Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-0752
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-08-14T06:21:27.574712Z
status: published
lens: trace
sector: health
headline: Artificial Intelligence Can Direct Patients Toward a Complaint-specific Musculoskeletal Provider
dek: Introduction Patients with musculoskeletal complaints often search online to identify an appropriate healthcare provider. With the increasing availability of large language models (LLMs), these artificial intelligence (AI) tools can direct patients to providers. This study evaluated the ability of LLMs to recommend appropriate providers based on representative patient musculoskeletal queries. Methods Three LLMs (ChatGPT, DeepSeek, and Gemini) were prompted with standardized musculoskeletal queries for two US cit…
gain_title: Large language models directed patients with musculoskeletal complaints to currently practicing, specialty-appropriate providers in the requested city, with ChatGPT achieving 100% appropriateness in the tested queries.
problem_title: When recommending musculoskeletal providers, LLMs at times provided inaccurate phone numbers and contact information, potentially preventing patients from reaching the appropriate clinic.
trace_subject: LLM recommendations of local musculoskeletal providers for patient queries in Lynchburg VA and Trumbull CT
gain_reading: Large language models directed patients with musculoskeletal complaints to currently practicing, specialty-appropriate providers in the requested city, with ChatGPT achieving 100% appropriateness in the tested queries.
gain_evidence: LLMs showed potential to direct patients to local, specialized musculoskeletal providers | ChatGPT being most often appropriate (17/17, 100%)
problem_reading: When recommending musculoskeletal providers, LLMs at times provided inaccurate phone numbers and contact information, potentially preventing patients from reaching the appropriate clinic.
problem_evidence: specific contact information was at times inaccurate | Listed phone numbers were checked for accuracy
quick_read: Researchers prompted ChatGPT, DeepSeek, and Gemini with standardized musculoskeletal complaints for Lynchburg, VA and Trumbull, CT, and judged whether recommended physicians were currently practicing locally in the relevant specialty and whether phone numbers were correct. By the August 13, 2026 publication date, ChatGPT was appropriate in all 17 recommendations, while Gemini and DeepSeek were appropriate in 43% and 40% of recommendations respectively.

The findings matter because patients increasingly use LLMs to choose providers, so accurate specialty and location matching could improve triage, but inaccurate contact details could delay care. It remains uncertain how performance generalizes beyond the two cities, standardized queries, and three models tested, and how to ensure provider contact information is accessible to models.
limitation: Findings are limited to standardized queries for two US cities and three specific LLMs, restricting generalizability to other locations, complaints, or models.
tag: Dual reading
key_points: Study prompted ChatGPT, DeepSeek, and Gemini with standardized musculoskeletal queries for Lynchburg, VA and Trumbull, CT. | Appropriateness was defined as physician currently practicing in requested location and specialized in relevant area, with phone numbers checked for accuracy. | ChatGPT was 17/17 appropriate, Gemini 9/21, DeepSeek 4/10, with statistical comparison via Fisher exact tests.
rundown: Three models were tested with identical prompts for two cities; appropriateness required current practice in the requested location and relevant specialization, and listed phone numbers were verified.

Results showed variation by model: ChatGPT 17/17 appropriate, Gemini 9/21, DeepSeek 4/10, with authors noting potential utility but also contact inaccuracies as tools evolve.
sources:
- peer_reviewed | JAAOS: Global Research and Reviews | https://doi.org/10.5435/jaaosglobal-d-25-00339 | 2026-08-13
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

38e1e49b1c6a01f0d750c6b0f9567483ec525cbddc95e025b3eb39a416be7fe7

previous

0000000000000000000000000000000000000000000000000000000000000000