Data Science Education for Residents, Researchers, and Students in Psychiatry and Psychology: Program Development and...

Area of Science:

• Behavioral Health Sciences
• Data Science in Psychiatry
• Computational Psychology

Background:

• Artificial intelligence (AI) adoption in behavioral health is growing, but mental health professionals lack adequate training.
• AI education is crucial for clinicians to implement AI tools, collaborate with data scientists, and develop interdisciplinary research skills.
• Lack of AI training hinders the integration of advanced analytics in clinical practice and research.

Purpose of the Study:

• To develop, implement, and evaluate a virtual workshop for psychiatry and psychology trainees on AI applications in suicide prevention research.
• To equip trainees with foundational knowledge in natural language processing (NLP) and Python coding for mental health data analysis.
• To assess the workshop's impact on trainees' confidence and skills in data science for suicide prevention.

Main Methods:

• A 3-hour virtual workshop was conducted, covering NLP concepts (data characterization, standardization, concept extraction) and Python coding using Jupyter notebooks on Azure Databricks.
• Electronic health record chief complaints were processed to demonstrate real-world NLP applications in analyzing suicide-related encounters across demographics.
• The Kirkpatrick program evaluation model was used, assessing participant reactions (Level 1) and learning outcomes (Level 2) via pre- and post-workshop confidence surveys and open-ended feedback.

Main Results:

• Ten psychiatry and psychology trainees participated, reporting the workshop as helpful (mean 3.17/4).
• Significant increases in confidence were observed in NLP knowledge (P=.002) and coding abilities (P=.01).
• Post-workshop confidence scores rose substantially for both NLP knowledge and coding skills.

Conclusions:

• The tailored data science workshop effectively enhanced trainees' confidence and skills in applying NLP for suicide prevention research.
• Future workshops will incorporate advanced topics like large language models, thematic analysis, and diverse datasets.
• Further research will explore the impact of this training on participants' clinical practice and research outcomes.