Integrating artificial intelligence into surgical nursing workflows: Evaluation of an image-based Instrument...

Area of Science:

• Surgical Technology
• Artificial Intelligence in Healthcare
• Perioperative Management

Background:

• Effective postoperative instrument management is crucial for patient safety and operating room efficiency.
• Conventional manual counting (CMC) is time-consuming and prone to errors.
• Novel technologies are needed to streamline instrument management processes.

Purpose of the Study:

• To compare an image-based, artificial intelligence-driven instrument recognition and inventory system (AI-IRIS) with CMC for postoperative instrument management.
• To evaluate the efficiency, accuracy, and user adoption of AI-IRIS in simulated surgical settings.

Main Methods:

• A randomized crossover design was employed with 34 operating room nurses.
• Participants completed standardized counting scenarios using both AI-IRIS and CMC.
• User adoption was assessed using the Diffusion of Innovations Questionnaire (DOIQ).

Main Results:

• AI-IRIS demonstrated identical recognition accuracy to CMC under simulated conditions.
• AI-IRIS significantly reduced counting time by approximately 33-35% (26-28 seconds per scenario).
• AI-IRIS achieved significantly higher DOIQ scores, indicating improved user perception and adoption.

Conclusions:

• The AI-IRIS enhances efficiency and user perception in instrument management without compromising accuracy.
• AI-IRIS shows promise for improving safety-related counting processes and optimizing surgical workflows.
• Further research is recommended to evaluate AI-IRIS in real-world perioperative settings and with varied instrument configurations.