Validation of Insulin Dose Prediction Model Based on Artificial Intelligence Algorithm

NCT07066891 | Recruiting DMC

• 2 other identifiers: 202483682070918
• Study Type: interventional
• Target: 400
• Locations: 1 country
• Sites: 1

Brief Summary

The present study aims to conduct a prospective controlled trial comparing an LSTM-based artificial intelligence (AI) prediction model and clinicians' experience in the efficacy and safety of blood glucose control in hospitalized patients with type 2 diabetes mellitus (T2DM) receiving continuous subcutaneous insulin infusion (CSII) treatment in the Department of Endocrinology. The main question it aims to answer is: Is the prediction model superior to or (at least) non-inferior to clinicians' experience? Eligible patients who receive CSII treatment are randomly allocated into the prediction model group and the empirical group. Patients will:

1. 1.Receive CSII treatment as standard of care during hospitalization for 1-2 weeks, where the daily insulin dose regimen is determined by a prediction model or a clinician's experience.
2. 2.Use continuous glucose monitoring (CGM) for glucose tracking.
3. 3.Receive diabetes self-management education covering nutrition and physical activity.

Conditions

Diabetes Mellitus; Type 2 Diabetes

Keywords

long short-term memory; type 2 diabetes mellitus; continuous subcutaneous insulin infusion; intensive insulin therapy; insulin dose prediction model

Outcome Measures

Primary Outcomes (2)

• Time in range

  Time in range refers to the percentage time in range between 3.9mmol/L and 10.0 mmol/L in continuous glucose monitoring data during the continuous subcutaneous insulin infusion treatment period.

  During continuous subcutaneous insulin infusion treatment period（assessed up to 2 weeks, the treatment period between the initiation and suspension of the insulin pump)

• Time below range

  Time below range refers to percentage time of glucose\<3.9mmol/L in continuous glucose monitoring data during the continuous subcutaneous insulin infusion treatment period.

  During continuous subcutaneous insulin infusion treatment period（assessed up to 2 weeks, the treatment period between the initiation and suspension of the insulin pump)

Secondary Outcomes (2)

• Post-therapy fasting blood glucose

  12 hours after insulin pump suspension.

• HbA1c at 3 months after discharge

  3 months after discharge.

Other Outcomes (1)

• Insulin dosage during continuous subcutaneous insulin infusion treatment period

  During continuous subcutaneous insulin infusion treatment period（assessed up to 2 weeks, the treatment period between the initiation and suspension of the insulin pump)

Study Arms (2)

Prediction model group

EXPERIMENTAL

For the prediction model group, baseline information upon admission (including age, gender, body mass index \[BMI\], weight, waist circumference, fasting blood glucose before admission, glycated hemoglobin) will be put into the model, which will immediately return the insulin dosage for the first day of insulin pump treatment (basal rate and boluses of each meal). Physicians will then issue and execute these orders. On the following days, the model will adjust the basal rate and the boluses based on the blood glucose level on the previous day and the insulin dosages. This process continue iteratively for the whole CSII treatment period (about 1 to 2 weeks based on whether or not the patient is newly diagnosed or with different disease duration). The insulin pump is suspended after the administration of dinner bolus on the final day. Fasting blood glucose on the next day after insulin pump suspension will be recorded to conclude the study.

Drug: CSII in the prediction model group

Empirical group

ACTIVE COMPARATOR

During the CSII treatment, clinicians (mainly residents under the guidance of attending doctors) determine the everyday insulin dosages based on clinical experience and daily glucose monitoring, with patient data collection identical to the experimental group.

Drug: CSII in the empirical group

Interventions

CSII in the prediction model group DRUG

1. Everyday insulin dosage decided by AI prediction model. 2. CSII treatment continues for 1 to 2 weeks based on whehter or not the patient is newly diagnosed or with different disease duration.

Prediction model group

CSII in the empirical group DRUG

1. Everyday insulin dosage decided by clinicans' experience. 2. CSII treatment continues for 1 to 2 weeks based on whether or not the patient is newly diagnosed or with different disease duration.

Empirical group

Eligibility Criteria

• Age: 18 Years - 75 Years
• Sex: all
• Healthy Volunteers: No
• Age Groups: Adult (18-64), Older Adult (65+)

You may qualify if:

• Meets the diagnostic criteria of type 2 diabetes mellitus in the Chinese Guidelines for the Prevention and Treatment of Type 2 Diabetes (2020 edition).
• Insulin pump is used to control blood glucose during hospitalization, and the duration of CSII treatment period ≥6 days and \<30 days.

