Collaborating to cut hospital costs

Patients being wheeled in and out of operating rooms are unlikely to ever notice a significant source of wasteful healthcare spending: unused surgical instruments.

In a typical surgery, 70-80% of the instruments that enter the operating room aren’t used. But every re-usable instrument that enters an operating room must be sterilized again, even if it wasn’t used.

Hospitals also incur costs to purchase instruments and to lay them out in pre-planned surgical trays in advance of every procedure. The more instruments that go into an operating room, the higher the costs.

Surgical procedures account for more than 60% of the operating costs of a hospital and about 15% of those costs – or roughly 9% of total hospital operating costs – are spent on surgical instruments. For a medium-sized hospital, that’s $4 million to $7 million per year in total surgical instrument spending. Across the 6,000-plus hospitals in the U.S., the costs soar into the billions.

The problem isn’t new and a comprehensive, practical solution has been out of reach – until recently.

Researchers at UNC Kenan-Flagler Business School have developed an algorithm that enables software to automate most of the complex task of reducing unnecessary instruments in operating rooms.

It could save a typical hospital millions of dollars a year.

Vinayak Deshpande, Mann Family Distinguished Professor of Operations and Sandeep Rath, assistant professor of operations, and Nishanth Mundru, a former UNC Kenan-Flagler professor, developed the algorithm in collaboration with the founders of OpFlow, a healthcare software technology firm.

They based their work on data collected by OpFlow, which was founded by two UNC Executive MBA graduates and surgeons Dr. Benjamin C. Wood and Dr. Martyn Knowles (both MBA ’18).

Wood, Knowles and David Rowe started the firm to optimize surgical instrument and surgical tray use. They use software to collect data that feeds into the algorithm developed by the UNC Kenan-Flagler researchers.

They detail their work in “Data-Driven Surgical Tray Optimization to Improve Operating Room Efficiency,” which was a finalist for the prestigious 2022 Innovative Applications in Analytics Award by the Institute for Operations Research and the Management Sciences (INFORMS). Operations Research published their findings.

Diagnosing the problem

Before UNC researchers could tackle the problem of unused surgical instruments, they had to understand why the current system is so inefficient.

Surgeries – and the various instruments used during them – are planned in advance. Surgeons list the instruments they need on individual preference cards. Technicians then assemble one or more trays with the instruments, which takes 10-30 minutes.

Preference cards, even for the same surgery, can vary depending on an individual surgeon’s preferences. Instruments are sometime added to them, but rarely removed. As a result, hospitals deal with a vast number of surgical tray configurations.

Some hospitals have optimized trays by having surgeons and other experts go over the preference cards and take off unneeded instruments. But this process is labor intensive, not data driven and impractical given the number of preference cards a typical hospital deals with.

Before the UNC researchers could tackle the problem mathematically, they needed data. OpFlow had that data based on real surgeries at UNC Rex Hospital in Raleigh, North Carolina, where hospital personnel collect data on an iPad to quickly record which instruments were unused after surgical procedures.

Simplifying an unsolvable problem

The surgical tray optimization problem is hard. After they modeled the problem mathematically, the UNC Kenan-Flagler faculty found that it was “NP-hard” – a formal term from computer science that means, in layperson’s terms, that no computer, no matter how powerful and how much time it took, could find the best tray configuration for a real-world sized problem.

So the UNC professors broke the problem into parts and simplified it. While the algorithm’s solutions aren’t perfect – there are bound to always be some unused instruments – it dramatically reduced the number of unused instruments.

They developed the algorithm using instrument usage data for 1,448 surgeries conducted at UNC Rex Hospital between August 2018 and June 2020. Those surgeries produced 78,073 rows of data for UNC Kenan-Flagler researchers to work with.

Seventy percent of the data was used to develop the algorithm, while the remainder served as a test group to ensure the algorithm would still perform well on surgeries that hadn’t been used in its development.

The researchers’ approach first seeks to aggressively reduce the number of instruments on each tray. It then adds back instruments by creating new add-on trays to ensure that surgical teams have everything they need, even when some instruments are only used by a small number of surgeons or in a small number of procedures.

When they tested the solution on vascular surgery data, the algorithm reduced the overage –the number of unused instruments – from 114 instruments to 53, a 53.5% reduction.

Safety still comes first

Although the combination of real data and the optimization algorithm significantly reduced instrument usage, optimization efforts wouldn’t go anywhere if physicians or hospitals believed they increased risk for patients.

Some instruments, for example, are rarely used in most surgeries, but are critical in the case of rare medical emergencies. For UNC Rex Hospital, the OpFlow team worked with surgeons to evaluate the algorithm-derived instrument trays and then add to them any rarely used but critical instruments.

In addition to reducing surgical instrument costs, the method also has the potential for other positive impacts. For example, most hospitals devote substantial space to storing, preparing and sterilizing surgical instruments.

By cutting those costs hospitals might be able to reduce the amount of space they need for those processes, say Deshpande and Rath. That space could, in turn, be used for one or more new operating rooms, increasing hospital surgical capacity at minimal cost.

The U.S. spends more on healthcare per capita than any other country in the world. In 2019, healthcare spending grew 4.6%, but hospital operating margins have gotten tighter – dropping from 4.18% to 2.56%.

Surgical instrument optimization gives hospitals another tool in their battle against wasteful spending.