Bringing AI to Healthcare

Published: May 13, 2024


  • Healthcare facilities are gaining benefits from AI to manage data and make predictions based on staff member movement as they go about their tasks.
  • RFID data related to inventory, with AI and machine learning, ensures better replenishment forecasting, optimizes stock levels and ensures critical supplies and equipment are available when needed.

Of all the industries poised to benefit from AI, one of the most challenging is healthcare.

In hospitals, the fast-paced environment, complexity of the critical care of patients, and unpredictable nature of each event mean that current technology can fall short of solving the day-to-day challenges administrators and staffers face.

Because every single patient brings unique needs, individual staff members—and the entire hospital enterprise—must be prepared for the unexpected. And while the complexity has been growing, it is being met with more patients, fewer healthcare workers, and a system that is increasingly overtaxed.

AI offers the potential to bring some order and efficiency to this environment, but it needs good data to do so. Already at many hospitals, IoT and RFID systems are gathering data, and the result could be better management of the tasks healthcare workers take on, as well as better asset management, to name a few applications.

AI with RTLS for Staff Management

When it comes to the often chaotic healthcare environment, AI offers a way to ensure optimal orchestration of care, according to Navenio’s CEO Connie Moser.

By understanding patient needs based on service requests, hospitals can improve their reaction time, meet compliance issues and fill in gaps in patient care. To build this understanding, hospitals use a variety of real-time locating system (RTLS) technologies to capture data about clinicians, where they are, what they are doing and who they are serving.

Navenio, a workforce solutions company, has been offering an AI system to hospitals for the past two years, capturing data from Wi-Fi enabled smartphones or tablets carried by hospital personnel. It can leverage data from other RTLS systems, and even provide BLE beacons for assets when needed.

Mapping is Key

For the most part though, said Moser, “we are infrastructure free—we use Wi-Fi, we use the phone,” and a map of the organization.

Navenio typically maps the organization’s facility and applies it to a CAD drawing of the site. Users download the app on their device, sign into work with that device, and put it in their pocket to be tracked via Wi-Fi transmissions to the hospital’s existing Wi-Fi network. As they walk around the hospital, machine learning develops a 3D multi floor map of that individual’s movements, a process that takes about two weeks.

The AI system then can locate people through the phone to provide predictive support related to work strategies.

Deploying Staff

Navenio’s AI service tasking engine can be used in real time to assign staff members to a patient when that patient presses the nurse call button. The most appropriate nurse receives their task, based on location and movement patterns, and when they arrive in the room, the request is automatically turned off based on the nurse’s location.

The system can use proximity acuity, prioritization and workload balancing to sequence tasks for each care team member appropriately.

Since it was released two years ago, the solution is in use at 17 NHS hospitals in the UK as well as sites in the U.S. At NHS alone, Moser said, the AI based solution has decreased care team response times by 31 percent and “we’ve been able to improve task volume by 94 percent in some institutions,” according to Moser.

Lower Price Point

Additionally, the AI-based system can accomplish functions such as automatically triggering a bed clean when a transporter has been detected taking the patient out of the room. Discharge papers can be prompted as another way to expedite processes for patients.

Without the AI tasking engine, most RTLS solutions have typically been costly, and as a result they are deployed in limited locations within a single site (such as the emergency or operating room only). Now, using Wi-Fi based data and AI, Moser pointed out, the intelligence can be accessed at lower cost and with a more seamless deployment.

“The AI testing engine has real value because it takes chaos and makes it organized,” she said.

Inventory and Asset Management

For inventory and asset management in hospitals, AI provides vast benefits, according to Jerry Abiog, CEO and co-founder of Standard Insights. However, data capture techniques are critical.

“If you don’t have an accurate inventory, and you’re running that data through a predictive machine learning engine—it’s just not going to matter if you don’t have accurate data,” Abiog pointed out. ‘Think of it as “garbage in, garbage out’.”

Standard Insights is a five-year-old company based in Atlanta. Its team began building AI-driven management software, with predictive analytics inventory forecasting, for a healthcare client that is due to go live later this spring.

AI and its Role in RFID and IoT – the third in an ongoing series exploring how AI is impacting the RFID and IoT industries.

Examining The Interdependence of AI and RFID

AI on the Edge

Bringing Intelligence to Manual Processes

The healthcare markets Standard Insights target includes medical product manufacturers, distributors and healthcare companies. Until recently, Abiog noted that tracking supplies and equipment was still being accomplished on an Excel spreadsheet—”even if they have a viable WMS or ERP system, they often are not using it to its fullest capability.”

From tongue depressors to EpiPens, healthcare product manufacturers often had no way of knowing when their customers required restock.

But if RFID tags are applied to these products, AI data can illuminate that stock information. In fact, RFID tagged goods can help save an organization 20 to 25 percent of safety stock or over stocking. Abiog explained that AI can identify extra stock that’s unused, and other products that are approaching minimal levels.

Turning an Aircraft Carrier

Abiog noted that the healthcare industry has long been conservative when it comes to technology adoption. He likens hospitals and their management companies to giant aircraft carriers trying to pivot in deep waters.

The best way to introduce AI into these environments, he said, is to start implementing small changes first. Often, he said, a hospital may take on a small proof of concept (PoC) with minimal data analysis to gain a comfort level with AI.

Using a baseball analogy, Abiog commented “we never want to go swing for the fences first, you don’t start with the home run.”

Layering AI

Instead, healthcare companies need to see the benefits, and have less interest in the flashy features that have put AI on the front pages of media sites. ‘Sometimes the boring stuff is really the highest value,” commented Abiog.

If RFID tags are already in place on many products, Abiog pointed out, adding software and AI analytics can be a logical next step for those who touch those tagged goods.

“What we’re seeing is that people get to a certain point where they just can’t do things manually anymore, and they can then start layering AI onto their [software management] to make predictions,” he said.