Artificial intelligence improved efficiency of the computed tomography program.

Area of Science:

• Medical Imaging
• Artificial Intelligence in Healthcare
• Workflow Optimization

Background:

• Computed tomography (CT) is a vital but time-limited imaging resource for both inpatients and outpatients.
• Efficient and accurate utilization of CT scanners is crucial for optimal patient care.
• Existing workflows often face administrative burdens and delays, impacting patient throughput.

Purpose of the Study:

• To enhance the accuracy and efficiency of CT imaging services.
• To reduce administrative workload associated with CT referrals and scheduling.
• To expedite patient care while maintaining rigorous clinical oversight through AI and human review.

Main Methods:

• Implemented a hybrid AI workflow integrating optical character recognition (OCR) for data extraction.
• Utilized natural language processing (NLP) for structured clinical concept extraction with human validation.
• Employed a rule-based engine to match clinical needs with scanner availability for protocol recommendation and scheduling.

Main Results:

• Increased annual CT examinations by 20% (from 10,000 to 12,000).
• Achieved significant time savings, approximately 10 staff hours per week.
• Reduced patient waiting times by 30% and improved patient satisfaction by 12%.
• Decreased complaint rates from 5% to 1%.

Conclusions:

• AI-assisted workflow integration is a cost-effective and efficient solution for CT services.
• The hybrid AI approach improved safety, reduced staff workload, and enhanced satisfaction for patients and staff.
• This model demonstrates the successful application of AI in optimizing high-demand medical imaging resources.