There are new achievements in non-invasive blood sugar measurement, and the test result value has met the national standard!
In addition to high blood pressure, diabetes has become one of the fastest growing health problems in the 21st century, and my country is the country with the largest number of diabetic patients. In 2021, the number of adult diabetic patients in the world will reach 537 million, and about one in ten adults will be affected; the number of diabetic patients in China will exceed 100 million, ranking first in the world.
Blood glucose monitoring is an important means of controlling diabetes. At present, the domestic blood glucose meter market size is about 5 billion yuan, and the market space exceeds 20 billion yuan. The current blood glucose monitoring is mainly invasive blood glucose monitoring that needs to be pricked with fingers, and non-invasive monitoring is the trend of future blood glucose monitoring
Recently, the team of Nie Zedong, a researcher at the Minimally Invasive Center of the Institute of Biomedical and Health Engineering, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, has made new progress in non-invasive blood glucose monitoring technology based on physiological information. The relevant research results were published in "IEEE Transactions on Neural Networks and Learning Systems" (IEEE Transactions on Neural Networks and Learning Systems) under the title of Noninvasive blood glucose monitoring using spatiotemporal ECG and PPG feature fusion and weight-based Choquet integral multimodel approach.
The national standard "General Technical Conditions for Blood Glucose Monitoring System for Self-Testing of In Vitro Diagnostic Testing System" states that the allowable deviation of blood glucose measurement value shall not exceed ±15%. In other words, the MARD value <15% is the national standard for my country's blood glucose monitors to be qualified. The MARD value of 13.42% of the Chinese Academy of Sciences already meets this standard.
The MARD value of the non-invasive blood glucose monitoring technology of the Chinese Academy of Sciences has reached 13.42%, which has met the national listing standards for blood glucose monitoring equipment.
As an important index to measure the accuracy of a blood glucose meter, the MARD value refers to the overall average absolute relative difference of the monitoring results, and is an important indicator of whether a blood glucose meter product can be marketed. The smaller the value, the higher the accuracy
Active blood glucose monitoring is one of the important means to effectively reduce diabetes and delay complications. At present, blood glucose monitoring is mainly through fingertip blood sampling or implantable blood glucose monitoring equipment (CGMS) based on electrochemical detection technology, but these methods have disadvantages such as pain, short service life, and high cost, which limit patient compliance. Therefore, the development of non-invasive, comfortable and convenient non-invasive monitoring technology is of great significance and clinical value to promote blood glucose monitoring.
Studies have shown that changes in blood sugar concentration will stimulate the autonomic nervous system of the human body and cause changes in physiological information such as electrocardiogram (ECG) and photoplethysmography (PPG). Considering that ECG and PPG can be obtained through smart wearable devices, they have Advantages such as quickness and low cost, as shown in the figure, a non-invasive blood glucose monitoring technology based on multimodal fusion of ECG and PPG is proposed. This technology obtains the spatiotemporal feature information of the above physiological information by using numerical calculation methods and deep learning algorithms, and uses the Choquet integral algorithm based on variable weights to realize the decision fusion of different modes. Researchers used this technology to obtain data from 21 volunteers for a total of 103 days. In the 10-fold cross-validation, the proposed multimodal fusion algorithm achieved a MARD value of 13.42% in blood glucose monitoring, and A+B of the consistency error grid Area > 99%. The above achievements provide an important theoretical basis and technical support for non-invasive blood glucose monitoring based on wearable health devices and home health devices, and have broad application prospects.