You may not qualify if:

• Diabetes other than type 2.
• Age ≥75 years who is not suitable for intensive insulin therapy.
• Hypoglycemic regimen other than CSII treatment, such as oral hypoglycemic drugs or multiple daily insulin injections during hospitalization.
• With severe infection or uncontrolled acute complications (including ketoacidosis coma, hyperosmolar hyperglycemia, etc.) , or any condition that the researcher believes not suitable for the study.
• Severe hepatic and renal insufficiency (ALT≥5 times the upper limit of normal, eGFR\<30ml/min/1.73m2) ), or patients at the acute stage of cardiovascular and cerebrovascular diseases considered unsuitable for study.
• Pregnancy.

Sponsors & Collaborators

• Sun Yat-sen University: lead

Study Sites (1)

Department of Endocrinology and Diabetes Center, The First Affiliated Hospital of Sun Yat-sen University

Guangzhou, Guangdong, 510080, China

RECRUITING

Related Publications (4)

• Liu L, Ke W, Li H, Li F, Fan G, Kuang J, Ma J, Zhang X, Ji B, Li S, Du Y, Xue Y, Lyu Z, Gao L, Qu S, Shi Y, Yan L, Deng W, Xu C, Dai P, Xu L, Liu J, Wan X, Wei G, Yu S, Hong S, Zhang P, Huang Z, Cao X, Liao Z, Xiao H, Mu Y, Handelsman Y, Li Y. Intense simplified strategy for newly diagnosed type 2 diabetes in patients with severe hyperglycaemia: multicentre, open label, randomised trial. BMJ. 2024 Oct 15;387:e080122. doi: 10.1136/bmj-2024-080122.

  PMID: 39406449 BACKGROUND

• Chen A, Huang Z, Wan X, Deng W, Wu J, Li L, Cai Q, Xiao H, Li Y. Attitudes toward diabetes affect maintenance of drug-free remission in patients with newly diagnosed type 2 diabetes after short-term continuous subcutaneous insulin infusion treatment. Diabetes Care. 2012 Mar;35(3):474-81. doi: 10.2337/dc11-1638. Epub 2012 Jan 6.

  PMID: 22228747 BACKGROUND

• Weng J, Li Y, Xu W, Shi L, Zhang Q, Zhu D, Hu Y, Zhou Z, Yan X, Tian H, Ran X, Luo Z, Xian J, Yan L, Li F, Zeng L, Chen Y, Yang L, Yan S, Liu J, Li M, Fu Z, Cheng H. Effect of intensive insulin therapy on beta-cell function and glycaemic control in patients with newly diagnosed type 2 diabetes: a multicentre randomised parallel-group trial. Lancet. 2008 May 24;371(9626):1753-60. doi: 10.1016/S0140-6736(08)60762-X.

  PMID: 18502299 BACKGROUND

• Li Y, Xu W, Liao Z, Yao B, Chen X, Huang Z, Hu G, Weng J. Induction of long-term glycemic control in newly diagnosed type 2 diabetic patients is associated with improvement of beta-cell function. Diabetes Care. 2004 Nov;27(11):2597-602. doi: 10.2337/diacare.27.11.2597.

  PMID: 15504992 BACKGROUND

MeSH Terms

Conditions

Diabetes Mellitus; Diabetes Mellitus, Type 2

Condition Hierarchy (Ancestors)

Glucose Metabolism Disorders; Metabolic Diseases; Nutritional and Metabolic Diseases; Endocrine System Diseases

Study Officials

• Yanbing LI, MD.&Ph.D

  First Affiliated Hospital of Sun yat-sen U

  STUDY CHAIR

Central Study Contacts

Zhimin Huang, MD. & PhD.

CONTACT

+86 13925057613; hzhim@mail.sysu.edu.cn

Yuping Cao, B.S.

CONTACT

+86 15107688525; caoyp5@mail2.sysu.edu.cn

Study Design

• Study Type: interventional
• Phase: not applicable
• Allocation: RANDOMIZED
• Masking: NONE
• Purpose: TREATMENT
• Intervention Model: PARALLEL
• Sponsor Type: OTHER
• Responsible Party: PRINCIPAL INVESTIGATOR
• PI Title: Professor, Department of Endocrinology and Diabetes Center, the First Affiliated Hospital of Sun Yat-sen University

Model Details: The insulin dose prediction model is constructed based on the database from the investigator's department, included all the type 2 diabetes mellitus (T2DM) patients who underwent short-term CSII therapy over the past 20 years. By combining patients' clinical characteristics, the investigators have input the daily capillary blood glucose (CBG) readings and insulin dose adjustment orders from the intensive therapy period into machine learning algorithms. Using a long short-term memory (LSTM) model, the system automatically extracted features and established memory patterns. Through iterative training and model optimization, the model was developed to predict the daily required insulin infusion doses (both basal rates and preprandial boluses) for any given blood glucose level.

Study Record Dates

• First Submitted: June 7, 2025
• First Posted: July 15, 2025
• Study Start: July 15, 2025
• Primary Completion (Estimated): June 1, 2026
• Study Completion (Estimated): June 1, 2026
• Last Updated: August 14, 2025

Record last verified: 2025-08

Locations

CN (1)