Artificial intelligence (AI) is paving the way for improved pneumonia treatment by aiding in antibiotic prescription decisions. Patients' symptom information will be analyzed by an AI system to determine whether their condition is caused by a bacterial or viral infection. Based on this analysis, the system will recommend if antibiotics, which are effective only for bacterial infections, are needed. It will also suggest the specific antibiotic and the appropriate dose tailored to the patient’s clinical data.
Development and Pilot Study
The AI solution, known as Augmented Intelligence in Infectious Diseases (AI2D), was developed by Singapore General Hospital (SGH) in collaboration with DXC Technology and Synapxe, Singapore's national health tech agency. Pneumonia, a leading global cause of death and the fourth largest cause of hospitalization in Singapore, was chosen as the initial focus due to its potential to become severe quickly, according to Dr. Piotr Chlebicki, a senior consultant in SGH's Department of Infectious Diseases.
AI2D was trained using data from 8,000 SGH patients treated between 2019 and 2020, which included:
- X-rays
- Clinical symptoms
- Periodic vital signs
- Trends in body responses to infection
The model was validated with data from 2,000 additional cases in 2023. During its pilot study, AI2D accurately identified the need for antibiotics upon a patient’s first diagnosis in 90% of cases.
Potential Benefits
- Early detection of bacterial infections: AI2D processes vast amounts of data to uncover patterns that may indicate a high risk of infection.
- Targeted antibiotic use: Reducing unnecessary antibiotic prescriptions helps combat antibiotic resistance.
- Personalized care: Recommendations for antibiotic type and dosage are tailored to individual patient profiles.
Dr. Chlebicki highlighted that this system leverages AI's ability to process and detect patterns in complex medical data, ensuring more precise and efficient treatment decisions. As pneumonia is a critical global health challenge, AI-driven approaches like AI2D hold the potential to transform infection management and patient outcomes.