Optimization of University Counseling Consent Forms With Large Language Models: Multidimensional Comparative...

Area of Science:

• Mental Health
• Information Science
• Artificial Intelligence

Background:

• University student mental health is a growing concern, often hindered by limited counseling resources and complex informed consent forms (ICFs).
• Existing ICFs frequently suffer from ambiguity, technical jargon, and incompleteness, impeding timely help-seeking and client understanding.
• Large language models (LLMs) present a potential solution for enhancing the clarity and accessibility of these crucial documents.

Purpose of the Study:

• To evaluate the effectiveness of LLM-based rewriting in improving university counseling ICFs.
• To compare the performance of two advanced LLMs, ChatGPT (GPT-5) and Grok-4, in rewriting ICFs.
• To assess improvements in structure, readability, content quality, and comprehensibility.

Main Methods:

• Comparative evaluation of 33 original Chinese university counseling ICFs against versions rewritten by ChatGPT (GPT-5) and Grok-4.
• A multidimensional framework assessed textual structure, readability (Lee-Yang Readability Index), expert-rated content quality, and volunteer-rated reading comprehension.
• Statistical analysis included Wilcoxon signed rank tests and linear mixed-effects models to validate findings.

Main Results:

• Both LLM-rewritten ICFs demonstrated significant improvements in readability, with decreased Lee-Yang Readability Index scores.
• Expert-rated content quality and volunteer-rated comprehension scores substantially increased, indicating enhanced clarity and acceptability.
• Grok-4 produced longer documents, suggesting a potential trade-off between content enhancement and document length.

Conclusions:

• LLM-based rewriting effectively enhances the readability, content quality, and comprehensibility of university counseling ICFs.
• LLMs offer a scalable method to optimize counseling documentation, improving student access to mental health services.
• Implementation requires consideration of practical factors like document length and the necessity of human oversight.