Mapping emotional perceptions from stakeholders’ survey using natural language processing in the management of chronic mental illnesses—perspectives from qualitative analytics

Abstract

Chronic mental illnesses (CMI), such as schizophrenia, schizoaffective disorder, bipolar disorder (BD), and major depressive disorder (MDD), place a substantial burden on individuals, family, society and the healthcare infrastructure. Existing treatment methods fall short in addressing the needs of patients thereby leading to inadequate care and less-than-optimal health outcomes. To address this gap, our study explores a patient-centric approach through leveraging text mining and natural language processing (NLP) techniques by analysing transcribed interviews from various stakeholders, including clinicians, researchers, and healthcare professionals. Using, sentiment analysis, we examined and categorized the emotions and sentiments expressed in CMI-related discourse and explores the application possibilities using the four different lexicons in the syuzhet package in R to analyse open-ended responses in management of CMIs within academic, social and medical frameworks. The findings indicate that NRC lexicon provided text analysis methods with valuable insights into participants' emotional and attentional focus, thereby deepening our understanding of patient experiences and their reactions to interventions. Additionally, we compare sentiment analysis with outcomes from qualitative content analysis to evaluate their effectiveness in routine scientific applications and policy making. Integrating sentiment analysis into CMI management has the potential to enhance patient-centred care, ultimately leading to improved treatment outcomes. This research emphasizes the importance of leveraging innovative, data-driven methodologies to supplement conventional psychiatric care and policy development, fostering a more holistic comprehension of CMIs.