This is a multi-center, open-labeled, parallel group, randomized controlled trial to access the effect and safety of the Artificial Intelligence Assisted Insulin Titration System (iNCDSS) in patients with Type 2 Diabetes Mellitus.

As one of the most common treatments for diabetes inpatients, insulin regimens often vary due to different physicians. Since 2016, Zhongshan Hospital has set up the Internet-based glucose management system (iGMS) to monitor plasma glucose of diabetes patients, and further set up an Artificial Intelligence Assisted Insulin Titration System (iNCDSS) to recommend insulin regime in 2019. Previous single-center clinical trial (NCT04053959) have demonstrated the efficacy and safety of iNCDSS in glycemic management in patients with type 2 diabetes.

This multi-center study enrolls 120 patients with Type 2 Diabetes from 20 wards of Zhongshan Hospital who are on treatment with insulin for at least 3 months. They are randomly allocated into 2 groups at a ratio of 1:1 after screening for the inclusion and exclusion criteria. Patients in the Intervention group (AI group) receive insulin regimen set by iNCDSS and patients in Control group receive insulin regimen recommended by endocrinologists.

This study will be conducted in the Department of Endocrinology, Zhongshan Hospital,Fudan University and consist of a 7-day intervention period. Patient allocation will be stratified by HbA1c, BMI and previous total insulin doses. The primary endpoint is the fasting plasma glucose level after the 7-days trial period.

Inclusion Criteria:

• Men or women aged 18-99 years old;
• Inpatients who had been diagnosed with type 2 diabetes;
• Subjects who are on treatment with insulin for at least 3 months;
• HbA1c: 7.0%-10.0%.

Exclusion Criteria:

• Patients who were diagnosed with T1D, gestational diabetes or other specific types of diabetes.
• Subjects with acute complications of diabetes such as ketoacidosis or hyperglycemic hyperosmolar state;
• Patients with severe cardiac, hepatic, renal or general diseases;
• Subjects that are, in the judgement of the investigator, unlikely to comply with the protocol.
• Absence of informed consent.