@giftiger wunsch
My previous response was perhaps to short. So, to elaborate, the major problems in glucose monitoring are not the methodology for measuring glucose in a sample of plasma or blood. The methods we have are low cost, quick, precise, accurate, easily automated, and reasonably immune from interference. They have been in use for so long that most interferences with the principal methods are assumed to be known (obviously unprovable).
The main problems in management of patients are specimen collection and compliance. For the latter, hemoglobin A1c is most common today. There are some very interesting studies on compliance that show patients with diabetes, particularly those with Type I, play games with the glucose results and their insulin doses. It is a complex subject involving psychology, biochemistry, immunology, and genetics and well is beyond the scope here. Remember, diabetes in clinical practice is not simply "not enough insulin."
As for specimen collection, that is an area of great research activity. Everyone wants a less painful or non-invasive procedure for monitoring, and those will likely happen. There has been a lot of publicity about using the forearm for samples. That has simply not proved to be the best option for all patients. Methods using sweat and saliva continue to be worked on, and I believe a modified sweat method has been approved in the US by the FDA. Something similar to a pulse oximeter but for glucose is probably not too far distant, at least there is hope for it.
John