Math-Physical Medicine

NO. 102

Sensor glucose behavior pattern analysis using GH-Method: math-physical medicine methodology

Corresponding Author: Gerald C. Hsu, eclaireMD Foundation, USA.

This article summarizes the author’s research findings on 33,523 glucose data collected via a continuous blood glucose monitoring device (CBGM) applied on his arm (Sensor).  The emphasis is to study glucose moving patterns or glucose wave shapes in comparison with 1,688 glucose data collected via finger-piercing and test strip (Finger) using correlation coefficients ( R ) from time-series analysis.  The dataset is provided by the author, who uses his own type 2 diabetes metabolic conditions control, as a case study via the “math-physical medicine” approach of a non-traditional methodology in medical research.

Math-physical medicine (MPM) starts with the observation of the human body’s physical phenomena (not biological or chemical characteristics), collecting elements of the disease related data (preferring big data), utilizing applicable engineering modeling techniques, developing appropriate mathematical equations (not just statistical analysis), and finally predicting the direction of the development and control mechanism of the disease.

Sensor data were collected over 453 days from 5/5/2028 – 7/31/2019 with 74 data each day.  Fasting plasma glucose (FPG) has 11 counts per day and PPG has 36 counts to cover three meals.  The remaining 27 counts are pre-meal and pre-bed.

The author used time-series analysis to calculate the averaged value of a dataset (wave or curve) or correlation coefficient between two datasets to study their pattern comparison.  He utilized daily data to calculate a more accurate average and 90-days moving average to observe the curve’s moving trend.  For each dataset, he used two different time windows: 5/5/2018-7/31/2018 and 8/5/2018-7/31/2018.

As observed in Table 1, the glucose values are quite different between Finger and Sensor with Sensor data on the upper end.  However, the data differences have been discussed in his earlier publications and presentations.  With this research, he focuses on the correlation coefficients (R ) between two glucose datasets (i.e. waves or curves) by using time-series analysis.

  • Generally speaking, finger data and sensor data do not have a close relationship (i.e. low R) which means these two glucose moving patterns are not similar. The Sensor’s peak values and averaged values have shown a high R of 82%.  Furthermore, the Sensor PPG has a rising speed of 32 mg/dL over 0-60 minutes after the first bite of meal and a dropping speed of 20 mg/dL over 60-180 minutes after the first bite of meal.
  • Two FPG results show low R. This is due to the Finger FPG between 6:00 am to 7:30 am each day, while the Sensor FPG is the averaged value between 00:30 to 07:45 with 11 measurements per day.  In fact, the Sensor FPG data show its lowest period around 3:00 am and 5:00 am during the deepest stage of sleep.
  • The traditional medical community’s consensus is to measure a patient’s PPG at two hours after the first bite of food. However, from the author’s previous papers, he has already indicated that the peak of PPG occurs around 45 minutes to 90 minutes after the first bite of food (with an averaged peak time at 60 minutes after the first bite).  Therefore, in this particular analysis, there are high R (57%-70%) between Finger PPG and Sensor PPG at 120 minutes after the first bite).  Essentially, the PPG results of the Sensor peak at 60 minutes and Finger at 120 minutes have no correlation at all.  In contrast, the Finger and averaged Sensor (60-120 and 60-180) show high R as well.
  • The Sensor glucose component averaged values and contributions are – FPG: 13% at 114 mg/dL; PPG: 54% at 136 mg/dL;  Pre-periods: 30% at 129 mg/dL; and Daily average: 100% at 131 mg/dL.  One special data consists of eating fruit within the Pre-meal period: 13% at 156 mg/dL.  Fruits are important source of nutrition for the overall general health.  The author avoids eating them during his three meals in order to not push his PPG higher.  He hardly eats any processed snacks.
Figure 1: Daily glucose
Figure 2: FPG
Figure 3: PPG
Figure 4: Finger & Sensor 120 minutes
Figure 5: PPG moving speeds
Figure 6: Finger and Sensor at 120 minutes
Figure 7: Average Sensor values
Figure 8: Sensor glucose component average values and contributions
Table 1: Sensor Glucose, especially PPG, and behavior patterns

This paper has further verified the complexity of glucose behavior patterns.  There are many questions related to these two glucose results.  For example, which blood test method should be used and is it reliable?  How to deal with the fact that the PPG peaks at 60 minutes after the first bite?  How to integrate the characteristics of both Finger and Sensor together?  Therefore, the author has attempted to develop a reasonable range of glucose variances and their movement patterns in order to be more accurate in describing glucose behaviors.

By understanding glucose, this is the foundation of understanding diabetes and its numerous complications.  Once we can numerically describe glucose behaviors, we can then calculate many diabetes outcomes precisely.


  1. Hsu, Gerald C. (2018). Using Math-Physical Medicine to Control T2D via Metabolism Monitoring and Glucose Predictions. Journal of Endocrinology and Diabetes, 1(1), 1-6. Retrieved from 2018/06/JEAD-101-1.pdf
  2. Hsu, Gerald C. (2018, June). Using Math-Physical Medicine to Analyze Metabolism and Improve Health Conditions. Video presented at the meeting of the 3rd International Conference on Endocrinology and Metabolic Syndrome 2018, Amsterdam, Netherlands.
  3. Hsu, Gerald C. (2018). Using Signal Processing Techniques to Predict PPG for T2D. International Journal of Diabetes & Metabolic Disorders, 3(2),1-3. Retrieved from
  4. Hsu, Gerald C. (2018). Using Math-Physical Medicine and Artificial Intelligence Technology to Manage Lifestyle and Control Metabolic Conditions of T2D. International Journal of Diabetes & Its Complications, 2(3),1-7. Retrieved from
  5. Hsu, Gerald C. (2018). A Clinic Case of Using Math-Physical Medicine to Study the Probability of Having a Heart Attack or Stroke Based on Combination of Metabolic Conditions, Lifestyle, and Metabolism Index. Journal of Clinical Review & Case Reports, 3(5), 1-2. Retrieved from