AI-Powered Nursing: Leveraging AI and NANDA-I for Accurate Chronic Pain Diagnosis

1. Background Chronic pain significantly impacts patient well-being and healthcare systems. Artificial intelligence (AI) and machine learning (ML) have the potential to improve nursing diagnoses, especially using NANDA-I standardized language. However, research in nonEnglish contexts, such as Italian clinical data, is limited. 2. Aim This research aims to develop AI-driven tools that assist nurses in making accurate NANDAI-based nursing diagnoses for chronic pain, using real Italian clinical records. 3. Method Two studies were conducted: the first applied ML and deep learning algorithms to six years of anonymized Italian clinical notes; the second compared different ML models, including Extreme Gradient Boosting (XGBoost) and Bidirectional Encoder Representations from Transformers (BERT), ensuring transparency through SHapley Additive exPlanations (SHAP). 4. Results Both studies showed that AI models, particularly XGBoost, effectively supported accurate chronic pain nursing diagnosis. XGBoost outperformed BERT, particularly when applied to real-world clinical data. 5. Conclusion AI and ML tools can enhance nursing diagnoses for chronic pain, especially when adapted to linguistic and contextual factors. Involving nurses in AI development ensures these tools align with clinical workflows, improving patient outcomes and care quality.