Case Study Related to Disasters: A Large-Scale Analysis Using Structural Topic Modeling

INTRODUCTION: Case study has been frequently conducted in disaster-related research, yet the trends and patterns of disaster-related case study are unclear. This large-scale analysis aimed at better understanding their thematic focus and reporting practices. METHODS: Based on systematic search strategy, publication metadata from 1901 to 2023 were obtained from Elsevier/Scopus and analyzed. Structural topic modeling was employed to identify the focus areas of the topics. The number of topics was determined based on content and diagnostic metrics such as held-out likelihood and semantic coherence. Hierarchical clustering was used to categorize the identified topics. The contents and reporting styles of the most-cited articles within each topic were further analyzed. RESULTS: This large-scale analysis included 18,782 publications, showing an increase in number. The number of topics was determined as 12. They grouped into 2 overarching categories: public health and social medicine; and earth science and environmental technology. There were variabilities in reporting. CONCLUSIONS: This study highlighted a growing trend in the publication of disaster-related case studies across diverse thematic areas. As variabilities in reporting exist, there is a need for standardization in reporting to enhance transparency in disaster-related case study.