Towards an {Integrated} {Approach} for {Sentiment} and {Emotion} {Analysis} in {Health}-{Related} {Texts}

Sentiment and emotion analysis is key to understanding how individuals convey experiences and emotional states through language. In the context of digital health, it provides insight into patient perspectives, though standard tools may not perform optimally on specialized textual data. This study explores emotional and sentiment patterns in posts written by individuals with multiple sclerosis (PwMS) on an Italian online blog. The objective is twofold: to gain a clearer picture of the lived experience of PwMS and to evaluate the reliability of existing sentiment and emotion analysis tools on domain-specific content. A selection of libraries, each addressing different emotional dimensions, was applied at sentence level to enhance granularity. Results suggest that including a neutral category would improve classification accuracy. Other cases involve the expression of multiple emotions, presenting classification challenges. Future research may consider broader units of analysis and examine strategies for adapting existing tools to health-related language contexts.