Characterization of Type 1 Diabetes Subgroup: An Artificial Intelligence Analysis of Clinical and Glucometric Features

Differentiated groups of people with type 1 diabetes defined through the analysis of clinical, analytical, and glucometric variables.

Cluster stability over time determined using the Jaccard index as a reliability criterion: persistence of clusters over 1, 2, and 3 years. The Jaccard index (JI) measures the degree of similarity between two sets, regardless of the type of elements. It takes values between 0 and 1, with the latter corresponding to complete equality between both sets

Presence of acute complications (such as severe hypoglycemia) and chronic complications (such as retinopathy, nephropathy, and neuropathy) across the different clusters.

Mean glucose reported in mg/dL

GMI (glucose management indicator) reported in percentage (%)

CV (coefficient of variation) reported in percentage (%)

Time in range expressed as percentage:

• % of time in glucose range 70-180 mg/dl (TIR) >70%
• % of time in glucose range 70-140 mg/dl (TTIR) >70%

• % of time <70 mg/dl (TBR1) <4%
• % of time <54 mg/dl (TBR2) <1%
• % of time >180 mg/dl (TAR1) <25%
• % of time >250 mg/dl (TAR2) <5%

Lab or point-of-care HbA1c

Laboratory mesured total cholesterol, triglycerids, LDL anb HDL

creatinine Laboratory measure. Units mg/dL

laboratory estimated glomerular filtration rate. Units mL/min/1.73 m2

Albuminuria, laboratory measure. Units mg/g

Use of Antihypertensive treatment and its relationship with the clusters.

Use of hypolipemiant medication (yes/no)

Association between use of hypolipemiant treatment and the clusters.

Type of insulin therapy: multiple daily injections, continuous subcutaneous insulin infusion systems, hybrid closed-loop systems

Association with the type of insulin therapy and the clusters

Weight in kilograms and its relationship with the clusters. Weight and height will be combined to report BMI in kg/m^2.

Height in centimeters and its relationship with the clusters. Weight and height will be combined to report BMI in kg/m^2.

Waist circumference in centimeters and its relationship with the clusters.

Tobacco consumption and its relationship with the clusters. Tobacco use will be reported: active tobacco use, past tobacco use, never smoker, unknown.

Acohol consumption and its relationship with the clusters. Alcohol consumption will be reported as: Low risk consumption, Risk consumption (>10 grams of alcohol in women, >20g of alcohol in men), known active alcohol disorder, Passed alcohol disorder, Unknown.

Patient age at diabetes diagnosis and its relationship with the clusters.

Diabetes duration and its relationship with the clusters.

Active pregnancy and its relationship with the clusters.

Parity status in women and its relationship with the clusters.

Menstrual cycle phase and its relationship with the clusters.

Reproductive stage in women and its relationship with the clusters. Reproductive stage will be reported as: Reproductive, Perimenopausal, Postmenopausal, Unknown

Patient-reported health-related quality of life will be assessed using a validated questionnaire for patients with type 1 diabetes. The Spanish version of the Diabetes Quality of Life questionnaire (EsDQOL) will be used. The score obtained from the questionnaire ranges from 0 to 100, where 0 represents the lowest possible quality of life and 100 the highest possible.

Correlation between patient reported health-related quality of life and the association with the clusters.

Date of birth

Sex assigned at birth

Race/ethnic background reported as: White, Mediterranean or Hispanic, African or Caribbean, South Asian (Indian, Pakistani, Bangladeshi, or other Asian), East or Southeast Asian (Chinese, Japanese, or Southeast Asian), Arab or North African (including Egyptian), Unknown