AI-Assisted Remote Insulin Pump Optimization in T1DM

Evaluating AI-Assisted Remote Optimization of Insulin Pump Settings in Type 1 Diabetes Using CGM

This study evaluates an AI-assisted remote parameter optimization tool for individuals with type 1 diabetes （T1D） using insulin pump and continuous glucose monitoring (CGM). The Artificial Intelligence (AI) algorithm analyzes regularly CGM and insulin data to generate personalized recommendations for adjusting pump settings. All algorithm-generated recommendations are reviewed and approved by physicians before being transmitted to the participant's mobile application for implementation.

Study Overview

• Status: Recruiting
• Conditions: Type 1 Diabetes Mellitus; Type 1 Diabetes (T1D)
• Intervention / Treatment: Device: AI-Assisted Insulin Pump Settings Optimization System; Device: Conventional Insulin Pump Therapy

Detailed Description

This study consists of two independent, parallel cohorts. Participants are assigned to only one of the two cohorts and do not crossover between cohorts.

Cohort 1 (Inpatient Phase): A separate group of hospitalized adults with T1D will undergo AI-assisted insulin pump parameter optimization over 5-7 days under close clinical supervision. This cohort is designed to evaluate the short-term safety, feasibility, and glycemic responses to the AI algorithm in a controlled inpatient setting.

Cohort 2 (Home-based Phase): A separate group of outpatients with T1D will be randomized in a 1:1 ratio to a two-period crossover design over 11 weeks. This cohort evaluates the efficacy and safety of AI-assisted remote parameter optimization compared to conventional therapy in the home environment.

• Study Type: Interventional
• Enrollment (Estimated): 50
• Phase: Not Applicable

Contacts and Locations

• Study Contact: Name: Jing Lu, PhD; Phone Number: 0086-025-83106666-61431; Email: lujingsnow@126.com

Study Locations

• China
  • Jiangsu
    • Nanjing, Jiangsu, China, 210008
      • Recruiting
      • Department of Endocrinology
      • Contact: Jing Lu, PhD; Phone Number: 0086-025-83106666-61431; Email: lujingsnow@126.com
      • Principal Investigator: Yan Bi, PhD
      • Sub-Investigator: Jing Lu, PhD

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Child; Adult; Older Adult
• Accepts Healthy Volunteers: No

Description

Inclusion Criteria:

1. Clinical diagnosis of T1D, meeting all three of the following criteria:

   1. Clinical medical records or a certification document issued by a specialist physician confirming the diagnosis;
   2. At least one of the following: age at disease onset < 15 years; presentation with ketosis or diabetic ketoacidosis (DKA) at onset; no overweight or obesity at onset; highest random C-peptide < 200 pmol/L;
   3. At least one of the following: continuous insulin therapy initiated after diagnosis; positive for any islet autoantibody.
2. Age 14 to 65 years, inclusive, male or female.
3. HbA1c < 11.0% at the screening visit.
4. At least 3 months of documented T1D duration, and at least 3 months of treatment with either multiple daily insulin injections (MDI) or insulin pump therapy.
5. Willing to wear the study device continuously throughout the trial and to use either insulin aspart or insulin lispro.
6. Willing to perform at least one fingerstick blood glucose measurement per day.
7. Willing to upload data from the study device.
8. Willing and able to provide signed informed consent (by the participant and/or legally authorized representative) and to learn basic diabetes management and study device operation.

Exclusion Criteria:

1. History of severe hypoglycemia within the 6 months prior to screening, defined as an event requiring medical assistance (i.e., physician office visit, emergency room visit, or hospitalization), coma, seizure, or hypoglycemia with loss of consciousness.
2. Unresolved acute complications of diabetes at screening, including diabetic ketoacidosis (DKA) or hyperosmolar hyperglycemic state (HHS), or hyperglycemia with severe circulatory impairment.
3. Unresolved adverse skin conditions at the intended device placement site (e.g., psoriasis, dermatitis herpetiformis, rash, staphylococcal infection) at screening.
4. Intolerance to subcutaneous infusion sets or adhesive tapes, or presence of edema.
5. Presence of significant acute or chronic complications, hepatorenal disease, or other systemic diseases at screening. Hyperthyroidism or hypothyroidism with unstable thyroid function at the time of screening, defined as thyroid-stimulating hormone outside the normal range with concurrent abnormal free triiodothyronine or free thyroxine levels.
6. Prior pancreas or islet cell transplantation.
7. Any of the following cardiovascular or cerebrovascular events within 1 year prior to screening: myocardial infarction, heart failure (NYHA Class III/IV), unstable angina, coronary artery bypass grafting, coronary stent implantation, angina pectoris, congestive heart failure, ventricular arrhythmia, transient ischemic attack (TIA), cerebrovascular accident (CVA), or other thromboembolic disease.
8. Diagnosis of any of the following: malignancy, tuberculosis or other chronic wasting disease, hematological disorder, psychiatric disorder, systemic lupus erythematosus, rheumatoid arthritis, autoimmune hemolytic anemia, or clinically significant malabsorption. Diagnosis of adrenal insufficiency, or current treatment for hyperthyroidism, or planned treatment for hyperthyroidism during the study period.
9. History of drug abuse or alcohol abuse.
10. Use of any oral, injectable, or intravenous glucocorticoids within 8 weeks prior to screening, or planned use during the study period.
11. Use of any of the following non-insulin glucose-lowering agents within 8 weeks prior to screening, or planned use during the study period: thiazolidinediones (TZDs), alpha-glucosidase inhibitors, sulfonylureas (SUs), glinides, dipeptidyl peptidase-4 inhibitors (DPP-4i), sodium-glucose cotransporter-2 inhibitors (SGLT2i), or glucagon-like peptide-1 receptor agonists (GLP-1 RAs). Participants currently using any of these agents must undergo a washout period of at least 7 days prior to enrollment.
12. Planned elective surgery requiring general anesthesia, or planned dialysis during the study period.
13. For women of childbearing potential: pregnancy, lactation, positive pregnancy test at screening, or planned pregnancy during the study period. Women of childbearing potential who are not using an effective method of contraception (e.g., oral contraceptives, intrauterine device, barrier method with spermicide, or surgical sterilization) and who do not agree to continue using such effective contraception during the study period will be excluded.
14. Current participation in, or treatment with an investigational drug or device within 2 weeks prior to screening (observational studies are permitted).
15. Any other condition that, in the investigator's judgment, would make the participant unsuitable for the study, including but not limited to: history of ocular trauma or other confirmed ocular disease causing blurred vision, deafness, eating disorders, celiac disease, or being underweight (Body Mass Index < 18 kg/m²).

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Treatment
• Allocation: Randomized
• Interventional Model: Crossover Assignment
• Masking: Double

Number of Arms

3

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
Experimental: Inpatient AI Optimization; During the inpatient stay, the AI-assisted decision support system will analyze continuous glucose monitoring and insulin pump data to generate daily recommendations for optimizing insulin dosage. All algorithm-generated recommendations will be reviewed and approved by physicians before implementation.	Device: AI-Assisted Insulin Pump Settings Optimization System; A cloud-based software system that analyzes continuous glucose monitoring and insulin pump data to generate recommendations for optimizing insulin pump settings. All recommendations are reviewed and approved by a physician prior to implementation.
Experimental: Home AI-First (Sequence A); Participants in this arm will complete a home-based crossover trial. In the first 4-week period, they will receive AI-assisted remote optimization of insulin pump settings. The AI algorithm will analyze weekly uploaded CGM and insulin data to generate parameter adjustment recommendations, which will be reviewed and approved by a physician prior to remote transmission to the participant's mobile app. In the second 4-week period, they will switch to conventional pump therapy without AI-assisted recommendations.	Device: AI-Assisted Insulin Pump Settings Optimization System; A cloud-based software system that analyzes continuous glucose monitoring and insulin pump data to generate recommendations for optimizing insulin pump settings. All recommendations are reviewed and approved by a physician prior to implementation.
Experimental: Home Control-First (Sequence B); Participants in this arm will complete a home-based crossover trial. In the first 4-week period, they will receive conventional pump therapy without AI-assisted recommendations In the second 4-week period, they will cross over to receive AI-assisted remote optimization of insulin pump settings. The AI algorithm will analyze weekly uploaded CGM and insulin data to generate parameter adjustment recommendations, which will be reviewed and approved by a physician prior to remote transmission to the participant's mobile app.	Device: Conventional Insulin Pump Therapy; Sensor-augmented insulin pump therapy using continuous glucose monitoring with dose adjustments made by the participant and/or physician per standard clinical practice, without AI-assisted algorithmic recommendations.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Time-in-Range (TIR)	Percentage of glucose readings within range of 3.9 to 10 mmol/l	the last 2 weeks of each 4-week intervention period

