To investigate the topics and the evolving characteristics across different periods of China's health consortium policies,a multi-faceted and multi-level topic model was constructed.By utilizing structural topic model(STM),latent Dirichlet allocation(LDA)and labeled latent Dirichlet allocation(Labeled LDA),the study quantitatively analyzed the health consortium policy documents from various granularities,including texts,topics,and individual words,through both unsupervised learning and supervised validation.The research revealed that,at the macro level,"health consortium"as a topic has garnered increasing attention.At the meso level,there is a growing emphasis on aspects such as"strengthening grassroots service capabilities".At the micro level,although it is not yet widely improved there is a trend towards optimizing the utilization of supply-based and demand-based policy tools.Additionally,the study found that governments,secondary and higher-level hospitals,as well as primary care facilities,are receiving increasing attention.The number of implementation details is increasing annually,whereas prescriptive policies are decreasing annually.These findings offer valuable insights for assessing the current state of China's health consortium policies,thereby contributing to their further refinement and improvement.
关键词
健康联合体/政策分析/主题模型/时间演化
Key words
health consortium/policy analysis/topic model/time evolution
维普
万方数据