Intelligent сhatbot MS-Assist for clinical decision support in multiple sclerosis

Multiple sclerosis (MS) represents a significant interdisciplinary challenge requiring accurate and timely diagnosis, as well as complex, long-term patient management. Amid shortages of specialized physicians and high disease variability, digital clinical decision support systems (CDSS) are becoming critically important tools. This article describes the development and testing of the MS-Assist intelligent chatbot, designed to support physicians and provide informational guidance to patients with MS in the Russian Federation. The system is based on a Retrieval-Augmented Generation (RAG) architecture, integrates current national clinical guidelines (KR739_MS), and utilizes a domestic large language model (GigaChat). The prototype, implemented in Telegram, includes modules for calculating the EDSS score, diagnostic algorithms based on the McDonald criteria, and selection of DMTs. Testing involving 388 users showed high usability ratings (4.6 out of 5) and strong demand among the target audiences. MS-Assist demonstrates the potential to reduce diagnostic time, automate routine tasks, and be scaled to other nosologies.