Future of Diabetes Technology

In the past two decades, we have seen a tremendous progress in the whole field of diabetes technology (DT), with its many different areas of research and development. When we look into the future, we have to be careful not to simply reflect our thoughts about the past, but to see which challenges are in front of us and which activities we have to initiate now to be able to handle these adequately. Paying attention to these should help us to transfer the opportunities that DT offers into pronounced clinical benefit of patients with diabetes. I’ll only briefly discuss different relevant topics subsequently that might initiate more detailed discussions and hopefully leads to specific actions.

Young Researchers

In several countries, the number of endocrinologists/diabetologists is decreasing, despite the increasing need for such specialists that is driven by the ever-increasing number of patients with diabetes. One should be concerned about the “despecialization” of the specialists who typically initiate DT usage.

One reason for this decrease in the number of specialized physicians is that the number of academic centers with a focus on diabetes is decreasing in the United States, Germany, and other countries. Therefore, the chances to get a high-level training in diabetes research in general and in DT research more specifically is low. In addition, the number of centers—might they be academic or more independent—that have a focus on DT in general is quite low worldwide, probably less than the number of fingers on one hand. This is somewhat annoying in view of the medical need and a market that has a size of several billion dollars/euros. If only a small portion of the money spend for DT would be invested into academic sites and training for young researchers, clinical studies, and outcomes research, a lot of ground and considerable benefit would be gained. Such an “investment” is needed and doable.

Without an adequate number of good young researchers coming from the medical field—many of those who develop technology come from nonmedical fields like engineering, computer science, and so on—it will be difficult to continue the clinical development of new products in the same speed as we have seen it in the past years. In addition, transfer of knowledge from the research level into clinical practice also requires well-trained physicians. Measures to improve this situation include:

• - Influence academic training during medical school, residency, and fellowship studies, increase the attention for treatment of patients with diabetes with a focus on DT

• - Establish additional academic centers that specialize in DT, including political activities to make legislators more aware of the need for such centers

• - Increase the funding of research activities in the DT area by devoting one-tenth of a cent of each dollar spent for DT to scientific research in this area; this would also enable more company independent research

Medical Associations and DT—What Is Their Role?

Support by diabetes medical associations is clearly needed to increase DT training and funding for independent research. However, national and international associations do not appear to have a clear interest in DT and many academic colleagues do regard DT research as “real” science. However, the associations might have realized that they are missing the train as documented by the high number of participants in DT meetings while those attending the large meetings is decreasing (see below).

Recommendations for diabetes and endocrine associations include:

• - Establishment of internal task forces to address aspects of DT that should be more actively addressed

• - Each task force should report directly to upper management in the association

Conferences About Diabetes Technology

In the past 10 years, specialized meetings like the Diabetes Technology Meeting (DTM; https://www.diabetestechnology.org/dtm/) in the United States and the Advanced Technologies and Treatments for Diabetes (ATTD; http://www.attd2017.com/) meeting in Europe (plus national meetings like the DiaTec in Germany; http://diatec-fortbildung.de/) have seen not only a steady increase in the number of participants, they also represent a platform for communication in all aspects when it comes to DT. One reflection of the success of the DTM and ATTD is the strong interest of the industry in DT as it becomes a key ingredient for improved diabetes care. Companies less directly connected to DT, such as insulin manufacturers, are also increasingly supportive of DT meetings.

The internal focus of various DT-oriented meetings differs. The DTM is more focused on a scientific exchange between academic scientists and researchers working for companies, whereas the ATTD is more focused on training of diabetologists from around the world. This is also reflected by the fact that at one meeting there is an industry exhibition (the ATTD) whereas there is none at the other (DTM). Also, only during the ATTD a number of sessions are organized by manufacturer that are interested in presenting their products to diabetologists that might use them in daily practice.

As an adjunct to the DTM, the Diabetes Technology Society (DTS) in the United States is a relatively small organization that organizes a series of additional small conferences dealing with specific topics that do also relate somehow to DT (eg, cybersecurity, digitalization, DT in the clinic). Such meetings enable also a more in-depths exchange of discussion and thoughts about current topics; they also offer the opportunity for some kind of political discussions. In principle, this is more the domain of academic meetings organized by medical associations.