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Percentage of Time with Glucose Below 3.9 mmol/L	Percentage of CGM readings below 3.9 mmol/L during each 4-week crossover period.	the last 2 weeks of each 4-week intervention period
Percentage of Time with Glucose Above 10.0 mmol/L	Percentage of CGM readings above 10.0 mmol/L during each 4-week crossover period.	the last 2 weeks of each 4-week intervention period
Mean Sensor Glucose	Mean CGM glucose level (mmol/L) during each 4-week crossover period.	the last 2 weeks of each 4-week intervention period
Glycemic Variability	coefficient of variation (%CV) of CGM glucose levels during each 4-week crossover period.	the last 2 weeks of each 4-week intervention period
Total Daily Insulin Dose	Mean total daily insulin dose (units/day) during each 4-week crossover period.	the last 2 weeks of each 4-week intervention period
Incidence of Diabetic Ketoacidosis (DKA)	Number of DKA events occurring during the study.	through study completion, an average of 12 weeks
Incidence of Severe Hypoglycemic Events	Number of severe hypoglycemic events (requiring assistance of another person) during the study.	through study completion, an average of 12 weeks
Incidence of Adverse Events and Serious Adverse Events	Incidence of all AEs and SAEs, including relationship to study device and severity.	through study completion, an average of 12 weeks
Physician Acceptance Rate	Proportion of algorithm-generated recommendations accepted by the physician	through the 4-week AI-Assisted Systme intervention
Physician Satisfaction with AI-Assisted System	Mean scores from a modified clinician satisfaction survey evaluating usability, reliability, and overall experience.	the end of the 4-week AI-Assisted Systme intervention
Patient Treatment Satisfaction	Mean scores from the Diabetes Treatment Satisfaction Questionnaire (DTSQ)	the end of each 4-week intervention period
Percentage of Time with Glucose Above 13.9 mmol/L	Percentage of CGM readings above 13.9 mmol/L during each 4-week crossover period.	the last 2 weeks of each 4-week intervention period
Percentage of Time with Glucose Below 3.0 mmol/L	Percentage of CGM readings below 3.0 mmol/L during each 4-week crossover period.	the last 2 weeks of each 4-week intervention period

Collaborators and Investigators

• Sponsor: The Affiliated Nanjing Drum Tower Hospital of Nanjing University Medical School

Study record dates

Study Major Dates

• Study Start (Actual): January 28, 2026
• Primary Completion (Estimated): December 31, 2027
• Study Completion (Estimated): December 31, 2028

Study Registration Dates

• First Submitted: June 21, 2026
• First Submitted That Met QC Criteria: July 26, 2026
• First Posted (Actual): July 29, 2026

Study Record Updates

• Last Update Posted (Actual): July 29, 2026
• Last Update Submitted That Met QC Criteria: July 26, 2026
• Last Verified: June 1, 2026

More Information

Terms related to this study

• Keywords: Continuous Glucose Monitoring; Artificial Intelligence; Type 1 Diabetes; Insulin Pump; Insulin Dosage Adjustment
• Additional Relevant MeSH Terms: Endocrine System Diseases; Metabolic Diseases; Autoimmune Diseases; Immune System Diseases; Glucose Metabolism Disorders; Diabetes Mellitus; Nutritional and Metabolic Diseases; Diabetes Mellitus, Type 1
• Other Study ID Numbers: 2025-1221-01

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No