Predicting Type 1 Diabetes Risk with Technology

Original Article: https://journals.sagepub.com/doi/abs/10.1177/19322968231209302

Studying Diabetes Risk

New research recently looked at how certain blood markers could indicate a higher chance of developing type 1 diabetes (T1D). Here, researchers focused on 42 healthy relatives of people with T1D, dividing them into two groups based on the number of markers found in their blood. They used a special monitoring device called continuous glucose monitoring (CGM) for a week and had standardized meals to see how their blood sugar changed. The goal was to create a system using machine learning—a kind of smart technology—that could tell apart those at low versus high risk for T1D.

Finding Diabetes Indicators

When comparing the two groups, they noticed differences in blood glucose levels after meals. The group at higher risk for T1D showed more time with high glucose levels, a wider range of glucose readings, and more variation in their glucose levels. Using a smart computer model called a linear support vector, they found it was good at distinguishing between the low-risk and high-risk individuals, with an accuracy score of 0.88.

Technology for Diabetes Prediction

This study suggests that using a special monitoring device along with smart computer systems could help predict the risk of developing type 1 diabetes. By analyzing blood glucose patterns, this technology might be a way for people to check their risk for diabetes in the future.