from human data capture and data entry," Gostine said. "I observed how people captured data and then documented it. I realized we use our senses, so we see things with our eyes, we hear things with our ears, we touch patients with our hands, we read patient vital signs and we read the medical records. We successfully ended up developing the technology to replicate all of those senses so computer vision to watch the ORs with algorithms in the camera that can determine when these events happen."
"We developed the infrastructure, the sensor layers, the processing layers, the artificial intelligence, that could essentially replicate all of human senses to do all of the data capture for us. That allowed us to let clinicians be clinicians, and let sensors to be sensors. That was a far more efficient way of doing this kind of capture and building these real-time coordination systems to run the ORs more efficiently," he added.
Gostine said Artisight's technology is "deeply ingrained" into clinical workflows. "We've seen an explosion in clinical adoption because it's not just about making the artificial intelligence, you have to actually go that last mile of deep workflow integration to make it super usable. The more passive you make it, the more adoption you're going to get. I always keep telling our staff, 'If it's not 100% passive, it's 100% useless to a doctor or nurse.' Our goal is to make everything fully hands-free, fully passive, autonomous documentation for the doctors and nurses."
RELATED: Nvidia outlines new AI projects in robotic surgery, autonomous imaging
The company says health systems using its platform have reported significant cost savings and more efficient operating room turnover. Artisight's operating room solution is already reporting a 16x return on investment and a 5% increase in scheduling procedures.
"From my perspective as a physician, I didn't go to medical school necessarily because I'm super interested in optimizing hospital P&L. I'm really interested in the amount of access and care that we've delivered to marginalized populations to make sure that there's not forgotten populations in the world who can't get access to life-saving procedures," Gostine said.
The Guthrie Clinic worked with Artisight to launch the Pulse Center, an advanced remote nursing program with a virtual command hub. The provider reported that nurse turnover dropped significantly, from 25% to 13%, nurses saved more than 30 minutes per shift on EHR data entry and the organization achieved $7 million in cost savings, according to a case study.
WellSpan Health implemented the company's smart hospital platform for virtual observation of patient rooms and virtual nursing. The health system reported a 52% reduction in patient falls and a drastic reduction in the need for in-person sitters.
Artisight's technology can be deployed in ORs, patient rooms, command centers, supply closets as well as parking lots and loading docks. "There is not a spot in the hospital that wouldn't benefit from better, more passive data capture," Gostine said.
The company is focused on enhancing its AI and computer vision technology. Advances with multimodal models and vision language models, which combine language and visual images, open up opportunities for smart hospital tech, he noted.
"We're getting very close to liberating nurses from the medical record system completely, where everything they do can be documented subject to the nurses' sign-off," he said.
The company raised $42 million in a series B round in January 2024 backed by multiple strategic and client health system investors as well as full participation from series A investors, including chipmaker Nvidia.