In contrast to DT focused meetings, the number of presentations that are DT related during the annual meetings of the EASD and ADA remains relatively low. These large meetings have a primary focus on basic research and clinical studies related to drug treatment of patients with diabetes (such meetings are heavily funded by pharmaceutical companies). For example, during the ADA literally hundreds of posters about new SGLT-2 inhibitors, and others, are on display, reflecting the massive interest of the pharmaceutical industry in such developments. Only a relatively small number of DT related sessions and posters appear at these meetings.

The number of participants in these larger conferences appears to be decreasing in recent years and may continue to diminish, driven by a number of factors, including the decreasing number of endocrinologists/diabetologists and restrictions placed on the pharmaceutical industry for bringing physicians to attend such conferences. In addition, the smaller and more focused DT meetings may better fulfill the need of diabetologists than large meetings with numerous parallel sessions with a focus on basic research, and others. Therefore, as mentioned before, the medical associations should be interested in seeking communication with the organizers of the DT meetings to lose not more ground to them.

Outcome Research

Attendees at specialized DT meetings might get the impression that everyone with type 1 diabetes will be using an artificial pancreas (AP) system by tomorrow or the next day, and that clinical decision support systems (CDSS) will make the treatment decision for the majority of patients with diabetes type 2 in the next year. However, despite high expectations, it is worth remembering that the first continuous glucose monitoring (CGM) systems were introduced into to the market some 15 years ago, and the insulin pumps that were introduced in the late 1970s are still not worn by the majority of type 1 patients in North America or Europe! It is obvious that these expectations failed by large and it would be worth a critical in-depth evaluation why this was the case. One reason why this dream has not become true was the poor performance of the first generations of CGM systems. Another one is that people are only now beginning to use CGM and other data in ways that improve glucose outcomes (see the recently published DIAMOND study). Other reasons are the relatively high costs associated with CGM and the cumbersome handling (see below).

Checking the literature reveals that in the early days of CGMs (and in a sense also still today) no systematic and independent evaluation was performed on how much patients benefit from CGM usage in daily practice when using this novel diagnostics option. So, no outcome research was performed, we rely fully on the benefits shown in randomized controlled trials (RCTs). Clearly this type of study provides the highest level of evidence when it comes to drug treatment, but diabetes technologies do not act as “mechanistic” as drugs (= metabolic effect without patient involvement). Usage of devices requires active involvement of the patients to induce a positive effect on glucose control. Several patient related factors come into the game; for example, without addressing adequate training of patients, application of a given technology might not induce any benefit at all. Therefore, I believe there is a need for both types of research, randomized controlled trials (RCTs), and those that are outcome related.

Apps

Many patients use apps for their diabetes therapy. There are many different types of apps; however, these represent a “brave new world,” these relatively small pieces of software represent a widely unregulated world. Nobody has an overview about the many different types of apps and their quality

Country Dependent Differences

RCTs performed in one country, for example the United States, might not tell the full truth when the given DT is applied in another country, that is, the benefit that can be achieved might be different in another health care system due to cultural differences or health care systems. This requires—at least to a certain extent—performance of studies in different countries/health care systems. In Germany, we have now the situation that CGM systems are reimbursed by the insurance companies; however, the hurdles we see now with the implementation into clinical practice shows that the insurance companies try to avoid the high costs associated by establishing additional bureaucratic burdens.

Focus on HbA1c

A critical aspect in such a discussion is the question, do most patients that are using DT show an improvement in glucose control? At first, we should define what doing better means: is this in any case a decrease in HbA1c numbers? In the future, it might very well be that a combination of hypoglycemia data with changes in HbA1c is used and/or postprandial glycemic excursions and/or time in certain glucose ranges. Regulatory agencies might have a different opinion on endpoints than insurance companies. In daily practice, interesting differences might show up between the interests of patients (they might want to live a life as normal as possible) and the expectations of their treating diabetologist/diabetes teams/payers. The latter ones are interested in optimizing glucose control to reduce the risk of long-term complications and thereby the high costs for the treating these. So, for them the primary outcome parameter of clinical trials is the HbA1c. But is this parameter so much of relevance for patients with diabetes, especially when it comes to those who have type 2 diabetes?

