Effects of a Personalized Retrieval-Augmented Generation Chatbot on Information Needs in Chronic Kidney Disease Patients: Mixed Methods Randomized Controlled Trial

This journal article reports a mixed-methods randomized controlled trial of a personalized retrieval-augmented generation (RAG) chatbot for patients with chronic kidney disease. The chatbot answers patients’ questions using their own clinical information and vetted medical guidance, and the trial measures its effect on patients’ information needs alongside qualitative accounts of how they used it.

The work was carried out with nephrologists and clinical researchers at Seoul National University hospitals. My contribution centred on the design and implementation of the RAG system and its patient-facing interface. A companion CSCW poster, Between the Clinic and the Kitchen, examines the same system through the lens of boundary infrastructure. The paper is accepted and in press at the Interactive Journal of Medical Research.