The researchers at St. Jude Children’s Research Hospital are using artificial intelligence to better understand the healthcare needs of childhood cancer survivors. Many survivors experience long-term physical and emotional challenges after treatment, but much of this information is buried in conversations between patients and doctors. The study demonstrates that AI can help analyze these discussions and identify patients who may need additional support.
A key finding is that large language models (LLMs) can process interviews with survivors and their caregivers to detect symptoms and assess how severely these issues affect daily life. Researchers tested different prompting methods and discovered that more detailed and information-rich prompts significantly improved the AI’s ability to interpret patient experiences accurately.
This is important because a large portion—around 40% to 60%—of clinical interactions consists of patients describing their symptoms and experiences. Traditionally, much of this qualitative data is underused. By leveraging AI to analyze these conversations, doctors can gain deeper insights into patient conditions and make more informed decisions about care and interventions.
Overall, the study highlights AI’s potential to transform survivorship care by making it more personalized and proactive. Instead of relying solely on standard assessments, healthcare providers could use AI-driven insights to identify hidden risks, monitor patient well-being, and deliver targeted support—ultimately improving long-term outcomes for childhood cancer survivors.
