Novel artificial neural network-based (ANN) approach for glycemic forecasting

Description:

Project ID:  D2009-05

 

Background:

Currently, glycemic control is obtained via point of care (POC) glucose monitoring and delivery of insulin based on a defined insulin infusion protocol. This discrete monitoring is disadvantageous as glucose concentrations are unknown when POC glucose values are not obtained. Neural network based models for real-time prediction of glucose in critical care patients using prediction horizons >60 minutes have been developed. These models utilize information derived from CGM devices, POC glucose monitoring, and other medical records routinely obtained during a patient’s stay in the intensive care unit. This predictive modeling can be implemented in bedside monitoring technologies and utilized for clinical decision support, intelligent therapeutic direction, recommendation, and automated glucose control. Using this technology prediction of glucose will provide a means to enhance patient safety, care, and outcomes.     

Invention Description:

Continuous glucose monitoring (CGM) system which can provide glucose concentration measurements every few minutes for significant insight for the determination of patterns existent in daily glucose fluctuations of critical care patients. This novel neural network glucose method has been reduced to practice and test data is available for review.

Advantages:

1.     System provides better glycemic control and subsequent avoidance of complications--enhancing treatment and outcomes and reducing health care costs

2.     Integrates with continuous glucose monitoring devices

3.     Real-time model execution in user interface of medical device or bedside monitor for intelligent therapy direction and/or automation

4.     Accurate means of gauging predictive success in real-time allows sufficient lead time to adjust therapy to avoid problematic glucose values

5.     Enables medical personnel to gauge the effect of various factors on future glucose concentration

6.     Provides sufficient time to adjust therapy to avoid problematic glucose values

 

Applications:
Type 1 & 2 diabetes; post-traumatic hyperglycemia; cardio-thoracic surgery; critically ill/critical care patients; military personnel

IP Status: US Patent 8,762,306,  US Patent 9,076,107

 

Patent Information:
For Information, Contact:
Seth Smith
Licensing Associate
The University of Toledo
419-530-6229
Seth.Smith3@utoledo.edu
Inventors:
Brent Cameron
Scott Pappada
Keywords:
CGM
Continuous glucose monitoring
Diabetes
Hyperglycemia