Automated Surgical Protocol Violation Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current surgical audit methods are inefficient and inaccurate due to the lack of an automated system to systematically review surgical protocols and environments, leading to potential surgical site infections (SSIs) and suboptimal patient care.
Innovation Solution
An automated system employing multiple cameras and machine-learning algorithms for real-time monitoring of surgical workflows, detecting non-compliance with protocols, and providing alerts and recommendations to enhance surgical safety and patient outcomes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If independent consultants are used to audit surgical procedures, then surgical safety review is performed, but the system is inefficient, inaccurate, and expensive due to inability to observe all surgeries
Solution Approach 1:
The patent replaces the manual mechanical auditing process performed by independent consultants with an automated computer vision system using machine learning models. The system captures video feeds from operating rooms, processes them through trained models to detect surgical milestones and activities, and automatically determines protocol compliance, thereby eliminating the inefficiencies of manual review while maintaining or improving accuracy.
Solution Approach 2:
The surgical audit system performs self-service by automatically monitoring its own environment without requiring external human consultants. The machine learning models continuously analyze video streams, detect protocols violations, and generate reports autonomously, enabling the system to serve itself in the auditing function it was designed for.
2Measurement precision
If consultants manually observe surgical procedures, then potential causes for SSIs are investigated, but the coverage is insufficient and recommendations are inaccurate
Solution Approach 1:
The automated system serves multiple surgical areas simultaneously through a networked camera and processing system. A single system can monitor multiple operating rooms, perform various types of surgical audits, and generate comprehensive reports, making the auditing capability universal rather than limited to single-point manual observation.
Solution Approach 2:
The system changes the parameter of audit coverage from individual consultant capacity (limited) to system capacity (unlimited). By transitioning from human observation to automated video analysis, the quantity of surgeries that can be audited increases dramatically while maintaining high detection accuracy through trained machine learning models.
3Reliability
If traditional surgical auditing is performed, then some protocol violations are detected, but real-time prevention of SSIs is not achieved
Solution Approach 1:
The system implements real-time feedback by continuously monitoring surgical videos, detecting protocol violations as they occur, and immediately alerting the surgical team. This closed-loop feedback mechanism allows for immediate correction of violations rather than post-hoc analysis, enabling real-time prevention of SSIs.
Solution Approach 2:
The system performs preliminary detection of potential protocol violations before they result in SSIs. By monitoring surgical activities in real-time and identifying compliance issues as they emerge, the system enables preventive action rather than reactive response, reducing the time loss between violation occurrence and correction.
Data Source
Figure 1
Figure 2
Figure 3A~3B
AI summary
The present disclosure relates generally to improving surgical safety, and more specifically to techniques for automated detection of non-compliance to surgical protocols in a surgical environment such as an operating room. An exemplary method comprises: receiving one or more images of the operating room captured by one or more cameras; detecting a surgical milestone associated with a surgery in the operating room using a first set of one or more trained machine-learning models based on the received one or more images; detecting one or more activities in the operating room using a second set of one or more trained machine-learning models based on the received one or more images; and determining, based on the detected one or more activities and a surgical protocol associated with the detected surgical milestone, that an instance of non-compliance to the surgical protocol has occurred in the operating room.