Cross-platform evaluation of LLM-generated educational texts on cardiac myxoma: quality, readability, and actionability using network analysis and latent profile analysis - Takeaways - MDSpire

Cross-platform evaluation of LLM-generated educational texts on cardiac myxoma: quality, readability, and actionability using network analysis and latent profile analysis

  • By

  • Bo Deng

  • Zhiqiang Wang

  • Tong Cheng

  • Zhiwen Zhang

  • Muwei Li

  • June 10, 2026

  • 0 min

Share

  • 1

    The study evaluated the quality, readability, and actionability of LLM-generated educational texts on cardiac myxoma using expert-curated questions.

  • 2

    Nine widely used LLMs generated 540 texts, revealing significant variability in information quality and readability among the outputs.

  • 3

    Latent profile analysis identified three text phenotypes: moderate-quality/low-readability, high-quality/high-actionability, and low-quality/easy-to-read.

  • 4

    Longer texts correlated with higher informational quality scores, while greater syntactic complexity was linked to lower actionability.

  • 5

    The findings indicate that LLMs currently produce limited patient education materials for cardiac myxoma, often lacking in readability and practical guidance.

Original Source(s)

Related Content