---
title: "Insulin-on-board Omissions in LLM Advice for Pediatric Hyperglycemia"
id: "medrxiv-14-accounting-for-insulin-on-board-in-large-language-model-advice-about-pediatric"
canonical_url: "https://medichelpline.com/clinical-feed/medrxiv-14-accounting-for-insulin-on-board-in-large-language-model-advice-about-pediatric"
content_type: "clinical_feed_article"
specialty: "Endocrinology"
source_name: "medRxiv (Clinical Preprints)"
source_url: "https://www.medrxiv.org/content/10.64898/2026.11.06.26367316v1?rss=1"
published_at: "2026-11-08T12:00:00.000Z"
evidence_level: "Verified Feed"
license: "CC-BY-NC-4.0 / Informational Use"
---
# Insulin-on-board Omissions in LLM Advice for Pediatric Hyperglycemia
## Provenance & Clinical Metadata
- **Canonical URL:** https://medichelpline.com/clinical-feed/medrxiv-14-accounting-for-insulin-on-board-in-large-language-model-advice-about-pediatric
- **Specialty:** [Endocrinology](https://medichelpline.com/clinical-feed/endocrinology.md)
- **Primary Source:** medRxiv (Clinical Preprints)
- **Source URL:** [Original Journal Publication](https://www.medrxiv.org/content/10.64898/2026.11.06.26367316v1?rss=1)
- **Published At:** 2026-11-08T12:00:00.000Z
- **Evidence Rating:** Verified Feed
## Executive GIST (TL;DR)
- The authors evaluated large language model (**LLM**) outputs in response to caregiver questions about hyperglycemia in **pediatric type 1 diabetes**. They measured a specific safety concern: recommendations to give insulin without advising to account for **insulin-on-board**. - Investigators tested multiple models: four queried through programmatic endpoints and two accessed via chat interfaces. The source reported occurrences of the safety failure across both testing approaches. - The frequency of the omission changed when the prompts included different contextual details about the patient’s recent meal, prior insulin doses, and the treatment regimen; the study team observed context-dependent variation in LLM behavior. - The authors emphasize a need for mitigation strategies or guardrails for patient-facing LLM applications used by caregivers managing diabetes, since LLMs sometimes offered dosing advice without explicit insulin-on-board warnings. - The preprint is not peer-reviewed; authors declared sponsored research funding from Dexcom for two coauthors and listed other funders. Data and code were made available at the project repository URL provided by the investigators. - The study did not report detailed numeric rates in the abstract for each model, nor did it specify clinical dosing algorithms or precise prompt structures in the abstract; those methodological details are in the full manuscript or supplementary material referenced by the authors. - The cohort and real-world clinical impact were not established in the abstract; investigators framed findings as safety signals requiring further work rather than definitive clinical guidance.
## Clinical Analysis & Structured Key Points
Accounting for Insulin-on-board in Large Language Model Advice about Pediatric Hyperglycemia | medRxiv Skip to main content Accounting for Insulin-on-board in Large Language Model Advice about Pediatric Hyperglycemia View ORCID Profile Elizabeth Healey , View ORCID Profile Sadie Lee , View ORCID Profile Christina M Astley , View ORCID Profile Kenneth D Mandl doi: https://doi.org/10.64898/2026.11.06.26367316 Elizabeth Healey Boston Children's Hospital Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Elizabeth Healey Sadie Lee Boston Children's Hospital Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Sadie Lee Christina M Astley Boston Children's Hospital Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Christina M Astley Kenneth D Mandl Boston Children's Hospital Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Kenneth D Mandl For correspondence: kenneth.mandl{at}childrens.harvard.edu Abstract Info/History Metrics Supplementary material Data/Code Preview PDF Abstract Large language models (LLMs) can be used by patients and caregivers to answer medical questions and the characterization of safety concerns in LLM outputs for patient-facing applications is an active area of research. In this work, we study LLM responses to caregiver questions about hyperglycemia in pediatric type 1 diabetes. We measure the frequency of a potential safety failure: advice to dose insulin in the absence of a clear warning to account for insulin-on-board. We found occurrences of the potential safety failure across four LLMs tested through programmatic endpoints and two LLMs tested through chat interfaces. The rates of the potential safety failure varied when different context was given about the patient's meal and insulin history and treatment regimen. These results highlight a need for guardrails for patients with diabetes who use LLMs to manage their care. Competing Interest Statement E.H. and S.L. receive sponsored research funding from Dexcom, Inc. Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes Data Availability The code and data are available at https://github.com/lizhealey/LLM_caregiver_questions.git. Funder Information Declared Eunice Kennedy Shriver National Institute of Child Health and Human Development , T32HD040128 American Diabetes Association, https://ror.org/04f6cgz95 , 1-26-ACE-0997 Advanced Research Projects Agency for Health , 140D042690019 Copyright The copyright holder for this preprint is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license . Back to top Previous Next Posted November 08, 2026. Download PDF Supplementary Material Data/Code Email Thank you for your interest in spreading the word about medRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following Accounting for Insulin-on-board in Large Language Model Advice about Pediatric Hyperglycemia Message Subject (Your Name) has forwarded a page to you from medRxiv Message Body (Your Name) thought you would like to see this page from the medRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share Accounting for Insulin-on-board in Large Language Model Advice about Pediatric Hyperglycemia Elizabeth Healey , Sadie Lee , Christina M Astley , Kenneth D Mandl medRxiv 2026.11.06.26367316; doi: https://doi.org/10.64898/2026.11.06.26367316 Share This Article: Copy Citation Tools Accounting for Insulin-on-board in Large Language Model Advice about Pediatric Hyperglycemia Elizabeth Healey , Sadie Lee , Christina M Astley , Kenneth D Mandl medRxiv 2026.11.06.26367316; doi: https://doi.org/10.64898/2026.11.06.26367316 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Areas All Articles Addiction Medicine (638) Allergy and Immunology (917) Anesthesia (350) Cardiovascular Medicine (5005) Dentistry and Oral Medicine (496) Dermatology (428) Emergency Medicine (676) Endocrinology (including Diabetes Mellitus and Metabolic Disease) (1693) Epidemiology (16265) Forensic Medicine (36) Gastroenterology (1241) Genetic and Genomic Medicine (7279) Geriatric Medicine (751) Health Economics (1093) Health Informatics (5289) Health Policy (1467) Health Systems and Quality Improvement (1832) Hematology (599) HIV/AIDS (1381) Infectious Diseases (except HIV/AIDS) (16490) Intensive Care and Critical Care Medicine (1203) Medical Education (700) Medical Ethics (156) Nephrology (752) Neurology (7598) Nursing (377) Nutrition (1130) Obstetrics and Gynecology (1296) Occupational and Environmental Health (1040) Oncology (3746) Ophthalmology (1093) Orthopedics (409) Otolaryngology (465) Pain Medicine (501) Palliative Medicine (143) Pathology (721) Pediatrics (1858) Pharmacology and Therapeutics (756) Primary Care Research (786) Psychiatry and Clinical Psychology (6131) Public and Global Health (9992) Radiology and Imaging (2532) Rehabilitation Medicine and Physical Therapy (1519) Respiratory Medicine (1269) Rheumatology (661) Sexual and Reproductive Health (804) Sports Medicine (597) Surgery (803) Toxicology (109) Transplantation (319) Urology (308)
## Related Clinical Research

