This research is the first transdisciplinary study that critically examines user-generated comments from Reddit to assess Canadians’ opinion on trust in health system. Instead of traditional questionnaire and public opinion polls, we design advanced embedding-based keyword filtering for sentiment analysis and topic modeling. This allows to identify key trends and reveals both systemic and individual health concerns across major Canadian provinces. Our results reveal that key topics of concern for Canadians are related to healthcare and government spendings, lack of trust in healthcare providers, access to mental health, and nurse training. Sentiment analysis highlighted prevalence of negative (44.9%) and neutral (46.4%) sentiment, with only 8.7% of positive comments. We also obtained insights on how geographic and temporal factors influence trust in health system and health policies. This research provides valuable guidance for policymakers and stakeholders seeking to enhance public health and respond effectively to these concerns and sets the stage for how public research universities can begin to identify community priorities.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Exploring Public Trust Through LLM-Driven Opinion Mining

  • Lily Dey,
  • Fahim Anzum,
  • Ulises Charles-Rodriguez,
  • A. S. M. Hossain Bari,
  • Jean-Christophe Boucher,
  • Aleem Bharwani,
  • Marina L. Gavrilova

摘要

This research is the first transdisciplinary study that critically examines user-generated comments from Reddit to assess Canadians’ opinion on trust in health system. Instead of traditional questionnaire and public opinion polls, we design advanced embedding-based keyword filtering for sentiment analysis and topic modeling. This allows to identify key trends and reveals both systemic and individual health concerns across major Canadian provinces. Our results reveal that key topics of concern for Canadians are related to healthcare and government spendings, lack of trust in healthcare providers, access to mental health, and nurse training. Sentiment analysis highlighted prevalence of negative (44.9%) and neutral (46.4%) sentiment, with only 8.7% of positive comments. We also obtained insights on how geographic and temporal factors influence trust in health system and health policies. This research provides valuable guidance for policymakers and stakeholders seeking to enhance public health and respond effectively to these concerns and sets the stage for how public research universities can begin to identify community priorities.