Nevertheless, the situation in principle also enables performance of an interesting historical evaluation: Hundreds and thousands of patients with diabetes (most of them with type 1 diabetes) are starting to use CGM over a relatively short period of time. To determine if this result in a change in their glucose control, the glucose control data stored in computers at specialized diabetes practices could be analyzed to determine if this technology has impacted disease management and glucose and A1c levels before and after CGM usage became more widespread. This outcome research might provide data to assist in the communication with insurance companies. However, this assures that a positive outcome from the increase in CGM usage will be demonstrated.

In a sense, another interesting experiment was performed in the past years in Germany/Europe. A system for glucose monitoring (not a “conventional” CGM) was brought to the market, the so called “flash glucose monitoring system” (called Freestyle Libre, Abbott). It appears as if one single factor, that no finger pricking for calibration is needed with this innovative product, meaning no finger pricking at all for diabetes therapy, makes this product so attractive for many patients (predominately with type 2 diabetes, but also for a lot with type 1 diabetes), that this system became a big success story on this side of the Atlantic. The lower price of this system in comparison to other CGM systems also supported its success. So, an analysis of data about the level of glucose control of the approximately 100 000 patients in Germany/250 000 patients in Europe before and after starting to use this innovative diagnostic system would be of interest.

For most patients with diabetes, optimization of glucose control (lower HbA1c values and less glucose variability) is not of utmost importance; they are more interested in reducing the risk of hypoglycemic events, that is, avoiding acute complications. In addition, all aspects of “patient-reported outcome” (PROs) should be more in the focus in the future, also in clinical studies. This would require that PROs are excepted as primary endpoints and that the studies are adequately powered for these.

The invention of medical devices that enables noninvasive monitoring of other vital parameters will not provide only data about physical activity and movement patterns, but also about heart rate, oxygen content in tissue, and so on. We’ll have to see how such a wealth of information can be used effectively. Probably a smart algorithm will suggest a patient in the future to reduce the insulin dose after intensive exercise. In reality, many patients are not “smart” enough (= not trained enough) to interpret for example self-monitoring of blood glucose (SMBG) data adequately, and may be done more efficiently by software in the future. If you are a nonbeliever in such developments, consider the success of bolus calculators in the past years. I was quite skeptical about the need for these some years ago; however, they are an integer part not only of each insulin pump nowadays, but also of blood glucose systems and apps.

DT and RCTs

One wonders if RCTs are the optimal way to show a benefit of DT in reality. Usage of a CGM system is something different than using a new antidiabetic drug or insulin. Studies with the latter products can be performed double-blind, crossover, randomized, and so on, which is more difficult with DT. We need adequate studies that take all considerations made above into account to have a fair and adequate assessment of a given product/class of devices while avoiding the risk of bias if such studies are sponsored by the manufacturer. Up to now most (to not say all) studies are sponsored by the manufacturer of DT, with the clear exception of the JDRF-CGM study. Unfortunately, neither the insurance companies nor regulatory agencies/political institutions/academic centers have invested to any meaningful degree into generating scientific evidence for usage of DT. We have a clear need for good “class studies” (= studies that address a general question and are not product related) that demonstrated quite clearly and beyond any doubts the advantage of using a given technology.

Need for Good Registers

Establishing high-quality diabetes registers, like those available for a number of years in Sweden, would provide us with data to state how many patients in daily practice benefit in a clinically relevant meaning from using DT. The Swedish register showed impressive benefits of continuous subcutaneous insulin infusion usage on the ultimate hard endpoint of mortality. ¹ A major obstacle for establishing such registers (which also means finding somebody who is willing to finance maintaining these) is the rapid development of DT. Before any hard data can be documented and published, the world has changed, that is, new device generations with improved features are available. Having such a moving target makes evaluation of DT difficult, at least as long as the focus is on a given product and not on answering a more a general question. The demonstration of such benefits in clinically relevant endpoints in class studies represents a crucial step.

If such registers get data automatically from the medical devices patients use in daily practice, this would also enable analysis of the data of the individual patients and see—for example—how often he does skip an insulin injection? Smart insulin pens (= monitoring when how much insulin is given) will be crucial to understand such a behavior better.

