Large Language Models (LLMs), through the use of text embeddings, allow for numerical representation of narrative text into high-dimensional vectors. In this project, the History of Present Illness (HPI) portion of chronically ill pediatric patient notes, along with a standardized review of systems library were converted to text embedding vectors. By calculating the Euclidean distances between these embeddings, repetitive patterns in organ systems were identified. The output of this automated analysis was cross-referenced with annotations from two resident pediatricians to evaluate its accuracy and clinical relevance. Preliminary results show high variability between visits and low distances correlating with problematic organ systems that led to hospitalizations.