Industry Spotlights | November 3, 2025

Using AI to improve patient diagnosis

Antimicrobial resistance (AMR) is a major global health challenge, making infections harder to treat and putting lives at risk. The overuse and misuse of antibiotics are key contributors to this problem, with studies showing that up to 50% of antibiotics prescribed in hospitals may be unnecessary or inappropriate. Recognizing the need for action, Singapore General Hospital (SGH) teamed up with DXC Technology and Synapxe to create AI-driven solutions that transform how antibiotics are prescribed.

At the heart of this collaboration is the “Augmented Intelligence in Infectious Diseases (AI²D)” system. AI²D analyzes patient-specific data, like x-rays, vital signs, symptoms and infection trends, to guide doctors in making smarter decisions about antibiotic prescription. Initially focused on pneumonia and urinary tract infections (UTIs), the system helps reduce unnecessary prescriptions while ensuring patients get the right treatment promptly.

DXC and SGH partnered to build an AI model that helps antimicrobial stewardship program (ASP) teams quickly and accurately evaluate antibiotic use. Using data from patients treated with seven broad-spectrum intravenous antibiotics for UTIs, the model applies advanced clinical reasoning to identify potential misuse cases. This AI solution has improved ASP team efficiency by enabling real-time insights into each patient’s condition and recommending the most suitable treatment approach.

SGH faced a dual challenge: ensuring patients receive timely and accurate antibiotic treatments to prevent complications, while minimizing unnecessary antibiotic use to fight AMR. Doctors often struggle to balance these priorities due to limited real-time information.

For instance, SGH’s ASP teams conducted around 12,000 audits of broad-spectrum antibiotic use in 2018, but manual reviews covered only a fraction of the cases. Expanding audit coverage and shifting to a proactive approach for stewardship became essential.

AI²D was created to help doctors determine whether antibiotics are necessary and recommend the most appropriate treatments based on clinical data. The AI²D platform:

• Uses anonymized historical data to build predictive models for pneumonia and UTIs
• Provides real-time insights to help doctors prescribe antibiotics more precisely
• Continuously improves through ongoing validation and studies, including a current project involving 200 inpatients

With automation, SGH’s ASP team can now focus on complex cases, expand their audit reach, and ensure better antibiotic stewardship. AI²D’s pneumonia model, tested on 2,000 cases in 2023, achieved 90% accuracy in identifying whether antibiotics were necessary. In a review of 2,012 pneumonia treatments, the AI model flagged 624 cases for further review, successfully identifying 87% of intervention cases.

• 90% accuracy in determining antibiotic necessity: Doctors are empowered with data-driven insights to make better decisions and reduce unnecessary prescriptions
• Increased efficiency: Automation allows ASP teams to review more cases in real time, reducing the workload of manual reviews
• Better patient outcomes: Ensuring safer and more effective antibiotic use helps minimize the risk of AMR
• Future-ready design: AI²D is scalable, with plans to enhance its capabilities for identifying optimal antibiotics and dosages for pneumonia and UTIs
• A strong partnership: The collaboration between SGH, DXC Technology, and Synapxe showcases the power of partnership in solving global healthcare challenges