Recommendations are:

• - To learn more about how patients use DT in daily practice, good registries should be set up and adequate ways to finance these should be established

• - This would also help to answer the question of whether investment into widespread usage of AP systems that we might see in the next years, beneficial in daily life, yes or no?

• - Set up standards for reporting of anonymous diabetes data such as CareLink for Swedish users to the Swedish diabetes registry

Heterogeneity of Access to DT

As front-runners for DT we tend to forget that in practice in many countries (also inside a given country), access to DT is not straightforward for many patients (and the Diabetes team). This will become even more difficult with the increasing cost pressure seen in all health care systems. In reality, only a relatively small proportion (how small?) of patients has access to all options that DT provides today!

It can be assumed that manufacturers of DT have a good understanding of reimbursement situation in the different countries, as this is of high relevance for them. However, they keep such information as proprietary to not support their competitors. At the end this leads to the disturbing situation that an independent association like the International Diabetes Federation (IDF) or other independent research institutes would have to repeat this evaluation to be able to present and publish such data. If international health care organizations like the IDF or the WHO would support an undertaking, this would be of great help as on a national level. The willingness of health care politics in a given country to support such an initiative are limited for obvious reasons.

Need for an Independent DT Research Institute

In the same line of thinking it is annoying that no research institute exists that is focused on DT and medical outcomes. Main focus of the work of such an independent institute shown be on the evaluation of medical products with a focus on safety and efficacy (also after market approval, see the work done by the Norwegian institute SKUP (www.skup.nu) which evaluates medical products before and after they come to the market) and more clinical research in general. Regulatory agencies are mainly involved in the approval process of medical products and not so much in surveillance of such products once they are on the market. In the United States, the FDA is interested in a surveillance program for blood glucose meters; however, not so in the EU. An issue in the EU is that not a single agency is taking care of medical products but different agencies in each of the member countries. The improvements in the CE marking process that soon will be establish will hopefully help to make sure that only medical products with sufficient quality can come to the market in the EU.

It appears as if the number of academic sites with an interest in basic research for DT is also quite limited; that is, centers looking into new glucose sensor development, development of new algorithms for AP systems, and so on. A lot of such research is driven by the manufacturer themselves. It is good that they are willing (forced?) to do so; however, they can keep control about the progress made, and if they are for some reasons not interested in translating this into a medical product, the scientific outcome of this work might never get published. Such an institute should also be able to perform clinical DT research independently from manufacturer support (like the JAEB center in the United States; www.jaeb.org).

Need for Standards

For regulatory agencies, it would be of help to have good standards to estimate the quality and performance of DT products. We have an accepted ISO standard for BG systems; however, the FDA has decided to have a different one for the United States. Until now we have no good standards for rtCGM systems. The standards for insulin pumps (and pens) are not widely known and some (patch) pumps do probably not fulfill them.

In the same line of thinking to have established standards for data downloading and data analysis would help practical work a lot. Still today the software developed by different manufacturer blocks each other on the computers used in specialized practices diabetes using devices by more than one manufacturer. For the users having data analysis software by a number of different manufacturers is not of help, as each requires a different handling, has different options, and others. Many patients (and physicians?) do not take full advantage of the data analysis options or they don’t use these at all probably because of this. They are not interested in looking at rows of data, see no values in using them. A standard for data analysis would facilitate communication between physicians and patients, but would also allow transfer of certain information to health insurance companies for documentation purposes. Hopefully we’ll see in the not too distant future that patients have not too think about data at all, the invention of clinical support systems hopefully provide them clear recommendations on what to do therapeutically.

Need for Training Programs for Patients

Glucose control of many patients is poor despite the availability of DT that in principal should support them in improving their control. Why is this the case and how to overcome this disturbing situation? To blame this on the limited abilities of patients to use the technology adequately is not of help. It is important to acknowledge that technology per se will not do the job (with the exception of AP systems)! Patients must be trained to use DT more smartly, for example, CGM systems. There is a need for an adequate and immediate reaction of the patient for example to an alarm given by a CGM system; he must interpret the information provided and adjust his insulin therapy (and probably his behavior) accordingly. This requires development of adequate teaching programs, like SPECTRUM for CGM. ² This program is focused on translating the information about the glucose profil into meaningful therapeutic action by the patients themselves. Such training programs have to be evaluated, that means it has to be documented that the investment into training of patients with such programs lead to relevant improvements in patient related outcomes (eg, glucose control). The need for such training programs was somewhat ignored when SMBG came to the market in 1978 or CGM in 1999. Hopefully we can avoid making the same mistake with each new technology (AP systems?) over and over again?

