Currently submitted to: JMIR AI
Date Submitted: Apr 21, 2026
Open Peer Review Period: Apr 28, 2026 - Jun 23, 2026
(closed for review but you can still tweet)
NOTE: This is an unreviewed Preprint
Warning: This is a unreviewed preprint (What is a preprint?). Readers are warned that the document has not been peer-reviewed by expert/patient reviewers or an academic editor, may contain misleading claims, and is likely to undergo changes before final publication, if accepted, or may have been rejected/withdrawn (a note "no longer under consideration" will appear above).
Peer review me: Readers with interest and expertise are encouraged to sign up as peer-reviewer, if the paper is within an open peer-review period (in this case, a "Peer Review Me" button to sign up as reviewer is displayed above). All preprints currently open for review are listed here. Outside of the formal open peer-review period we encourage you to tweet about the preprint.
Citation: Please cite this preprint only for review purposes or for grant applications and CVs (if you are the author).
Final version: If our system detects a final peer-reviewed "version of record" (VoR) published in any journal, a link to that VoR will appear below. Readers are then encourage to cite the VoR instead of this preprint.
Settings: If you are the author, you can login and change the preprint display settings, but the preprint URL/DOI is supposed to be stable and citable, so it should not be removed once posted.
Submit: To post your own preprint, simply submit to any JMIR journal, and choose the appropriate settings to expose your submitted version as preprint.
Warning: This is an author submission that is not peer-reviewed or edited. Preprints - unless they show as "accepted" - should not be relied on to guide clinical practice or health-related behavior and should not be reported in news media as established information.
Clinical Accuracy and Safety of a Locally Hosted Large Language Model for Pediatric E-Consults: A Blinded Multi-Subspecialty Evaluation
ABSTRACT
Background:
Electronic consultations (e-consults) improve access to pediatric subspecialty care, particularly in rural settings, but rising consult volume contributes to subspecialist documentation burden, creating interest in whether large language models can safely assist with draft response generation.
Objective:
To evaluate the clinical utility, safety, and accuracy of a locally hosted, open-source large language model (LLM) in drafting pediatric subspecialty e-consult responses.
Methods:
We compared AI-generated consult drafts (Qwen3-30B, hospital-hosted) with human subspecialist-written e-consults for 50 real pediatric cases. Blinded pediatric subspecialists (n=50 case ratings) and generalists (n=20 case ratings) assessed accuracy, appropriateness, communication quality, and safety using structured rating instruments. Reviewer free-text comments underwent thematic analysis.
Results:
Among 50 cases, 60% of AI-generated drafts were rated as reasonable medical advice compared with 98% of physician-authored consults. False statements were identified in 39% of AI drafts, incorrect details in 58%, and potentially harmful omissions in 30%. Despite these errors, 70% of AI drafts were considered safe and potentially useful as initial drafts under specialist oversight. Performance varied by subspecialty, with neurology drafts most frequently rated reasonable (90%) and infectious disease and endocrinology drafts rated lower (40%-60%). Generalists found AI drafts understandable and comfortable to act upon in 80% of cases.
Conclusions:
While locally hosted LLMs show promise as drafting assistants to improve efficiency, high rates of clinical inaccuracies preclude their autonomous use. Specialty-specific guardrails and rigorous human oversight remain essential for safe implementation.
Citation
Request queued. Please wait while the file is being generated. It may take some time.
Copyright
© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.