Singapore General Hospital (SGH) is developing an artificial intelligence solution called "Augmented Intelligence for Infectious Diseases" (AI2D), aimed at determining the necessity of prescribing antibiotics, reducing antibiotic usage, and identifying the most suitable antibiotics for each patient. This project is in collaboration with DXC Technology and currently covers pneumonia cases.
The AI2D model is built on de-identified clinical data from approximately 8,000 SGH patients between 2019 and 2020, including X-rays, clinical symptoms, vital signs, and infection response trends, covering seven commonly used broad-spectrum intravenous antibiotics. The research team conducted preliminary validation studies of the AI model in 2023, comparing it with 2,000 pneumonia cases.
In the study, SGH and DXC noted that AI2D could reduce the number of cases requiring review by one-third (from 2012 to 624). The AI model also increased the likelihood of identifying cases needing intervention in the reviewed cases to nearly 12%, compared to only 4% for traditional manual reviews. Furthermore, the analysis time for a case was reduced from 20 minutes of manual review to "less than a second."
Research shows that the AI model achieved an accuracy rate of 90% in determining whether antibiotics were needed for pneumonia cases. The study also revealed that nearly 40% of antibiotic prescriptions in these cases may be unnecessary.
SGH stated that pneumonia accounts for 20% of all infections in the hospital and is the most frequently prescribed infection type for antibiotics. The average hospital stay for patients ranges from 2 to 9 days, with the government subsidizing up to 5,000 SGD (about 3,500 USD) per patient. According to an antibiotic usage audit in 2018, SGH found that 20% to 30% of broad-spectrum intravenous antibiotics were unnecessary, while about 30% of hospital-acquired infections in Singapore are considered resistant to broad-spectrum antibiotics.
To address this global issue, the hospital is establishing an antimicrobial stewardship program to prevent the overuse of antibiotics and identify more appropriate times for recommending narrow-spectrum antibiotics. By utilizing automation and artificial intelligence, real-time insights can be better provided during prescribing to help identify cases needing review and prioritize them.
The research team is currently conducting a comparative study for 200 SGH inpatients to test the effectiveness of the AI model in reducing antibiotic usage, with plans to develop similar models for urinary tract infections in the future.
Key Points:
🌐 AI technology assists in determining the necessity of antibiotic use, reducing misuse cases.
📉 The AI model shows a 90% accuracy rate, with nearly 40% of prescriptions potentially being unnecessary.
🏥 SGH's antibiotic stewardship program aims to tackle the global resistance issue.
