The dispute over the reimbursement or non-reimbursement of blood sugar test strips is apparently over. After the negative benefit assessment by IQWiG, the G-BA's final assessment was published on March 17th of this year - also negative, albeit with some restrictions. No wonder that the Roche Diagnostics information event, which has been established in the specialist media for years, mainly revolved around this current topic, on which there is also new evidence: However, it probably comes a little too late, because the BMG did not object to the decision and went to it no new study situation at all.
>> Anyone who looks at the history of blood glucose measurement (BZM) has seen a triumph since 1965, when the first test strip for blood glucose self-determination came onto the market from the Ames-Miles laboratories. PD Dr. described it as a “milestone in diabetes therapy”. Bernhard Kulzer (Diabetes Center Bad Mergentheim, FIDAM Diabetes Research Institute) in his lecture “The G-BA decision on the reimbursement of urine and blood sugar test strips and its consequences” describes this invention, which has become the basis of all modern therapy strategies over the years.
The G-BA has now spoken out in favor of restricting the ability to prescribe urine/blood sugar test strips for diabetics who do not require insulin. Experienced diabetes specialists are not just concerned about the fact that the decision was negative. Because in the justification and corresponding press release from the G-BA of March 17, 2011, a picture is conveyed from the doctor's perspective that does not correspond to reality, although it can probably be seen that way from an economic perspective.
In its justification, the G-BA says that the restriction on the ability to prescribe urine and blood sugar test strips only applies to diabetics who are not dependent on insulin because - and now comes the original G-BA - "these patients with milder forms of diabetes mellitus - This includes a large proportion of type 2 diabetics - they can get their disease under control with appropriate changes in diet, losing weight and increasing physical activity as well as taking oral medication (so-called anti-diabetics). And further: “This group of patients cannot benefit from self-monitoring of blood glucose because it has no direct consequences for therapy.”
Past reality
Unfortunately, this reasoning, even if it may even be approximately correct in fact, misses the point of reality. If it were so easy to get your diabetes under control by changing your diet, losing weight, increasing physical activity and taking oral medication, there would be around 4.7 million type 2 diabetics without insulin in Germany alone. Of these, 1.7 to 2 million are treated with insulinotropic medications, especially sulfonylurea tablets, which can trigger hypoglycemia; these patients must therefore definitely be described as being at risk of hypoglycemia.
Of course, many of these diabetics could get their disease under control with different eating habits and more exercise. But: do you do that? Not at all!
If that were the case, diabetes would not have become an epidemic and Germany would not have taken a sad leading role across Europe in the body mass index, which is often associated with diabetes. That is the sad fact of reality.
Now the G-BA speaks of urine and blood glucose measuring strips in its justification. Nobody can really answer why this is the case, because urine testing has almost disappeared from the market and, compared to BZM, accounts for far less than 1% (see p. 13) of the market.
But that is a small matter compared to the passages within the G-BA resolution, which is about “far-reaching exceptions” - which are quite positive from the patient's point of view - and which were made in coordination with the patient representatives who advised on the G-BA . Accordingly, doctors can continue to prescribe urine and blood sugar test strips if there is an “unstable metabolic situation”. This could be the case in the case of “intercurrent illnesses, initial treatment or change in therapy for oral antidiabetics with a high risk of hypoglycemia”. If this is the case, doctors can still prescribe up to 50 test strips per treatment situation.
In addition, according to the G-BA, a prescription of up to 50 test strips, which is exempt from the prescription restriction, is possible when initially switching to or changing therapy for oral antidiabetics with a high risk of hypoglycemia as part of participation in structured training programs in DMP. The G-BA does not reveal exactly when. Always and at every training session? Or only during the initial setting or when changing therapy?
Much more important, however, would be the answer to the question of what exactly constitutes an unstable metabolic state and an intercurrent illness. Even better would be the additional answer to the question: How is a doctor supposed to recognize an unstable metabolic state if his patient does not measure his blood sugar before seeing the doctor? Should he then send him back home, write out a prescription, have it measured for a few days and order it in again?
