The impact of missing sensor information on surgical workflow management

Area of Science:

• Ophthalmology
• Surgical Workflow Management
• Data Science in Medicine

Background:

• Sensor systems in operating rooms can experience intermittent data loss, potentially degrading the performance of surgical workflow management systems (SWFMS).
• Incomplete or erroneous sensor data can negatively impact SWFMS functionalities like decision support, device parameterization, and information presentation during surgery.

Purpose of the Study:

• To assess the robustness of surgical process models when faced with partial or missing sensor information.
• To evaluate the impact of sensor data loss on SWFMS performance and accuracy in tracking surgical interventions.

Main Methods:

• Developed a generalized surgical process model (gSPM) from individual surgical process models (iSPMs) of 100 cataract procedures.
• Simulated sensor data loss by removing information from specific tasks within iSPMs.
• Measured the impact of data loss using metrics such as successful path relocation, convergence steps, and pathfinding failures.

Main Results:

• A gSPM constructed from 30% of iSPMs achieved 90% accuracy in identifying the correct surgical path.
• Instrument information was identified as the most critical sensor data for maintaining model accuracy.

Conclusions:

• Generalized surgical process models (gSPMs) are robust and accurate for SWFMS, even with missing sensor data.
• SWFMS can effectively provide surgical guidance in the operating room under conditions of data loss.
• Sensor systems for surgical process tracking should be evaluated based on the stability and accuracy of their functional and spatial output.