Neonatal and Birth Risk Factors for Type 1 Diabetes Mellitus: Prediction Using an Artificial Neural Network

Area of Science:

• Pediatrics
• Endocrinology
• Data Science

Background:

• Type 1 Diabetes Mellitus (T1D) is a complex autoimmune disease with potential links to early-life environmental factors.
• Understanding neonatal and perinatal influences is crucial for identifying at-risk populations.

Purpose of the Study:

• To investigate the impact of neonatal and perinatal factors on T1D risk.
• To evaluate the efficacy of predictive models for T1D risk assessment.

Main Methods:

• Analysis of a cohort of 327 patients using correlation, General Linear Model (GLM), and Artificial Neural Network (ANN).
• Key factors examined included Apgar score, birth weight, feeding type, sex, and delivery type.

Main Results:

• T1D patients exhibited higher birth weight, lower Apgar scores, formula feeding prevalence, and increased cesarean deliveries.
• GLM identified nutrition type, birth weight, and Apgar score as significant predictors (R² = 0.88).
• ANN demonstrated high accuracy (84.2%), AUC (0.95), sensitivity (89.1%), and specificity (76.7%) in discriminating risk.

Conclusions:

• Early-life factors significantly correlate with T1D risk, with complex interactions identified.
• Developed predictive models, particularly ANN, show strong potential for early T1D diagnosis and personalized management.
• Further validation in larger, independent datasets is recommended to confirm predictive performance.