The use of terminology is also strange. When the G-BA talks about “mild” and “severe” forms of diabetes, the hair of a diabetes specialist like Kulzer begins to stand on end. The G-BA's division into “mild” type 2 diabetes and “severe” type 1 diabetes has very little to do with the reality of care. According to Kulzer, type 2 is the much more complicated disease because there are no directly noticeable consequences over many years. “There are people at work who don’t deal with the subject matter as well as they should,” says Kulzer, assessing the work of IQWiG and G-BA. He, who was present at various hearings at IQWiG, actively participated in statements and even personally spoke to G-BA board member Dr. Rainer Hess was “sustainably frustrated” that, on the one hand, it was always just about costs and, on the other hand, “scientific arguments were being put forward”.
By this, Kulzer primarily means the way in which studies were selected by IQWiG, which are included in the evaluation and which are not. And how weighting was then carried out within these evaluated studies to determine which effects were now classified as significant.
The fact that IQWiG based its negative assessment solely on the results of just five RCT studies bothers him deeply. Not so much because there is also a discussion about whether the results of these studies are clinically relevant or not, but because the type of study design has been put above the research purpose.
Kulzer is primarily concerned with the usefulness of the studies on blood sugar self-control (BZSK). In most of the studies included in the IQWiG assessment, only BZM itself was researched. Kulzer: “Without training, it makes little sense to actually measure self-control.”
Pointless without consequence
In his opinion, it should be the case that these studies - if they are used to determine the future of BZM - would at least have ensured that patients are well trained, know how to use BZM and are then able to do so to take the consequences of this measurement yourself. Kulzer: “It would actually be the minimum requirement for studies of this kind that a patient knows how he is testing, why he is testing and what he is doing after the testing.” Because otherwise testing itself would not make any sense, which would simply result in a value , which you can “give yourself as a gift without consequences”.
This is exactly the opinion of the IQWiG and consequently also the G-BA, which follows the apparent lack of evidence previously stated by the IQWiG. The fact that the RCT studies used for the evaluation usually have small numbers of cases and are also short-term studies of six, at most twelve months, also bothers the critics enormously.
In the included studies, the patients sometimes received no training at all and were only allowed to write down the test results and then hand them over to the doctor, who then drew the conclusions for them. “This is not self-management, but rather self-testing with external control,” says diabetologist Kulzer angrily. Because something like this has very little to do with the state-of-the-art of diabetes self-management and the real training situation in Germany. No wonder: The situation in international studies is not so easy to transfer to the reality of care for patients in Germany.
After all: Despite the incorrect setting, almost all studies have shown an effect, although some of them were very small, resulting in a total of 0.23 percentage points of HbA1c reduction between the intervention and the control group. However, since this value was below the usual non-inferiority margin of 0.4%, the statisticians simply assumed that this effect was unlikely to be clinically relevant. Even the only good study from Germany, in Kulzer's eyes, was only slightly higher with a HbA1c reduction of 0.44 percentage points - but still had the greatest effect included in the IQWiG assessment.
But does small automatically mean not clinically relevant? “You can of course ask yourself whether a 0.23 percentage point reduction in HbA1c is clinically relevant or not,” Kulzer also knows. For IQWiG, the difference apparently doesn't matter. Kulzer: “If a certain confidence interval is not exceeded, there is no significant effect because the statistician doesn’t care whether the difference is clinically significant or not.” But is it that simple?
Clinically relevant or not?
Exactly not. A statistician may like it, but a diabetes specialist cannot care less about the difference. This is because he can extrapolate even a minimal effect of 0.23 percentage points in HbA1c reduction, resulting in a roughly five percent reduction in all diabetes-related endpoints and an approximately eight percent reduction in microvascular complications. Kulzer: “I would definitely see that as a quite significant effect.” Compared to the effect sizes of diabetes medications, this isn’t such a bad result. Even training itself has an effect on the HbA1c value of (only) 0.8 percent, which has been very well proven in studies.
