Operating rooms (OR) in the US open billions of dollars worth of sterile surgical items (SSI) on the scrub table that will not be used. Eliminating unnecessary waste and conserving SSI is an imperative in health systems afflicted by supply-chain constraints. It can also reduce costs and prevent waste.
Researchers Kaz Barker, Mohomad El Moheb, Daniel Sery, and Aidan Herzig undertook a capstone project that researched operating room sterile waste identification and reduction.
Opened but unused SSI in the OR is well-evidenced, but studies are limited in scope and generalization. Collecting actionable intraoperative data is resource intensive, usually requiring a trained observer in the OR for the entire case.
The team developed, texted, and validated a variety of computer vision (CV) models to facilitate efficient, asynchronous identification of SSI and classification into used and unused buckets. Their pilot pipeline demonstrates a proof of concept for generating stabilized video mosaics, detecting surgical items, and identifying their removal over time, laying the groundwork for future CV-assisted waste-reduction tools
While further refinement is needed to improve accuracy and classification specificity, this work lays the groundwork for future integration with surgical inventory systems and data-driven waste reduction initiatives. Ultimately, this approach has the potential to improve operational efficiency, reduce costs, and promote more sustainable surgical care.
Researchers: Kaz Barker, Mohamad El Moheb, Aidan Herzig, Daniel Sery
Sponsors: Matthew Meyer
Advisors: Bill Basener