AI Clinical Support in Disaster Medicine: A Delphi Study

Artificial intelligence (AI) is increasingly explored for its potential to revolutionize disaster medicine. A recent Delphi study involving 131 international experts has shed light on the specific clinical challenges that practitioners believe AI could address. This study provides a global perspective on how AI can enhance clinical decision support in disaster situations, focusing on logistical and operational needs.

Background: AI in Disaster Medicine

Since the 1950s, AI technologies have evolved significantly, but their integration into disaster medicine has been limited. The introduction of advanced AI tools like ChatGPT in 2022 sparked renewed interest in applying AI to disaster response. However, much of AI development is driven by vendors in high-income countries, often overlooking the practical needs of practitioners in diverse regions.

This study aims to bridge the gap between AI development and real-world requirements by gathering insights from experts worldwide. By understanding the specific challenges faced by disaster medicine practitioners, AI can be tailored to provide more effective and relevant clinical support.

Methods: A Global Delphi Study

The study employed an online Delphi method, engaging 131 disaster medicine experts across three rounds. In the first round, participants were asked to identify clinical questions or problems they would like AI to address. Their responses were analyzed and compiled into statements for further evaluation.

In subsequent rounds, participants rated these statements on a 7-point scale, indicating their importance. This iterative process helped refine the most critical areas where AI could make a significant impact.

Results: Key Areas for AI Support

In the first round, participants proposed 539 statements, which were consolidated into 47 key statements for later evaluation. By the third round, consensus was reached on five critical areas where AI could provide valuable support:

Hospital distribution of disaster patients
Estimating the size of affected populations
Hazard vulnerability analysis
Resource acquisition and distribution
Transportation routing

Experts showed less enthusiasm for AI involvement in areas requiring complex decision-making, such as ethics, mental health, cultural sensitivity, and treatment choices.

Conclusion: Prioritizing User Needs

The Delphi study highlights the importance of aligning AI development with the practical needs of disaster medicine practitioners. While AI shows promise in enhancing logistical aspects of clinical duties, its application in more nuanced areas like ethics and mental health remains contentious.

Future AI development for clinical decision support in disaster medicine should prioritize user needs and incorporate a global perspective to ensure its relevance and effectiveness.

Implications for AI in Healthcare

This study underscores the potential of AI to transform disaster medicine, but it also emphasizes the need for a user-centric approach. By involving practitioners in the development process, AI can become a more powerful tool for managing complex disaster scenarios.

As AI continues to advance, its role in healthcare will likely expand, making it essential to address the ethical and practical challenges identified in this study.