Diabetes Technology: Many Steps From an Idea to a Product

Abstract

The development of technology, for example in the field of information technology or biotechnology and genetic engineering, has been breathtaking in recent decades. Development has also been rapid in the field of diabetes technology. However, no “technical cure” for diabetes has yet been achieved. Patients and also diabetes experts ask themselves why the availability of modern technologies does not automatically become visible in medical technology. They feel a gap between high technology and diabetes technology. However, the development and marketing of any product is an ever-increasing challenge. This will be mastered the faster; the higher the general dissemination, the higher the amortization of investments will be. It is generally necessary to understand the multiple steps from the idea of a product to its application to people.

Rapid Development of Technology in General and Diabetes Technology

In the last decades, we have experienced rapid development in many areas of technology. A symbol is smartphones, 20 years ago, these have been unimaginable. Then, in the 1990s, there were the first large, bulky, location-independent phones. In 1994, the first smartphone from IBM (developed in 1992) was available, a phone with a touchscreen and calendar function but weighing around 1 kg. In 2007, a breathtaking development with constantly new devices began with the iPhone from Apple. Today’s smartphones are mini-computers that can be used, among many other things, to make phone calls, with Internet access, high-resolution cameras, and an unmanageable number of apps that accompany us and offer data access in almost every situation of our life.

This example shows two things: first, the development appears to us to be faster than it was. If you look closely, the period from 1992 to 2007 is longer than it appears to us today. Second, it was the constant addition of innovations to the smartphone, which also makes the development seem so fast to us. At Apple alone, this has led to the 21st generation of devices to date (even if Apple presented the 14th iPhone generation recently).

From this point of view, the development of diabetes technology (DT) has also progressed rapidly. The first insulin pens were available to people with diabetes (PwD) in the mid-1980s (1983: the MADI pen from the Czech Republic, 1984: Novo-Pen 1 from Novo-Nordisk). Up to this point and for a while afterward, a syringe had to be filled up from a vial to inject insulin subcutaneously. The first devices for measuring blood glucose were available already some 40 or 50 years ago, but they were large and not portable, so they were not suitable for constant use under everyday conditions. In addition, a blood volume of several microliters was required for the measurement (eg, with the Ames Eyetone 1970), and the measurement quality was poor. The rarely used insulin pumps had the size of a small box of chocolates, into which a mounted insulin syringe was inserted (eg, Mill Hill Infuser 1976). If you look at the modern DT level, with systems for continuous glucose monitoring (CGM), Smart Pens, insulin pumps, and systems for automated insulin dosing (AID), which enable automatic adjustment of insulin administration controlled by current glucose values, then this development can also be described as rapid.

This editorial aims to highlight the complex way from an idea to a product and at how many crossroads the wrong direction might have been taken.

Perception of Patients With Diabetes

Understandably, PwD does not always feel this way: a smartphone is purely a consumer product, and the steps taken in the last 20 years to bring it to its current level are truly impressive. Medical products used for diabetes therapy, on the contrary, are about reducing diabetes-related suffering, relieving PwD from many daily burdens, and improving their quality of life. The wishes of PwD regarding technical development are specific and far-reaching. Having the speed of the general technical development in mind, it is understandable that this leads to high expectations. PwD have a feeling that DT is not evolving as fast as other tech branches, but certainly slower than one would like.

An example of a group of PwD that had such a feeling is the “loopers.” These patients build AID systems themselves because both insulin pumps and CGM systems were available at the end of 2013 in the United States, but not AID systems. Dana Lewis took an insulin pump and a CGM system, paired the two systems, and programmed a mathematical algorithm to infuse insulin in relation to current glucose readings. In doing so, she founded a movement under the motto: “We don’t want to wait” (#WeAreNotWaiting).

Behind this development, there was undoubtedly a certain criticism of the industry, which from the point of view of these PwD did not implement their wishes quickly enough. The objection to this is, among other things, that a correspondingly precise glucose measurement is necessary for reliable control of insulin delivery. This was only the case from 2012/2013 onwards, although CGM systems had been available in principle since 1999. In addition, the development of AID systems at companies such as MiniMed (today Medtronic) was already a vision when the company was founded.

This example shows that expectations regarding the development of DT are higher than the observed pace of development. Orientation toward the faster advances in some areas of the consumer goods industry does the rest. So what are the hurdles to faster availability of DT?

Many Steps Determine the Effort and Development Speed of DT

There are various steps, some of them quite extensive, to be mastered before a product is available on the market. This is already the case in general, but it becomes even more time-consuming and complex when it comes to a medical product. Clinical studies are often a prerequisite for the regulatory approval of such a product because safety guarantees must be given for its successful use. Some of these development steps are very difficult and extend the development time.

