A Natural Language Processing-Based Chatbot as a Medium for Consultation and Education on Direct-Contact Infectious Diseases

Direct-contact infectious diseases such as influenza, diphtheria, tuberculosis (TB), scabies, varicella, impetigo, herpes simplex, and HIV remain public health threats. Limited access to accurate information encourages the development of chatbots as educational media. This study aims to design and build an NLP-based chatbot named SerMediCare to provide consultation and education on infectious diseases. The Research and Development (R&D) method with an iterative approach was used, including needs analysis, data collection from journals and medical books, and interviews with healthcare workers; system design; model training; and implementation on a web platform. The dataset was prepared in JSON format, including patterns, responses, and tags, and trained with a Transformer-based model to accurately recognize user intent. Evaluation results show that SerMediCare achieves 86% accuracy, indicating its ability to provide relevant responses to user queries. Black box testing confirmed that all features function properly. This chatbot is expected to be an effective digital tool for improving health literacy and facilitating public access to reliable information about infectious diseases.