Information Extraction Framework for Disability Determination Using a Mental Functioning Use-Case

Area of Science:

• Computational linguistics and health informatics.
• Application of artificial intelligence in mental healthcare.
• Medical informatics and clinical decision support systems.

Background:

• Natural language processing (NLP) transforms health data but lags in mental health due to complexity.
• Existing information technologies for mental health are underdeveloped.
• NLP's potential in mental health management remains largely unexplored.

Purpose of the Study:

• To present a framework for advanced NLP methods in mental health assessment.
• To identify, extract, and organize mental health information for decision-making.
• To inform the disability determination process for work disability claims.

Main Methods:

• Developed a framework with 4 dimensions: temporal sequence/duration, severity, context, and information source.
• Applied NLP to extract mental health information relevant to disability claims.
• Utilized a use-case guided by the US Social Security Administration (SSA) disability process.

Main Results:

• Identified key aspects and NLP approaches for each of the 4 assessment dimensions.
• Demonstrated NLP's potential for creating functional timelines and assessing symptom severity.
• Highlighted gaps in current NLP applications and annotated datasets for mental health.

Conclusions:

• The proposed NLP framework offers a structured approach to assessing mental health and functioning.
• Findings will aid SSA adjudicators and have broader relevance for mental health assessments.
• Significant opportunities exist for NLP in mental health decision-making beyond disability claims.