To demonstrate that a single video question can effectively predict self-reported depression, anxiety, and trauma through text and voice analysis, addressing the need for efficient mental health assessments.
Key Findings:
Achieved 64.6% reduction in assessment time (78.4 s vs 221.7 s) while screening for all three conditions, indicating significant efficiency.
Only 1.4% of participants were unwilling to use video-based screening, suggesting high acceptance.
Demonstrated strong predictive performance and demographic consistency across age, gender, and race/ethnicity, reinforcing the model's applicability.
Interpretation:
The study supports the feasibility of efficient multi-condition mental health screening using brief video responses, particularly in the context of increasing provider shortages.
Limitations:
The model does not directly assess trauma exposure despite predicting PCL-5 scores, which may limit its comprehensiveness.
Further validation may be needed across larger and more diverse populations to ensure generalizability.
Conclusion:
The study presents a scalable and efficient method for mental health assessment that could alleviate the burden on healthcare providers and improve patient engagement.
