Recognizing warnings in adults at risk of life-threatening diagnoses using AI

Researcher, Unity Health Toronto & Clinician Scientist, Sinai Health

Delays, and often omissions, of life-threatening diagnoses are common, especially with hospitalized patients. To render the correct diagnosis, healthcare providers need sufficient uninterrupted time to review a patient’s medical record and speak with the patient directly, but the chaotic inpatient environment presents many challenges: interruptions, task-switching, limited time, lack of resources. Michael’s project proposes exploring the use of artificial intelligence (AI) to identify life-threatening diagnoses before they occur. AI diagnostic-prediction tools have the potential to help hospitals and providers at all stages of training realize the proactive ideal of medicine instead of the often-reactive reality. By building a tool that can identify life-threatening diagnoses in a timely manner, we can afford healthcare providers time to prepare for the outcome, prevent the outcome, and communicate with the patient, making more space for crucial elements of compassionate, patient-centered care: empathy, communication, and shared decision-making.