Market introduction of AP systems most probably will not solve metabolic issues of patients with diabetes; for example, if they eat a large carbohydrate rich meal and drink a classic Coke, such extreme situations most can’t be handled by technological systems at least for a period of time.

Payers (and regulatory bodies interacting with them) are not able and willing to waste high amount of money on new devices with a clear documentation of the benefits, for obvious reasons they are skeptical against far reaching promises made by the manufacturer (“revolution in diabetes therapy”). As the devices themselves will not lead to massive improvements in glucose control, training of patients is an important topic for all parties involved.

Need for Training Programs for Physicians/Diabetes Team

How good are our health care professionals when it comes to usage of DT? Physicians and also nurses/CDEs are usually not well trained during their studies with respect to technology, more focus is on drugs, biochemistry, and so on. Many are not technology-savvy, to say the least. The rapid development of DT overwhelms many members of the diabetes team. As many of members of this are older (see above), they not be too willing to learn something new.

Devices are handled by technicians and do not have a good reputation. Most training activities are provided by the manufacturer of a given device, that is, they are product-oriented. There is a need to develop teaching courses for the diabetes team that is adequate for these.

Focus of DT on Patients With Type 1 Diabetes

It appears as if the relevance seen for DT is quite different between countries, also depending on the resources a given country is able to allocate for DT. In such discussions, most often no distinction is made between patient groups and this can lead to severe misunderstandings. For example, the idea cannot be that each and every patient with type 2 diabetes will get access to expensive technological solution like AP systems. At least with the current cost level for such solutions, all health care systems can’t cover such costs. Even with patients with type 1 diabetes (which are a much small number of patients), not each and every patient is an optimal candidate for such technologies. However, the question remains: who is willing (able?) to make a selection in this respect when a patient is sitting in front of him? Probably questionnaires can be of help to select patients with a certain confidence.

It might also be that with certain developments it is better to focus on specific patient (sub)groups, for example, on pregnant women. Such patients might benefit much more than the totality of patients. As clinical studies are usually performed with, for example, patients with type 1 diabetes, it is questionable if the study outcome/benefits can be declared for all patients with diabetes.

DT and Digitalization

The Internet and the invention of smartphones have introduced rapid change of all our lives in the past 10 to 15 years. This rapid development will continue and most probably we’ll not type on our smartphone to get a message out in 5 or 10 years from now. Treatment adjustments for most of the patients will be done by smart algorithms. Probably patients will get more detailed information about the carbohydrate content automatically at restaurants via the local network. In the Internet of things, everything talks to everything. We will simply talk to the device (which might be smaller, more integrated in our clothes, etc) and this will tell what we should do with respect to diabetes treatment (if this is still necessary). All data will automatically into the Cloud (or however this will be called by then), are fully analyzed and interpreted immediately and clear messages are forwarded back to the patients.

This will also have an impact on how a diabetes practice will look like in the future. Most probably it will become an “AP service center.” Mostly patients with type 1 diabetes, pregnant women and patients with type 2 diabetes with special issues will be treated in such practices. The “average” patient with type 2 diabetes will be treated by his general physician; however, the level of treatment/outcome for many of these patients appears to be mediocre. Probably DT can be of help at this level as well. Patients will often not show-up in the practice in person but via modern communication channels for their visits (“doc on demand”).

Thus, digitalization and probably other “disruptive” technologies will change diabetes treatment in general. One such development can be smart insulins; if the release of such insulins is automatically adjusted to the current needs under all circumstances and without safety concerns, it might be that many—if not all—of the current discussions come to an end from one day to the other (I’ll retire immediately in this case). In a sense this would bring diabetology to an end as a medical discipline.

In the future, also aspects like screening for diabetes might be handled differentially. This can be done at a Walmart or pharmacy around the corner.

