Diabetes is a chronic disease that affects millions of people worldwide, with the number of newly diagnosed cases continuing to rise at an alarming rate. Patients diagnosed with diabetes require routine monitoring of their blood glucose levels, which is commonly done through invasive and minimally invasive techniques. However, these methods are often painful, prone to infection, and costly in the long run. Therefore, it is necessary to develop non-invasive, pain-free, affordable glucose monitoring techniques. Near-infrared spectroscopy has emerged as a promising technique among the non-invasive techniques for glucose monitoring, with several studies reporting its efficacy in detecting glucose concentration. This study designed and developed a near-infrared system based on an 1150nm wavelength for continuous non-invasive glucose monitoring. The system gave the output voltages that vary with increased glucose concentrations. It was ascertained from the Clark error grid analysis that the glucose concentration falls within region A and region B. This implies that the device is suitable for measuring glucose concentration non-invasively. As per Bland Altman's analysis, the majority of the data points were within both limits of agreement except one data point that was outside the upper limit of agreement, indicating 93% degree of accuracy, and the residual plot indicates that the regression model for the system is appropriate for predicting non-invasive glucose measurements.
Read full abstract