Smart Health Assist: An Intelligent Medical Chatbot for Patient Support and Health Record Management

Authors:

Shivam Tripathi1, Mr Dharmendra Roy2, Saumyadeep Pal3, P Yawan Kumar4, Valluri Srujana5, Md. Mosaraf Hossain6

1Computer Science and Engineering & HITAM 2Computer Science and Engineering & HITAM 3Computer Science and Engineering & HITAM 4Computer Science and Engineering & HITAM 5Computer Science and Engineering & HITAM 6Computer Science and Engineering & HITAM

Abstract - In this work we introduce Smart Health Assist, a web-based chatbot that helps patients find reliable healthcare information and manage their medical data effectively. Smart Health Assist is a web-based chatbot that helps patients find reliable healthcare information and manage their medical data effectively. The Smart Health Assist system merges artificial intelligence with retrieval-augmented generation to provide relevant and evidence-based answers to user questions about symptoms, medications and general health advice. We use the large language model through the Ollama framework along with a ChromaDB-based vector store to make Smart Health Assist work. This setup allows for retrieval of important medical knowledge during response generation for Smart Health Assist. Smart Health Assist offers storage and retrieval of electronic medical records, such as prescriptions and diagnostic reports. It also includes a module for scheduling appointments with healthcare providers for Smart Health Assist users. The Smart Health Assist architecture has a distributed design. User-facing services and sensitive data are managed on a frontend node while the resource-heavy inference and retrieval tasks run on a backend node for Smart Health Assist. This separation helps protect data privacy and enhances system performance and scalability for Smart Health Assist. Tests on the Smart Health Assist prototype show a response time of 7.8 seconds per query. This indicates the Smart Health Assist systems potential for real-world applications. Overall, the approach demonstrates how combining RAG-based models with secure health data management can improve accessibility lessen administrative tasks and foster better patient engagement, in healthcare systems using Smart Health Assist.

Key Words: medical chatbot, conversational AI, retrieval-augmented generation (RAG), healthcare informatics, patient engagement, electronic medical records (EMR), appointment scheduling, distributed systems, Mistral, ChromaDB