Artificial intelligence applications in social media for depression screening: A systematic review protocol for...

Area of Science:

• Mental Health Research
• Computational Linguistics
• Digital Epidemiology

Background:

• Social media platforms facilitate the formation of user groups centered around mental health conditions, notably depression.
• These platforms provide a rich source of data for understanding and predicting users' self-reported depression.
• Artificial intelligence (AI) methods are increasingly utilized for analyzing sentiment and detecting mental health indicators in user-generated content.

Purpose of the Study:

• To conduct a systematic review examining the content validity of AI-driven health surveillance models for depression detection on social media.
• To evaluate these models against standard diagnostic frameworks and clinical assessment tools.
• To provide a normative judgment on the strengths and limitations of current AI applications in social media-based depression screening.

Main Methods:

• Systematic review of English and German publications from 2010-2020 across multiple databases (PubMed, APA PsychInfo, etc.).
• Inclusion of various study types (cohort, case-control, cross-sectional, randomized controlled) and conference proceedings.
• Utilizing the PICOS tool for refining criteria, independent risk of bias assessment, and thematic synthesis of extracted data.

Main Results:

• The review will synthesize findings on the use of AI and machine learning for online depression surveillance.
• It will analyze the clinical validation procedures and content validity of these AI algorithms.
• Methodological quality will be assessed using the COSMIN framework.

Conclusions:

• The study will offer a comprehensive overview of AI's current role in social media-based depression screening.
• It aims to critically evaluate the clinical utility and validity of these computational methods.
• The findings will inform future research and development of AI tools for mental health monitoring.