Large Language Model-Assisted Decision Support for Improving Village Doctors' Integrated Management of Patients With Type 2 Diabetes and Hypertension: A Randomized Controlled Study
Trial Snapshot
- Phase
- Not Applicable
- Status
- Not yet recruiting
- Sponsor
- Enrollment
- 20
- Primary Endpoint
- Percentage of case assessments in which the LLM-assisted management plan is preferred over the non-assisted management plan according to a blinded expert preference assessment form
Study Overview
Brief Summary
This is a prospective, assessor-blinded, randomized controlled study designed to evaluate whether large language model (LLM)-assisted decision support can improve village doctors' abilities in the comprehensive management of patients with type 2 diabetes mellitus and hypertension.
For each standardized clinical case, two village doctors will independently review the same patient information and develop management plans. An independent statistician will randomly assign one doctor to receive LLM-assisted decision support, while the other will complete the assessment without assistance. All management plans and questionnaire responses will be de-identified and evaluated by an independent expert panel blinded to group allocation. The study will compare the quality of clinical decision-making, risk assessment, treatment recommendations, and follow-up planning between the LLM-assisted and control groups.
Study Design
- Study Type
- Interventional
- Allocation
- Randomized
- Intervention Model
- Parallel
- Primary Purpose
- Other
- Masking
- Single (Outcomes Assessor)
Eligibility Criteria
- Ages
- 18 Years to 80 Years (Adult, Older Adult)
- Sex
- All
- Accepts Healthy Volunteers
- Yes
Inclusion Criteria
- •Certified village doctors or primary care physicians practicing in rural primary healthcare institutions;
- •At least 1 year of experience in the management of chronic diseases, including type 2 diabetes mellitus and hypertension;
- •Able to complete the case assessment tasks required by the study;
- •Willing to participate in the study and provide informed consent.
Exclusion Criteria
- Not provided
Outcomes
Primary Outcomes
Percentage of case assessments in which the LLM-assisted management plan is preferred over the non-assisted management plan according to a blinded expert preference assessment form
Time Frame: Immediately after completion of case assessment
Blinded endocrinology experts will independently compare paired management plans using a standardized Expert Preference Assessment Form. For each case, experts will classify the paired plans as "LLM-assisted preferred," "No meaningful difference," or "Non-assisted preferred." The primary outcome is the percentage of non-tied case assessments in which the LLM-assisted management plan is preferred.
Secondary Outcomes
No secondary outcomes reported