In the beginning, there is always an idea. It may be required as part of an order to the research department of a company, but it may also result from an existing solution that does not have the desired properties. Sometimes ideas arise spontaneously and are not infrequently the start of a major technical innovation. If a cool thinker is then at work, then overarching solutions can emerge. Who could have imagined that a computer without major storage media, without immediate access via an interface (eg, USB) would be accepted by consumers? Steve Jobs, however, thought of the Internet and created the iPad.

What must follow an initial idea is the elaboration of the project. If we assign the number 1 to the idea as a measure of the effort, then the project elaboration costs at least 10 times the effort. This includes sketches for a device, the examination of patent literature, thoughts about the materials to be used, the question of finished products that could be incorporated into an own device, and a literature search, not only concerning the subject matter but also to find out about possible cooperation partnerships. During this phase, the first experiments with a test device will certainly be carried out to see whether what has been conceived works in principle. An example of this is a blood glucose meter that uses a physical method, say the absorption of infrared light. One will find a corresponding light source and a light sensor and in the absorption spectrum the signals to be assigned to glucose. However, one will also find that the glucose signal is masked by interfering variables and will try to separate the glucose signal. The effort to be expended has thus increased by a further factor of 10.

In the next step, a functional sample is created, a prototype. This will have many characteristics of the final product, especially on the functionality side. In the case of measuring devices, this concerns the measuring method, possibly the sensor electrodes, ensuring the operating voltage (if necessary) or micromotor and fine gear in the case of insulin pumps. Ideas for the design are included as well as considerations for the functionality. Furthermore, this concerns the software and data transfer. What follows is the first series of devices, the so-called pilot series, with which further tests are carried out. Again, the factor for the effort is 10 to 50 times higher than in the previous stage.

The product has to be tested. Several aspects arise in the process. For example, with CGM systems, the quality of the measurement is important on one hand. From today’s point of view, measured value deviations of less than 10% compared to a reference measurement are to be aimed for. On the other hand, at the beginning of the application, the proof of safety and efficacy of CGM played a role when PwD used it to manage their therapy. For this purpose, numerous clinical studies had to prove that the users benefit from it. This had to be demonstrated by an improvement in metabolic control (better HbA1c, fewer hypoglycemic events) and/or an improvement in quality of life. This second aspect was also important in the approval of CGM systems by payers. However, it also applies to other DT products if they lead to a new form of therapy, ie, insulin pumps with which sensor-assisted pump therapy is created via coupling with CGM, or automated insulin delivery (AID systems). Other examples include smart pens supported by CGM or apps planned as digital health applications (DiGAs).

At this stage, many developments have failed. Here are some examples:

1. Several times, most recently via Google, a contact lens was announced with which the glucose in the eye water was to be measured. Because of the lack of measurement accuracy and also because it is difficult to always wear a contact lens 24 hours a day, which means that glucose measurement cannot be guaranteed all the time, the efforts were discontinued.

2. In the early 2000s, the American company Sensys Medical introduced a system for the bloodless measurement of glucose based on the absorption of infrared light (near infrared). The laboratory measurements were optimistic, and they had well calibrated the glucose sensor. When the device was applied to subjects with diabetes, it showed satisfactory measurement quality when an employee of the company managed the measurements. If the support of the test person was left to a study nurse, the results were significantly worse and quite poor if the test person had to manage the measurement alone. The problem was that the pressure of the finger on the test field, the texture of the fingertip, and so on significantly influenced the measurement.

3. The Swiss company Pendragon caused a sensation in the early 2000s with a glucose measurement system called “Pendra,” which was also non-invasive. It determined the glucose concentration from the conductivity in the environment of the body cells. After extensive calibration, the measurement results were sometimes very good, but they were not the same for the same test person on another day. Because of various other influences (humidity, temperature, etc), it was never clear when the measurement was accurate and when it was inaccurate. At the same time, the company had already CE-marked, started marketing, and even prepared to go public.

Any number of similar examples can be found, including many insulin pump projects that show how complex and costly successful development can be. This certainly increases the effort by a factor of 10 to 50.

