Objective:

To explore the development of a multilingual, voice-enabled chatbot for patient education in retinal detachment.

Key Findings:

• GPT-4o outperformed other LLMs in BLEU, ROUGE, and BERTScore metrics, indicating better alignment with reference answers.
• The chatbot's design includes screen-reader compatibility and high-contrast features to improve accessibility.
• Variability in LLM performance highlights the importance of model selection for clinical applications.

Interpretation:

The chatbot represents a significant advancement in patient education for retinal detachment, addressing limitations of traditional information delivery methods.

Limitations:

• The chatbot is still a research prototype and has not undergone clinical validation.
• Further studies are needed to assess usability, trust, and impact on patient outcomes.

Conclusion:

AI-driven conversational tools like this chatbot could enhance patient communication in ophthalmology, providing scalable and accessible education.