- [Type 1 Diabetes Care: Hypoglycemia Burden and Beta-Cell Replacement Research](https://medichelpline.com/clinical-feed/endocrine-news-0-type-1-diabetes-setting-the-scene-and-the-evolution-of-type-1-diabetes-care.md)
- [MiniMed 780G Performance in Asian Adults With Type 1 Diabetes: 24-Week Exploratory Study](https://medichelpline.com/clinical-feed/pubmed-43056643.md) (DOI: 10.2196/94885)
- [WHO issues first guidelines on adult type 1 diabetes care](https://medichelpline.com/clinical-feed/who-0-0-who-issues-first-guidelines-on-type-1-diabetes-for-adults.md)
- [Diabetic ketoacidosis admissions fell 6.9% annually in a 10-year Norwegian cohort](https://medichelpline.com/clinical-feed/plos-one-18-trends-and-outcomes-in-diabetic-ketoacidosis-a-10-year-retrospective-study-at-a.md)
- [Diabetes After Solid-Organ Transplant: Multi‑Organ Registry Findings from ENDO 2026](https://medichelpline.com/clinical-feed/endocrine-news-0-looking-across-the-organs-diabetes-after-transplant.md)

## Navigation
- [← Back to Endocrinology Feed](https://medichelpline.com/clinical-feed/endocrinology.md)
- [← All Clinical Specialties](https://medichelpline.com/clinical-feed.md)
## Medical & Regulatory Disclaimer

> [!CAUTION]
> MedicHelpline content is structured for research, educational, and professional discovery purposes. It does not constitute individual medical advice, clinical diagnosis, or treatment recommendations.
> Always verify dosing, contraindications, and regulatory alerts against official product labeling and primary regulatory sources before clinical decision-making.