AI-Based Mental Health Screening and Intervention in Disaster Settings: A Proposal for Emergency Medical Teams

Abstract

Background/Introduction:

Integrating mental health in disaster responses is crucial. Mental disorders are often neglected in the aftermath of a crisis. However, unmanaged psychiatric disorders increase individual and societal strain, making it essential for EMTs to be equipped with fast, reliable, and easy to use tools for identifying and treating such concerns, especially as disasters can intensify psychological conditions.

Objectives:

Since diagnosing mental health problems is a labor-intensive, time-consuming, and challenging process, primarily because mental health conditions are complex and often lack quantifiable tests or measures for accurate assessment, we aim to provide a decision support system for EMTs to support the detection of mental disorders.

Method/Description:

We developed an AI model trained on the health records of 812,853 people to detect depressive disorder (specificity 0.999, sensitivity 0.944). This model will now be extended to identify additional mental health conditions that are pertinent in disaster situations, such as anxiety, mood and stress disorders, and addictive behaviors. For this, it is crucial to recognize that the presentation of mental disorders is culture-specific and varies between emergency situations and normal daily lives as well as the quality and completeness of the recorded data.

Results/Outcomes:

Considering increasing rates of post-disaster related mental health disorders, our novel models help to early recognize, and phenotype diseases, streamline EMT operations, and reduce costs and suffering associated with mental health burden.

Conclusion:

Incorporating AI can significantly enhance the detection of mental health issues in disaster situations, providing timely medical intervention by EMTs and supporting the Sustainable Development Goals of health, equality, and well-being.