The goal of this clinical trial is to evaluate the intervention effects on blood sugar levels using an AI diet app for personalized meal control and the continuous glucose monitoring system (CGMS) in patients with type 2 diabetes (T2D)

The main questions it aims to answer are:

1. Effect of glycemic control on hemoglobin A1c (HbA1c) in T2D patients through continuous glucose monitoring system (CGMS)

2 Effect of blood sugar management intervention using AI diet app

◎ Research method

<study arms>

• Negative Control group (NC) 20 people (*24 people) checking only hemoglobin A1c (HbA1c) three times (at the beginning of the study, after the 1st 12 weeks, and then after the 2nd 12weeks)
• Positive control group (PC) 30 people (*36 people) wearing the CGMS patch for 12 weeks and checking HbA1c three times
• Intervention group (IG) 30 people (*36 people) using CGMS patch for diabetes, AI diet app, and checking HbA1c three times (* Actual number of recruits considering dropout)

This study's aim is to identify dietary factors for each patient that can affect blood sugar control based on the Continuous Glucose Monitoring System (CGMS) in patients with T2D and to personalize the amount and type of meal intake. It is planned to confirm the effect of blood sugar control intervention through a nutrition counseling coaching system.

The sex-stratified random assignment method was used to recruit 20, 30, and 30 study participants in the control group (negative), control group (positive), and intervention group, respectively. Namely, each study group is allocated according to the endocrinology department visiting patients, so the ratio of participants by gender is similarly allocated.

The group allocation of subjects uses a double-blind method in which a researcher who is not at all related to research planning, analysis, and interpretation of results assigns subjects according to random sampling numbers.

This study aims to identify patient-specific dietary factors that can affect blood sugar control based on the continuous glucose monitoring system (CGMS) in T2D patients and to personalize the amount and type of meal intake.

[Research method]

<study arms>

• Negative Control group (NC) 20 people (*24 people) checking only hemoglobin A1c (HbA1c) three times (at the beginning of the study, after the 1st 12 weeks, and then after the 2nd 12weeks)
• Positive control group (PC) 30 people (*36 people) wearing the CGMS patch for 12 weeks and checking HbA1c three times
• Intervention group (IG) 30 people (*36 people) using CGMS patch for diabetes, AI diet app, and checking HbA1c three times

The sex-stratified random assignment method was used to recruit 20, 30, and 30 study participants in the control group (negative), control group (positive), and intervention group, respectively. Namely, each study group is allocated according to the endocrinology department visiting patients, so the ratio of participants by gender is similarly allocated.

The group allocation of subjects uses a double-blind method in which a researcher who is not at all related to research planning, analysis, and interpretation of results assigns subjects according to random sampling numbers.

[Research design]

Parallel design with three arms

◎ Effect evaluation criteria

• Based on the time of participation in the intervention study, the effect of CGMS and diet app intervention on blood sugar control in T2D patients will be evaluated by comparing HbA1c at each time point of CGMS intervention and AI diet intervention.
• In order to evaluate the difference in blood sugar control according to the use of CGMS, continuous blood sugar records of all CGMS users (PC + IG) for the first 12 weeks will be downloaded daily through the AI diet app and checked to observe irregularities in blood sugar control during the day. In addition, by checking the cumulative data for a week or a month, individual changes in blood sugar will be recorded and evaluated.
• To evaluate the blood sugar change according to the intervention using the AI Diet App, IG users' meal records will be downloaded every day, the meal characteristics according to the blood sugar change of CGMS will be identified, and the GL (glycemic index) or GL (glycemic load) will be calculated.
• This process includes counseling in which a professional nutritionist sends a text message for individual blood sugar control meals based on an AI diet app and a CGMS app. In addition, the investigators will evaluate the effect of these interventions on glycemic control by measuring HbA1c
• The effect of dietary intervention without CGMS for another 12 weeks after the first 12 weeks with CGMS is evaluated in T2D patients by measuring HbA1c before and after implementing only dietary intervention.

< Data items to be collected >

(1) Surveys three times*

• Demographic information: abbreviation of name, gender, date of birth, occupation, education, income, marital status, number of children, female history (menopause, uterine/ovarian removal)

• Disease history*: comorbidities, complications, antidiabetic drugs, duration of illness

• Nutrition practices*: Number of meals/days, whether or not to eat out, dining out location, main food intake

• Lifestyle*: sleep, smoking, alcohol consumption, exercise, physical activity

• Diabetes Self-Management Activities*: Execute meal plan, exercise, blood sugar test, diabetes medication, foot care (In the case of *, it is conducted at the beginning, after 12 weeks, and at the end.)

  • App: Daily meal information
  • Blood sugar: Continuous blood sugar information through CGMS patch

<Consideration of safety for research participants>

1. Basic procedures to secure research ethics

   This study complies with the Declaration of Helsinki and ICH-GCP and will be conducted after IRB approval.

2. Personal information plan for research subjects

Personal information that can be collected in this study includes name, initials, sex, year of birth, contact information, and account number.

Personal information is used for research through deletion or anonymization and is only accessible to authorized researchers.

Individual identification through the use of the data built up is impossible, and the information obtained from the research will not be used for purposes other than confirming or researching the occurrence of diseases in the future.

<Research participant incentives>

Negative control group: 40,000 won for clinical participation Positive control group: 60,000 won for clinical participation Intervention group: 100,000 won for clinical participation

Inclusion Criteria:

- Patients diagnosed with T2D who visited Kangwon National University Hospital between August 2023 and March 2024.

(Among outpatients between 18 and 69 years old who can continue to participate for six months)

• Individuals who can use an IoT device to check blood sugar in real time and send photos using a mobile phone and have no communication barriers
• Subjects who actively need dietary and nutritional counseling from their doctor due to poor blood sugar control with glycated hemoglobin (HbA1C) of 6.5% or higher

Exclusion Criteria:

1. Patients with severe diabetes and have a lot of medication changes
2. Visitors to the emergency room of Kangwon National University Hospital
3. Patients with infectious diseases, including COVID-19
4. Vulnerable subjects (pregnant women, terminally ill patients, those who cannot communicate, etc.)
5. Patients with adverse reactions (allergies) to the Libre Continuous Glucose Monitoring System (CGMS)

Criteria for suspension and elimination

1. Patients who withdrew their consent to the study during the study period or after the study was completed
2. Patients who need to change diabetes medication during the study period
3. Patients who arbitrarily stopped using the diabetic CGMS patch during the study period or did not participate in the meal survey app for more than 7 consecutive days

Dropout Prevention Measures The dropout rate is expected to increase in the order of negative control group < positive control group < intervention group. In general, the dropout rate when selecting subjects in clinical trials is about 10% to 15%, but the dropout rate of subjects in this study is as high as 16.7%. Therefore, in order to prevent the dropout of the intervention group, which is expected to have the highest dropout rate, real-time monitoring is conducted on the collected blood glucose data. We are going to prevent the dropout as much as possible by proceeding in the direction of encouraging the target.