JMIR Publications

ABSTRACT

Background:

Diabetes is a chronic condition that necessitates regular monitoring and self-management of the patient's blood glucose levels. People with type 1 diabetes (T1D) can live a productive life if they receive proper diabetes care. Nonetheless, a loose glycemic control might increase the risk of developing hypoglycemia. This incident can occur due to a variety of causes, such as taking additional doses of insulin, skipping meals, or over-exercising. Mainly, the symptoms of hypoglycemia range from mild dysphoria to more severe conditions, if not detected in a timely manner.

Objective:

In this review, we report on innovative detection techniques and tactics for identifying and preventing hypoglycemic episodes, focusing on type 1 diabetes.

Methods:

A systematic literature search following the PRISMA guidelines was performed focusing on the “PUBMED”, “Google Scholar”, “IEEE Xplore” and “ACM” digital libraries to find articles about technologies related to hypoglycemia detection in type 1 diabetes patients.

Results:

The presented approaches have been utilized or devised to enhance blood glucose monitoring and boost its efficacy to forecast future glucose levels, which could aid the prediction of future episodes of hypoglycemia. We detected nineteen predictive models for hypoglycemia, specifically on type 1 diabetes, utilizing a wide range of algorithmic methodologies, spanning from statistics (10%) to machine learning (52%) and deep learning (38%). The algorithms employed most are the kalman filtering and classification models (SVM, KNN, random forests). The performance of the predictive models was found overall to be satisfactory, reaching accuracies between 70% and 99% which proves that such technologies are capable to facilitate the prediction of T1D hypoglycemia.

Conclusions:

It is evident that CGM can improve the glucose control in diabetes but predictive models for hypo- and hyper- glycemia using only mainstream noninvasive sensors such as wristbands and smartwatches are foreseen to be the next step for mHealth in T1D. Prospective studies are required to demonstrate the value of such models in real-life mHealth interventions.