A Computational Map of the Disaster Mental Health Research Landscape
Summary
- Researchers Soheil Shapouri and Saber Soleymani used Claude Sonnet to conduct an LLM-assisted review of 7,859 PubMed articles on disaster mental health.
- The study found that earthquakes and epidemics are heavily studied, while heat waves, cold waves, aggression, and suicide receive less attention.
- Machine learning, agent-based modeling, causal inference, and NLP each appear in less than 1% of the analyzed disaster mental health literature.
- Research is geographically concentrated in the United States, Japan, and China, leaving many disaster-prone regions underrepresented.
Why it matters
This computational mapping reveals critical methodological and geographical gaps in disaster mental health research, highlighting an opportunity for AI practitioners and researchers to introduce advanced computational tools into public health domains.