Effectiveness of Using Interactive Consulting System to Enhance Informed Choice

Effectiveness of Using Interactive Consulting System Based on Large Language Model to Enhance Informed Choice of Cataract Patients: a Non-inferiority Randomized Controlled Trial

To compare the performance of an interactive chatbot versus senior ophthalmologists for enhancing informed decisions made by cataract patients. The chatbot was built based on large language models, and could generate medical expert-level responses.

Cataract surgery is the only way to treat age-related cataract. For patients at early or moderate stages, cataract surgery is an elective surgery without objective indications. Therefore the patients are uaually not aware whether they should receive the surgery. The number of trained and qualified ophthalmologists are limited in China and lots of patients cannot have detailed consultation with ophthalmologists. As a result, many patients still have doubts about whether they should receive cataract surgery or not. In this study, we aim to assess the effectiveness of an interactive chatbot based on large language models versus senior ophthalmologists to enhance informed decisions made by cataract patients.

Inclusion Criteria:

1. Individuals aged 45-80 years who have been resident in the selected study district for more than six months;
2. A definite diagnosis of age-related cataract;
3. Having not received cataract surgery;
4. Willing to participate the study and provide informed consent.

Exclusion Criteria:

1. Bilateral blindness (presenting distance visual acuity worse than 3/60);
2. Having ocular, hearing or mental disorders precluding reading or telephone interview;
3. Ocular disorders other than cataract leading to permanent vision loss that could not be corrected through cataract surgery;
4. Having cataract surgery contraindication.