Artificial intelligence–based analysis of thrombotic and infectious risk factors in peripherally inserted central catheters (PICCs)

The study highlights the value of AI-based tools to accelerate article selection and data extraction in medical literature. However, human validation remains essential to avoid errors and misinterpretations, particularly with acronyms or heterogeneous definitions. Hybrid approaches combining AI with expert review save considerable time and are expected to improve further with multimodal models and multi-agent strategies.