Related Experiment Video
Updated: Dec 9, 2025

Improving IV Insulin Administration in a Community Hospital
Published on: June 11, 2012
Insulin dose optimization using an automated artificial intelligence-based decision support system in youths with
Revital Nimri1, Tadej Battelino2, Lori M Laffel3
1The Jesse Z and Sara Lea Shafer Institute for Endocrinology and Diabetes, National Center for Childhood Diabetes, Schneider Children's Medical Center of Israel, Petah Tikva, Israel.
An artificial intelligence decision support system (AI-DSS) effectively managed type 1 diabetes by adjusting insulin pump settings. This AI-DSS was found to be as safe and effective as physician guidance for glucose control.
Area of Science:
- Endocrinology and Metabolism
- Artificial Intelligence in Healthcare
- Diabetes Technology
Background:
- Most individuals with type 1 diabetes (T1D) do not achieve glycemic targets despite advanced insulin pump and continuous glucose monitoring (CGM) use.
- Clinical inertia and time constraints in analyzing complex CGM data may hinder optimal insulin dose adjustments.
- Automated decision support systems offer a potential solution to improve glycemic control in T1D management.
Purpose of the Study:
- To evaluate the efficacy and safety of an artificial intelligence-based decision support system (AI-DSS) for automated insulin dose adjustments in T1D.
- To compare the performance of AI-DSS guided adjustments against physician-guided adjustments in a randomized controlled trial.
Main Methods:
- The ADVICE4U trial was a six-month, multicenter, multinational, randomized controlled non-inferiority trial involving 108 participants (aged 10-21 years) with T1D using insulin pumps.
- Participants were randomized to receive remote insulin dose adjustments every three weeks, guided by either an AI-DSS or by physicians.
- The primary efficacy endpoint was the percentage of time spent within the target glucose range (70-180 mg/dL).
Main Results:
- The AI-DSS arm achieved a statistically non-inferior percentage of time in target glucose range (50.2% ± 11.1%) compared to the physician arm (51.6% ± 11.3%).
- Rates of severe hypoglycemia (<54 mg/dL) were non-inferior between the AI-DSS arm (1.3% ± 1.4%) and the physician arm (1.0% ± 0.9%).
- No severe adverse events related to diabetes were reported in the AI-DSS arm, versus three in the physician arm (two severe hypoglycemia, one DKA).
Conclusions:
- Automated insulin dose adjustments using an AI-DSS are non-inferior to physician-guided adjustments for glycemic control in adolescents and young adults with T1D.
- The AI-DSS demonstrated a favorable safety profile, with no severe adverse events reported in the intervention arm.
- AI-driven decision support tools hold significant promise for optimizing insulin pump therapy and improving outcomes in T1D management.
Related Concept Videos
Insulin: Dosing Regimen and Adverse Effects
The basal dose constitutes about 40%-50% of the total daily dose, with the rest as premeal insulin. The mealtime insulin dose should mirror...
Diabetes Mellitus: Overview and Type I Subtype
Type 1 diabetes is an autoimmune disease in which the immune system mistakenly attacks and destroys the insulin-producing beta cells in the pancreas. As a result, the body is unable to produce sufficient insulin, and individuals with...
Insulin Formulations: Types and Delivery
Short-acting insulins are divided into...
Diabetes: Management and Pharmacotherapy
Insulin remains the cornerstone of treatment for most patients with type 1 and many...
Drug Dosing: Obese Patients
Diabetes Mellitus: Type 2 and Gestational

