Using large language models to support resident-centered decision-making about transfers from long-term care to hospital

When a person living in long-term care (LTC) becomes unwell, transfer to a hospital can provide rapid access to effective treatments. However, if the person has a short time to live transfer to hospital may result in a medicalized death instead of the peaceful, comfort-focused end-of-life experience that most people desire. Daniel’s project will utilize large language models trained in serious illness conversation skills and personalized mortality prediction to support LTC residents, care partners, and physicians in making compassionate, person-centered medical decisions about hospital transfers.