GHX, a supply chain management solution company, built an artificial intelligence model to predict intravenous solution shortages after Hurricane Helene temporarily shut down a Baxter International manufacturing plant in North Carolina.
The system alerts healthcare organizations to supply chain disruptions and identifies alternative products. It also can issue alerts for other events that could affect medical supply shipments, like the Baltimore bridge collapse in March 2024, and analyzes patterns and data in global supply chains that could be related to trade policies and tariffs, logistical challenges and regulatory shifts.
Related: Kaiser Permanente fund, Yale New Haven back supply chain startup
“It’s using AI to not only predict the impact of such an event, [but to] aid them in a very proactive way to make those decisions for avoiding that kind of disruption to their clinical care flow,” said Archie Mayani, the company's chief product officer.
Companies throughout healthcare are developing tools that use AI to help with diagnoses, make doctors more efficient and streamline back-office functions. Now, like GHX, they are increasingly leveraging the technology to help providers better anticipate shortages, quickly find alternative suppliers when disruptions occur and ensure products are used before they expire.
AI predictions
Sg2, a Vizient company, offers an annual forecast of inpatient and outpatient services by analyzing patient-level data and local market trends, helping healthcare providers determine their supply needs.
The 2024 forecast released in June predicted care at home to grow by 22% and that by 2034, 23% of evaluation and management visits would be virtual. It also anticipated inpatient bariatric surgery to decrease by 15% in the next decade.
“If we are able to quantify that based on the trends that we've seen in the past, but also things that we know about that are happening today or in the future, we can build our forecast models to help calibrate towards that future demand,” said Beth Godsey, senior vice president of data science and product and platform strategy at Vizient.
Another company, Premier, developed its AI model during the COVID-19 pandemic to analyze supply chain data to predict personal protective equipment shortages. The company expanded its model to predict shortages in lab supplies and other medical products affected by the pandemic, then adapted it again to manage supply strain from the surge in delayed procedures.
The model relies primarily on direct supply chain metrics such as inventory level, transit times and product substitutions. Severe weather and geopolitical events are accounted for in the AI model, but those factors don't have a strong impact on its predictions, said Matthew Shimshock, vice president of supply chain technology at Premier.
Inventory tracking technology
Inventory tracking technologies, including one that uses AI to identify expiring products, are helping to improve supply chain management from within.
By using barcode and radio-frequency identification technology, hospitals can track medical supplies in real time. Cardinal Health and Oracle can alerts customers to products that are expiring or being recalled and can generate replenishment orders. The predictive analytics used by Cardinal Health's WaveMark technology identifies products that are expiring and recommends other locations they could be transferred to so they are not wasted.
WaveMark has been in use for over 20 years, but AI and machine learning were incorporated into it in 2021. In September, Oracle partnered with hardware providers Zebra Technologies and Terso Solutions and radio-frequency identification technology provider Avery Dennison to offer its solution.
Without such technology, “you're leaving it to human eyes and that can be problematic," said Kristen Miles, Oracle's vice president of healthcare product strategy.”
“From an inventory management solution, the answer there is they buy more than they need, they store more than they need, and they throw out expired goods and chalk that up to a cost of doing business,” she said.
Columbus, Ohio-based OhioHealth implemented WaveMark in one of its catheterization labs in 2016 and in December 2023 began implementing it across the 170 catheterization labs and operating rooms in the system. WaveMark is in those locations because that's where many of its expensive supplies such as heart valves, stents and implants are stored.
When supplies arrive, the health system's inventory coordinator attaches a radio-frequency identification sticker to each box and registers it in the WaveMark system. The products are either put in a smart cabinet that continuously tracks the inventory in real time or on a shelf, where the inventory coordinator waves the Cardinal Health mobile radio-frequency identification reader over them twice a week to count them.
Catherine Roush, director of supply chain logistics at OhioHealth, said she expects the rollout to be completed by the end of 2026 and anticipates the return on investment to be six to 10 times the project's cost. The savings come from reduced waste, because since the system receive alerts before a product expires, it can relocate supplies elsewhere in the system.
"[We can] easily find and track down a supply that's expired versus manually looking in 1,500 bins," Roush said. "Then within the bin, there could be 100 items [and you're] looking for that expiration date on every supply.”
They also save by reducing inventory because they are able to follow a just-in-time approach to supplies, having on hand just what they need.
