Detailed Investigation of 4 Annual Datasets from the Continuous Glucose Monitoring Sensor Glucoses Collected by a Long-Term Type 2 Diabetes Patient Using the Concept of Distributional Data Analysis to Develop a Specific Analysis Method Regarding Glucose Density Based on GH-Method: Math-Physical Medicine (No. 508)

Abstract

Gerald C Hsu

The author read an article, “Glucodensities: a new representation of glucose profiles using distributional data analysis,” dated August 19, 2020, from arxiv.org (see Reference 1). He decided to perform a research task using the Glucodensity (GD) concept but with his own developed software algorithm and collected glucose data via a continuous glucose monitoring (CGM) sensor over four pseudo-annual periods of 2018, 2019, 2020, and 2021. In clinical practice, most medical doctors use HbA1C as the golden standard to evaluate the disease conditions of their type 2 diabetes (T2D) patients. The HbA1C value represents the average glucose value of all glucoses over the past 90 to 120 days or perhaps 115 days based on the red blood cell’s lifespan; however, the A1C alone cannot tell doctors additional information other than the mean value. Other biomarkers such as the glucose variability (GV) or the glucose fluctuation (GF) can provide more data regarding the damage of a patient’s internal organs via glucose excursion which causes many diabetic complications. Furthermore, the American Diabetes Association (ADA) issued guidance on time in range (TIR), time above range (TAR), and time below range (TBR) which can offer a general idea of how glucoses are distributed in three different ranges: TIR for normal conditions, TAR for hyperglycemic conditions, and TBR for hypoglycemic conditions.

PDF