Efficacy and Safety of an Intelligent Insulin-Dosing Decision Support System for Glycemic Control in Patients With Type 2 Diabetes: A Randomized Controlled Study

This study evaluates an intelligent insulin dosing decision support system designed to support basal insulin titration in adults with type 2 diabetes. The system operates on an internet-of-things platform that synchronizes glucose data and insulin dose records from a smart glucometer and an insulin logging device to a cloud-based server. Patients and clinical staff access the system through a mobile interface, which enables real time data transmission, automated dose calculation, and remote supervision.

The system contains four major components. The first component is a comprehensive assessment module that collects demographic information, clinical characteristics, and a full panel of islet function indicators including fasting glucose, postprandial glucose, glycated hemoglobin, fasting insulin, and multiple C-peptide values. A validated hypoglycemia risk prediction model classifies each participant into low, intermediate, or high risk categories. Competency in insulin injection and dose verification is assessed using a standardized questionnaire.

The second component determines the initial basal insulin dose. This calculation follows guideline recommendations and incorporates the islet function results and hypoglycemia risk category to generate an individualized starting dose.

The third component evaluates insulin sensitivity and glucose response patterns. Continuous transmission of glucose values and insulin doses allows the system to estimate the participant's insulin sensitivity level and to update these estimates as additional data are obtained.

The fourth component produces dose adjustment recommendations. Titration is structured into three time-based phases with clearly defined monitoring frequencies and clinical review points. During the initial titration phase from day three to day fourteen, fasting glucose is measured daily and the system generates dose recommendations every three days using the mean of the most recent fasting glucose values. Nurses review glucose patterns and safety concerns on day four and at the end of the first week, while physicians perform dose verification on day four. During the stable titration phase from week three to week twelve, fasting glucose is measured every three days and the system generates weekly dose recommendations based on the most recent value. Nurses conduct follow up assessments and provide stepwise guidance at week one, week two, week four, week eight, and week twelve. Physicians review system-generated recommendations at the same time points. During the long term maintenance phase from week thirteen to week twenty four, fasting glucose is measured weekly and the system continues to produce weekly dose recommendations. Participants who demonstrate sufficient accuracy in dose verification by week twelve may be authorized to review system-generated dose units independently from week thirteen onward under remote supervision. Physicians complete a final evaluation at week twenty four.

Safety supervision is integrated throughout the entire study. The system continuously evaluates glucose fluctuations and recalculates hypoglycemia risk at each dose adjustment. When a fasting glucose value falls below the predefined safety threshold or when a proposed adjustment exceeds the predetermined limit, a safety alert is automatically triggered. Nurses are required to conduct follow up within twenty four hours and document the event. Both self-reported and device-captured hypoglycemia events are recorded and reviewed across the study period.

The mobile platform provides automated reminders for glucose monitoring, displays dose recommendations, and delivers educational content related to insulin use. Patients may submit questions and receive guidance through text or image messages. The clinician interface summarizes glucose trends, allows review of system-generated dose recommendations at scheduled verification points, and supports real time management of safety alerts.

The intervention incorporates a structured empowerment-based management process that aligns with the titration timeline. The five components include identifying patient challenges during insulin initiation, addressing concerns and emotional responses, establishing short term and long term treatment goals, developing individualized action plans, and evaluating outcomes at specific time points. Nurses and physicians conduct these steps at the designated visits that coincide with the initial, stable, and maintenance titration phases. The process is designed to gradually strengthen the participant's ability to verify dose adjustments, prevent hypoglycemia, and maintain long term adherence to basal insulin therapy.

Outcome measurements are collected at baseline and week twelve. These include glycemic indicators, hypoglycemia frequency, empowerment ability, self-management behaviors, and attitudes toward insulin therapy. Through this integrated system of algorithm-guided titration, continuous data monitoring, structured follow up, and staged empowerment education, this study aims to determine whether the intelligent dosing system can improve the precision and safety of basal insulin titration and enhance behavioral readiness for long term insulin self management.

Inclusion criteria

Patients who

• were diagnosed with T2DM according to the Guideline for the Prevention and Treatment of Diabetes Mellitus in China (2024 edition);
• were prescribed basal insulin therapy for the first time;
• were aged 18 years or older;
• voluntarily participated in this study; and
• had access to mobile devices and the internet.

Exclusion criteria

Individuals who:

• had severe diabetes-related complications (e.g., cardiac, hepatic, cerebral, or renal insufficiency);
• were unable to participate in regular follow-ups;
• had communication or cognitive barriers;
• had diagnosed psychiatric disorders; or
• discontinued basal insulin therapy for non-glycaemic reasons.