According to Kulzer, you can now begin to calculate that this 0.8 effect includes the 0.23 percentage point HbA1c reduction in blood sugar self-control, as well as 0.3 from the increase in physical exercise and another 0.1 from the increase in motivation and a further 0.3 percent reduction comes from the associated weight reduction. This means that anyone who begins to break down the positive overall effect of training, which has been well proven in studies, into individual effects will ultimately come to the conclusion that weight loss is of no use, physical exercise is of no use either, motivation can be left alone and self-monitoring of blood sugar anyway. “If you divide a system into small parts, each part does little on its own, but as a whole it does a lot,” says Kulzer. This applies even more to a diagnostic instrument like the BZSK, which in itself cannot have any effect at all as long as nothing is done with the values that can only be obtained through it.
IQWiG was made aware of all these facts - with little success. “They were all methodologists who knew very little about diabetes,” says Kulzer, disappointed. The positive results of the STeP study (Structured Testing Program) were also not included in the IQWiG assessment and G-BA resolution. According to Roche, the 6-month interim results and a synopsis were brought to IQWiG's attention in autumn 2009, but were not included in the assessment. No wonder, because for “a sufficient assessment” the IQWiG considers “at least a publication containing all the essential details of the design and results (usually in a scientific journal) or a study report” to be necessary, as the deputy director of the institute PD Dr. Stefan Lange informed us upon request.
And the G-BA only relies on the IQWiG assessment. According to Roche, there was no opportunity for the G-BA to take the new evidence into account in the decision, despite information from the Association of Diagnostic Manufacturers (VDGH) in the comment process and after the study was completed.
A few weeks ago, however, a study initiated by Roche Diagnostics was published that clearly shows that self-measuring blood sugar can actually significantly reduce the HbA1c value.
One can criticize that, in case of doubt, more evidence was provided far too late. This is only partly true, as Ralf Kai Decker, Head of Marketing & Sales Diabetes Care Germany at Roche Diagnostics Germany, explains. After all, planning the study began shortly after the G-BA commissioned the IQWiG to assess the matter on February 22, 2005. But first it should be designed as a comprehensive work by many providers of blood glucose monitoring systems; with the effect that no consensus was reached for a long time.
Until Roche finally started the study alone, acquired doctors and patients in the USA for a year, observed them for a year and evaluated them for another year. The study was then submitted to the US journal “Diabetes Care” and finally published in February 2011 - but before the G-BA decision!
And what is the BMG doing with the new evidence that is finally officially available? Answer: Nothing at all. The BMG waved the directive resolution through and only noted that it assumed that the exceptions formulated in the resolution would be taken into account.
The STeP study is, after all, a prospective, two-arm and multicenter RCT study. However, in one control arm a patient group with blood glucose measurement (BZM) was not measured, in the other one without BZM. But that, in turn, doesn't quite fit into the grid of the studies that were included in the assessment by IQWiG: The five studies evaluated were zero studies: blood sugar testing was compared with no blood sugar testing.
But Roche assumes that BZM is useful per se and therefore just wanted to find out which measurement approach is better: the measurement that has previously been common in diabetes therapy (at rather undefined times and without instructions as to what conclusions can be drawn from the values ) or a fairly new approach in the form of structured blood sugar testing. It is no longer measured every day, but rather in a “structured blood sugar measurement” on three days in a row at specific times. According to Kulzer, this should make the results obtained much more meaningful than before, because the resulting daily profiles result in an even denser data base: The patient (and also the doctor during consultations) can immediately see and visualize whether there is hypoglycemia, whether there is fasting blood sugar, etc. - and can therefore adapt behavior and therapy much better than before.
In STeP, two groups performed blood glucose self-monitoring; some are more structured than others. In the analysis (Fig. 1) you can see the effect: On average, one group was able to reduce HbA1c by -1.3, the other by -0.8. Kulzer: “That means we have a much more significant HB1Ac reduction in both groups than was stated on average in the IQWiG report.” That was only - 0.23!
It also shows that not only HB1Ac, but also the mean pre- and postprandial glucose values have improved. For Kulzer, the results are particularly interesting because the new testing approach has a positive human as well as economic effect: “Patients have to test less and yet the results are better.” There is finally what he calls “intelligent Testing” means: structured with analysis and consistency, which also “significantly reduces the daily number of tests”. <<