Crucial Step: Manufacturing

The crucial next step is to create the technology for manufacturing the systems. This involves the installation of production facilities and test tracks. If large-scale production of the product is envisaged, then the interaction of a complex of equipment becomes crucial. This also includes the people who will be in charge of the equipment. An example from the microelectronics industry will illustrate this: if a new level of technology is targeted (eg, memory circuits of 50 Gbit), then at the beginning the yield is only a few percent. After about two years, this is then over 80%, although nothing has been changed in terms of technology, equipment, materials, or auxiliary materials. The reason for the higher yield is the experience curve of the overall process. Furthermore, it should be noted that the technology must meet extensive legal requirements. These include occupational health aspects, safety, environmental protection, and so on. The effort involved is thus increased once again by at least a factor of 1000 (compared to the idea, the effort has now already increased to several tens of millions of times).

Once all these hurdles have been overcome, the products still have to be approved. A CE marking is still relatively easy to obtain. The European Medical Device Regulation (MDR), which has been in force since 2021, poses a much higher hurdle. For example, in Germany, there is a fundamental need to demonstrate to payers, especially the German National Association of Statutory Health Insurance Funds (called Central association of the statutory health insurance funds; CA SHIF, in German: SV-GKV), that a therapy carried out with this technology is of benefit to patients. If this is a new form of therapy support, such as CGM or AID in the recent past, then conclusive studies must be presented. If there is any doubt, then the matter is referred to the Joint Federal Committee (JFC, in German GBA), which then often has the studies reviewed by IQEHC (Institute for Quality and Efficiency in Health Care, in German: IQWiG). These only accept statements from randomized, controlled studies over 6 months. Such studies usually provide good evidence but are time-consuming and very cost-intensive. Again, the overall effort increases.

Here is an example that shows how long such a process can take. One of the authors, at that time who was the scientific director of the company Medtronic Diabetes Germany, discussed for the first time in February 2007 with the SV GKV about the approval of CGM for everyday use by the PwD. Many studies followed, which were repeatedly presented to this committee. The SV GKV turned to the German Diabetes Society (DDG). In their Working Group on Diabetes Technology (AGDT), the positive evidence for the use of CGM was summarized in a basic publication. This was finally followed by IQWiG’s review of the evidence and its positive decision. Finally, in June 2016, after various hearings, CGM was approved for insulin-treated patients with type 1 and type 2 diabetes. This description of an approval process lasting more than nine years for a new examination and treatment method is not to be understood as a criticism of the authorities. In particular, providing the required evidence was lengthy. Also, the companies with CGM systems did not always have a uniform view on the fields of application of the method, and there was not always the same view among physicians either. All this is normal. But it shows how such an approval process can delay the availability of innovative products for years.

Once DT products have been approved, they still have to be marketed by the manufacturing company and/or its trading partners. Both the diabetes teams and the PwD must be convinced that their use will bring them a significant advantage. They will only recognize this if it not only improves diabetes control but also simplifies diabetes management in everyday life. Only then will the products become widespread and ensure the economic success of the manufacturing company.

One regrettable example is the GlucoWatch from the American company Cygnus. This was a watch with two electrodes and a pad attached to its underside. The pad contained an enzyme (glucose oxidase) for the conversion and measurement of glucose, comparable to a test strip. An ion current flowed through the skin via the electrodes, thus collecting fluid from the tissue, which also contained the glucose, and transporting it to the other electrode. The amount of glucose was measured on the pad. The disadvantage was that a pad could only measure up to 12 hours, but before that a run-in period of two hours was necessary. In addition, many users experienced current marks on the skin. Although the company was able to present a wide range of studies and achieved market approval as early as 2002 as the first CGM system to display current glucose values, commercial success failed to materialize. The GlucoWatch disappeared again.

The AID systems can be cited as a positive example. Adjusting the basal insulin delivery rate to the user’s glucose level ensures in most cases that the morning glucose values are in the euglycemic range. During the day, glucose levels are also usually within the target range, for which glucose levels between 70 and 180 mg/dL (3.9-10.0 mmol/L) are currently considered. Users of such systems very often spend more than 70% of the time in this target range, which would require enormous efforts in any other form of therapy (frequent monitoring of glucose levels, administration of correction doses if glucose levels are too high, attention to the risk of hypoglycemia) and can therefore hardly be achieved in the broad sense. AID systems, especially those with automatic correction doses, take care of this without the PwD having to intervene. The patient’s management is relieved and, in addition, the target glucose values are achieved.

Summary

A large number of necessary steps in the development of DT products that are then actually available to PwD are a challenge. If these are innovations, ie, products that are being developed for the first time of their kind, then the challenges increase to an extreme level. Only progressive companies with their large capital base or in partnership with other companies or investors can make the necessary decisions and complete the development process. It is good that this courage is present in some companies that will continue to provide new products in the future. After all, there is still plenty of development potential. However, the approval processes must also adapt to the speed of innovations.