Review of Specialist Recommendation System Based on User-Reported Symptoms

Abstract

Finding the right medical specialist according to symptoms is one of the main problems that healthcare patients face. This can lead to delayed diagnoses, increased healthcare costs, and compromised health outcomes. A variety of dig-ital solutions have been developed over the years, such as rule-based symptom checkers, machine learning models, and assistants driven by large language mod-els(LLMs). This paper reviews these approaches used within the healthcare sector for the identification of medical specialty. Machine learning models, large language models, and retrieval-augmented generation frameworks applied for patient interaction, symptom analysis, and specialist recommendation have been explored. In addition, the paper discusses challenges related to ethical considera-tions and data privacy. Finally, it outlines emerging research directions with the aim of improving the robustness, reliability, and ethical deployment of AI-driven healthcare assistants.