Copy to clipboard 
Ask any orthopedic supply chain expert about how to create an accurate forecasting system that never fails, and you’re likely to elicit laughter and a healthy amount of eye rolls.
Forecasting is equal parts art and science that requires true collaboration between multiple stakeholders. The very nature of the industry — procedural volatility, uncertain supply chains and complex SKUs — makes maintaining stable inventory levels one of the most difficult challenges that medical device manufacturers face.
The number of instruments, implants and other devices that are needed for a day’s worth of surgeries in just a single O.R. adds another layer of complexity to the inventory-stabilizing process. Companies must be ready to provide surgeon customers, who often operate at multiple hospitals and ASCs, with various amounts of instrumentation.
“Every time I operate, a small box truck arrives with around 40 cases of oversized bins to support my equipment needs,” said Jonathan Danoff, M.D., a joint replacement specialist at Northwell Health on Long Island who performs more than 700 cases per year.
On any given day, Dr. Danoff completes eight to 10 cases, each requiring a full complement of instruments and implants. “Orthopedic companies are shipping truckloads of very expensive instrumentation each year just to support my needs,” he said.
With the amount of equipment flowing back and forth between manufacturers and health systems, having a firm grasp on the frontline use filters all the way back to impact how OEMs and contract manufacturers coordinate production levels to meet future demand. It’s an imperfect science to be sure, but practical steps can keep supply shelves full and surgical schedules on track.
Down the Lane
Accurate forecasting requires accurate data, and one of the greatest disconnects in orthopedics is the lack of clean, usable information that companies can access to make informed decisions. Instead, companies often rely on what they hear from field reps who may only have anecdotal data and partial information about how many procedures are performed and what equipment was used.
“You’re relying on feedback that is at best incomplete, and at worst, a complete disaster,” said Chris Riedel, CEO and Co-founder of ConnectSX, which automates field inventory management for medical device teams. “For many companies, there are significant gaps in what they think they know about inventory in circulation and the amount that’s actually in the field.”
According to Riedel, forecasting down to the SKU level based on the amount and types of products you’re manufacturing is a must. Companies also need to collect field inventory and utilization data to dive into the metrics.
“Empower and incentivize your reps, or even your surgeons, to relay constant updates about surgical schedules,” Riedel said. “Finding ways to access that data is tremendously useful because you’ll know what’s being scheduled at the facility level.”
Marrying the volume of cases surgeons are performing with the amount of instrumentation they use is a powerful combination. Understanding the equipment preferences of individual surgeons is also paramount. Local reps often stock more of what their surgeon customers want to use so they don’t run the risk of losing out on a scheduled case.
Combining order-based forecasting with case-based data provides several advantages. “If you know the types and amounts of cases that are being booked six weeks in advance, you’ll have a head start on determining the number of products that need to be produced,” Riedel said.
Added Accuracy
Riedel is bullish on the potential of AI to transform the forecasting process. Large-language models can process massive amounts of data and identify patterns at a speed that’s beyond human capacity, and the technology will help companies better understand how to allocate product distribution.
“AI can find ways to redistribute nonmoving products based on volume,” he said. “If volumes are high in locations where loaners are being sent every week and instrument kits are sitting on a shelf in a hospital three states away that haven’t been used in six months, you need to move those kits to the busy location. You then won’t need to send loaner sets every week, and you’ll be able to turn inventory more effectively.”
As important as it is for companies to find ways to incorporate AI in their workflows, it’s just as important to clearly articulate how and why they are implementing it. Informing your team that you’ve assessed the forecasting and demand planning processes and discovered they’re spending 100 hours a month processing spreadsheets to collect accurate usage data — and letting them know that using AI could cut the time down to 20 minutes a month — is a goal teams can understand and buy into.
Dr. Danoff also believes in AI’s potential to build upon the platforms some surgeons are already using to preoperatively predict the implant sizes that are needed for individual patients and said that information will eventually infuse forecasting models with increased accuracy.
“We’re creating more predictive models that incorporate AI,” he added. “If you combine those models with data on what’s actually being used, you can forecast for specific patient demographics and pinpoint exactly how much of each implant size needs to be manufactured.”
As companies continue to lean into AI, expect to see some revolutionary changes in orthopedic forecasting precision. Having good data is one thing, but it must be shared effectively with all the stakeholders involved in manufacturing and supply chain management. That’s why it’s important to get all parties involved in forecasting decisions together at least monthly to look for ways to improve information sharing.
“Set up channels across teams and establish clear KPIs that show whether you are hitting the mark when it comes to forecasting accuracy,” Riedel said. “Review the KPIs regularly and drill down to the reasons you’re nailing them or falling short.”
From the Start
Dr. Danoff envisions a not-too-distant future in which the digitization of orthopedics changes the entire care process and, in turn, demand planning. In the clinic, he could pull up a hip replacement patient’s x-ray and rely on AI’s predictive analysis to determine the exact size of the acetabular cup, stem and femoral head that’s needed for the case.
“In that scenario, from the moment I saw the patient, I’ve already started predictably modeling,” Dr. Danoff said. “We’re moving toward that kind of system. Imagine evaluating a patient and instantly having AI suggest the best implant size and type, while also notifying the manufacturer weeks in advance. That level of integration will improve supply chain efficiency, reduce waste and lead to better patient care.”
