Singapore General Hospital (SGH) is leveraging artificial intelligence to address the growing global threat of antibiotic resistance, according to a report in Healthcare IT News. The hospital’s Augmented Intelligence in Infectious Diseases (AI2D) system, developed in collaboration with DXC Technology, aims to minimize unnecessary antibiotic prescriptions, ensure appropriate drug selection, and improve patient outcomes.
Study Demonstrates AI’s Potential to Improve Antibiotic Stewardship
SGH researchers recently conducted a pilot validation study to assess the effectiveness of the AI2D system, which is currently focused on pneumonia cases. Trained on deidentified clinical data from approximately 8,000 patients treated between 2019 and 2020, the model analyzes X-rays, symptoms, vital signs, and infection response trends to determine the necessity of antibiotic use. The system focuses on seven commonly prescribed broad-spectrum intravenous antibiotics.
In the study, AI2D was validated against 2,000 pneumonia cases from 2023 and achieved a 90 percent accuracy rate in determining whether antibiotics were warranted. Notably, the analysis revealed that up to 40 percent of antibiotics prescribed in these cases could have been avoided. The AI system also streamlined workflows, reducing the number of cases requiring manual review by threefold and increasing the identification of actionable cases.
The Urgency of Combating Antibiotic Overuse
Pneumonia accounts for 20 percent of all infections treated at SGH, making it a key driver of antibiotic use at the hospital. This contributes to a broader concern about antimicrobial resistance, which poses a serious challenge to the effective treatment of infections worldwide. According to a 2018 audit, 20 to 30 percent of broad-spectrum intravenous antibiotics prescribed at SGH were unnecessary. In Singapore, up to 30 percent of hospital-acquired infections are resistant to broad-spectrum antibiotics.
Antimicrobial resistance is not only a local concern but a global one, with many acute care hospitals prescribing inappropriate antibiotics. In response, hospitals are implementing antimicrobial stewardship programs designed to reduce hospital stays, prevent deaths and readmissions, and cut healthcare costs. AI tools like AI2D could further enhance these programs by offering real-time insights and expanding the scope of antibiotic audits. By prioritizing cases requiring intervention, such tools help healthcare teams make informed decisions faster and more efficiently.
Expanding AI’s Role in Infection Management
To build on these promising results, SGH researchers are now preparing a larger study involving 200 inpatients to evaluate AI2D’s effectiveness in reducing antibiotic use in a clinical setting. Future plans include enhancing the model to identify the most effective antibiotics for pneumonia treatment and extending its application to other hospital-acquired infections, such as urinary tract infections (UTIs).
This work reflects a broader trend of integrating AI into infection management. For example, Taiwan’s China Medical University Hospital has implemented an AI-powered platform, the Intelligent Anti-Microbial System, which tracks drug-resistant infections, predicts sepsis, and optimizes drug dosages. As healthcare institutions worldwide adopt similar technologies, AI is poised to play an increasingly critical role in combating antimicrobial resistance.
Expert Perspective
Dr. Piotr Chlebicki, a senior consultant in SGH’s Department of Infectious Diseases and a member of the AI2D project, emphasized the complexity of prescribing antibiotics. “It is often hard to tell definitively that patients will benefit from [antibiotics] based on clinical assessment, patient-specific factors, or condition severity alone,” he explained. “If not prescribed promptly for those who genuinely need them, it may lead to dire complications. Yet, misusing antibiotics contributes to antibiotic resistance, posing challenges for future infection treatment.”
AI-powered solutions like AI2D could help clinicians navigate these challenges, ensuring antibiotics are used judiciously while mitigating the risk of resistance.
Antimicrobial resistance is a growing concern, and Singapore General Hospital is exploring innovative approaches to address it. A recent study validated the AI2D system, which achieved 90% accuracy in assessing the need for antibiotics in pneumonia cases. The findings suggest up to 40% of prescriptions could have been avoided. This research highlights the potential for AI tools to enhance antibiotic stewardship and contribute to global efforts in managing resistance.