SickKids researchers are progressing rapidly with efforts to explore the use of artificial intelligence (AI) to reimagine the future of paediatric care through automation, prediction and early detection. The goal? To uncover diagnostic answers, provide faster care, prevent harm, and improve the patient and family experience.
Supported through the AI in Medicine for Kids (AIM) initiative, teams have successfully deployed four proof-of-concept machine learning (ML) models into silent trials, which evaluate the proposed model on patients in real-time while clinicians are blinded to the ML predictions, meaning the AI does not influence clinical decision-making. This is a key step prior to deployment of AI to ensure models are accurate, safe, and equitable. To support the teams, AIM worked with SickKids’ Information Management and Technology team to design a production and testing environment to allow the ML models to interface with data.
What kinds of silent trials are taking place? Here’s one: To improve the patient experience in the Emergency Department (ED), a team led by Dr. Devin Singh, Staff Physician and Clinical Artificial Intelligence and Machine Learning Lead in the Division of Paediatric Emergency Medicine, is pioneering ML algorithms that can safely identify which diagnostic imaging and lab tests a patient needs before they are assessed by a physician. Based on initial research, the team anticipates that the models will expedite aspects of care for over 22 per cent of patients in the ED and save patients, on average, two to three hours of time spent in the ED.
The work on this silent trial, alongside ML projects related to predicting heart arrythmias before they happen, reducing surgeries for hydronephrosis and supporting medical directives for emergency care, are helping inform the future of AI-supported care at SickKids.