Research on the Real-World Community Application of Large Language Models

NCT06966882 | Not Yet Recruiting

• 1 other identifier: 2025KYPJ020
• Study Type: observational
• Target: 314
• Locations: 0 countries
• Sites: N/A

Brief Summary

There is an imbalance between the supply and demand of eye care services, especially in local communities and remote areas. To address this, it's important to use new intelligent technologies to expand the reach of eye disease screening and treatment. Large language models (LLMs) are a type of deep learning technology that can learn from large amounts of text and generate human-like language to help with medical tasks such as diagnosing diseases and answering health-related questions. The investigator's team has previously developed a localized LLM capable of answering ophthalmology-related medical questions. Building on this, this study plans to use a screening-based trial design to explore how accurately the LLM can make referral decisions for eye diseases, diagnose conditions, recommend appropriate tests, and receive user feedback in real-world community settings. The goal is to improve the ability to screen for eye diseases in grassroots and regional areas.

Conditions

Ophthalmic Diseases (Specific Types Not Restricted)

Outcome Measures

Primary Outcomes (1)

• Metrics for Evaluating Referral Accuracy of Large Language Models: Sensitivity, Specificity, Accuracy, Positive Predictive Value, Negative Predictive Value.

  through study completion, up to 1 year.

Study Arms (2)

Negative group

Patients manageable at community level；Individuals without ocular pathology

Positive group

Patients requiring specialist referral

Eligibility Criteria

• Sex: all
• Healthy Volunteers: Yes
• Age Groups: Child (0-17), Adult (18-64), Older Adult (65+)
• Sampling Method: Non-Probability Sample

Study Population

community sample or primary care clinic

You may qualify if:

• Participants of any age and gender
• Belonging to one of the following ophthalmic categories: Patients requiring specialist referral；Patients manageable at community level；Individuals without ocular pathology
• Voluntary participation with written informed consent

You may not qualify if:

• Investigator-determined clinical contraindications

Sponsors & Collaborators

• Zhongshan Ophthalmic Center, Sun Yat-sen University: lead

Central Study Contacts

Haotian Lin

CONTACT

86-13802793086; haot.lin@hotmail.com

Mingjie Luo

CONTACT

86-18200202414; zoc_mjluo@yeah.net

Study Design

• Study Type: observational
• Observational Model: OTHER
• Time Perspective: PROSPECTIVE
• Sponsor Type: OTHER
• Responsible Party: SPONSOR

Study Record Dates

• First Submitted: April 23, 2025
• First Posted: May 13, 2025
• Study Start: May 1, 2025
• Primary Completion (Estimated): December 1, 2026
• Study Completion (Estimated): December 1, 2026
• Last Updated: May 13, 2025

Record last verified: 2025-05