Human Factors

The FDA now requires a thorough evaluation of human factors with each new medical product; however, this is not yet an important topic for market approval in the EU. Initially manufacturers complained about this additional regulatory burden, but it appears as if they rapidly realized how helpful it is to carefully check that products developed by engineers can be handled adequately by patients (and physicians/nurses/CDEs). If medical products including AP systems are too cumbersome to use, patients will not use them. Smartphones have impressively demonstrated how complex tasks can be simplified through optimized user interfaces when there is a clear will to do so. The message from patients is clear: keeping treatment simple is quite important!

Focus Is on Hardware, But Not on Daily Life of Patients With DT

Talking about technology, this is most often interpreted as “hardware” only, that is, devices that you can take in your hands, work with, and so on; however, this has at least some other aspects:

• - Psychosocial factors—are patients using the different technological systems at all/adequately? They want lifestyle products, not medical devices. It should not be obvious that they are “ill” by having a number of devices attached to them. Discretion is important!

• - Software—a lot is behind this single word, regulatory aspects, many different technologies, and so on, so telemedicine (= doctor on demand), mHealth, big data, and so forth.

Upcoming Resources for DT

In the past decades manufacture of medical devices develop new devices and brought them to market, this will not be sufficient in the future anymore. They will not only have to document that their product is of true help to patients, that the investment into a given technology pays off at the end, and so on, but that the given device fits into the whole therapeutic scheme of the patients. Different technologies have to be brought together in a holistic approach to support the patients ideally in their individual set of requirements. In view of this and the size of the diabetes market it is obvious that new manufacturer will come to the market. The massive interest of companies like Google, Apple, and others in the health care sector and especially in diabetes already started to change the situation. Such companies not only simply have more resources, they also have a different way of acting and pushing developments forward.

Telemedicine

One of the greatest roadblocks for the further development of novel approaches like telemedicine (TM) is the lack of reimbursement. This in turn mean that the evidence for using one of the many different technologies that we somewhat summarize when we say TM (or mHealth) is simply missing. Most of the companies interested in TM are simply not able to perform “clinical studies” that can provide the scientific evidence required by health care insurance/payers to reimburse the given approach. Again, the issue is that providing evidence is a time requiring business, that is, it takes years and a lot of money. At the same time technology is evolving like “hell,” a given approach might be outdated not being attractive any more for the patients when the all steps for getting market approval/reimbursement have been fulfilled. How can we solve this catch-22 situation?

Costs of DT

Many patients can’t afford to the costs or copays for expensive technology. Our knowledge about the situation in this respect in the different health care systems (see above) is quite scare when it comes to DT. This is somewhat fascinating to see as one would expect that patients with diabetes have the same needs everywhere on the globe.

DT comes along with many promises, the question is, what do we deliver at the end? It will have to be shown in good cost-benefit models that DT provides better service and outcomes to patients for lower costs. As such economic considerations are key in our modern world, this will require even more attention in the future. The systematic and critical performance of outcome research would not only help to understand to which extent patients in daily practice benefit from all the progress made with DT, it will also strengthen the communication with health care insurances when it comes to costs.

How to Accelerate Progress of DT Development?

Have we made all the progress in the past 10 or 20 years that in principle we could have made? I believe it would be of interest to analyze critically the different areas of DT development over time and see which hurdles slowed the progress. Quite often each of the parties blame it on the other ones, preferably on regulatory agencies. At least when it comes to AP systems the speed with which the FDA approved these speaks somewhat against this “habit.” The level of communication that was established by, for example, the JDRF but also by the DTS has surely supported the whole approval process. I envisage that establishing Round Tables for different topics that bring all relevant parties together can be of help in the future to speed up development. As regulatory agencies have limited knowledge about DT and clinical needs, they can benefit from such an approach as well. Talking not blaming each other, try to understand the different needs and points of view, will be helpful in achieving the best for the patients with diabetes.

Patients (or the parents of kids with diabetes) are not willing to wait for years until products like AP systems become available commercially while all components are accessible already today; they want solutions now (www.nightscout.info; openaps.org).

Summary

For somebody like myself it is impressive to see how many different aspects are relevant to make optimal use of DT. It is not simply applying a given technology and the job is done (“fire and forget”), without taking all the different aspects into account (at least some of them are discussed in this editorial); one should not be surprised if the intended outcome is not achieved. Hopefully my comments act somehow as a “call for action” to follow on some of them in a systematic and structured manner.