Sentiment analysis in public health: a systematic review of the current state, challenges, and future directions

Introduction Sentiment analysis, using natural language processing to understand opinions in text, is increasingly relevant for public health given the volume of online health discussions. Effectively using this approach requires understanding its methods, applications, and limitations. This systematic review provides a comprehensive overview of sentiment analysis in public health, examining methodologies, applications, data sources, challenges, evaluation practices, and ethical considerations. Methods We conducted a systematic review following PRISMA guidelines, searching academic databases through Semantic Scholar and screening studies for relevance. A total of 83 papers analyzing the use of sentiment analysis in public health contexts were included. Results The review identified a trend toward the use of advanced deep learning methods and large language models (LLMs) for a wide range of public health applications. However, challenges remain, particularly related to interpretability and resource demands. Social media is the predominant data source, which raises concerns about data quality, bias, linguistic complexity, and ethical issues. Discussion Sentiment analysis offers the potential for gaining public health insights but faces significant methodological, data-related, and ethical challenges. Reliable and ethical application demands rigorous validation, improved model interpretability, the development of ethical frameworks, and continued research to support responsible development and deployment.