In a recent development that could potentially change the patient-doctor relationship, a team at OSF HealthCare in Illinois is using artificial intelligence (AI) to determine which patients have a higher chance of dying during their hospital stay. The team has developed an AI model that can predict a patient’s risk of death within five to 90 days after admission to the hospital. The goal is for clinicians to be able to have important end-of-life discussions with these patients.
Dr. Jonathan Handler, OSF HealthCare senior fellow of innovation, explained that the organization aims to have advanced care planning discussions documented for every patient they serve, especially during sensitive times like the end of their life. By using AI to predict mortality risk, the hope is to prevent situations where patients die without receiving the full benefit of hospice care.
The AI model was tested on a dataset of over 75,000 patients, and the results showed that for patients flagged by the model as more likely to die during their hospital stay, the mortality rate was three times higher than the average. The model takes into account various factors such as clinical trends, organ functionality, healthcare visits, and age to make predictions about a patient’s likelihood of death.
While the AI model is not perfect and can result in false positives or false negatives, the goal is to stimulate conversations between clinicians and patients about end-of-life care. Dr. Harvey Castro, a board-certified emergency medicine physician, acknowledges the potential benefits of the AI model but highlights the risks and limitations, including potential distress for patients and their families due to incorrect predictions and delays in crucial end-of-life discussions.
It is important to pair AI predictions with human interaction to ensure compassionate care, and continuous monitoring and feedback are essential to ensure the accuracy and ethical use of AI in healthcare.
