Direct-to-Consumer Testing: Relevance for Diabetes Therapy

Abstract

The performance of laboratory measurements by the people with diabetes (PwD) themselves (“direct-to-consumer testing”; DTCT) is conceptually not new for diabetology. The number of parameters for which such tests are available might increase in the next years, ie, it might go beyond glucose and glycated hemoglobin (HbA_1c). One has to consider several pro and con arguments for DTCT. Appropriate training of the users is a key issue for the meaningful usage of DTCT. Artificial intelligence (AI) might change the way the results of such tests might be used in daily life for the optimization of diabetes therapy. The decision to what extent DTCT will be used will not be made by the health care professional (HCP) but by the PwD themselves. If they see a relevant advantage for themselves in DTCT, this option will become popular.

Introduction

There have recently been several publications on a new trend in laboratory medicine, namely the performance of laboratory measurements by people with diabetes (PwD) themselves (referred to here as “consumers”).^{1 -4} Such self-measurement of laboratory parameters by the user (“direct-to-consumer testing”; DTCT) is conceptually not something new for diabetology, but one wonders whether this trend is relevant in a broader population. Corresponding products are also presented at trade fairs, such as “Consumer Electronics,” which takes place every January in Las Vegas. The relevant companies see new sales markets here, also driven by the desire of many people to monitor as many of their bodily functions as possible on an ongoing basis.

When self-measurement of the capillary glucose concentration using test strips (self-measurement of blood glucose [SMBG]) (actually an early variant of DTCT) by PwD was introduced around 50 years ago, this represented a revolution. The fact that a PwD himself should be able to determine a laboratory parameter with a sufficient measurement quality and, above all, derive and implement therapeutic measures from it, which can already have serious and immediate consequences, was viewed extremely critically by many diabetologists and was therefore vehemently rejected. Today, such discussions seem like the darkest Stone Age, but they were only a few decades ago. The reality today is that PwD can measure the glucose concentration in a capillary blood sample with a small measuring system in a very short time and with a quality that was previously reserved for good laboratory systems. In this sense, the use of continuous glucose monitoring (CGM) systems for diabetes therapy, especially in PwD with type 1 diabetes and those with type 2 diabetes who are on intensified insulin therapy, is also a DTCT application. The DTCT products (like CGM systems) are also being utilized in persons without diabetes for monitoring glucose trends. This has been studied in the population of those people living with prediabetes in regard to the benefits of delaying progression to type 2 diabetes mellitus. The CGM is also now widely utilized by those without diabetes/prediabetes who are targeting optimizing their health through lifestyle/behavioral modifications.

The diagnostic option “glucose measurement” is probably by far the most frequently performed “laboratory measurement” in the context of diabetes therapy; however, there are other laboratory parameters that are measured frequently, like the glycated hemoglobin (HbA_1c). To date, there are probably very few PwDs who carry out HbA_1c measurements themselves under home conditions (ie, with capillary blood samples), presumably also because of the associated costs. Other parameters of interest are the measurement of lipid parameters (also in capillary blood samples) and ketone bodies, although this is done more for safety reasons and can also be done continuously. ⁵ Ketone bodies are often also monitored by the “Health-inclined” person targeting ketosis as part of a dietary plan. The measurement of antibodies, C-peptide, insulin, and so on does not fall within the target area of DTCT, as they require the collection of venous blood samples.

This commentary aims to critically examine the pros and cons of DTCT and discuss why this trend is relevant for diabetology. Many aspects can be viewed from both perspectives. There may be “killing arguments” for or against DTCT for some readers; these may be due to their position in connection with the diabetes therapy of the respective PwD.

Pro

• Independence of the users: ie, the PwDs can carry out the measurements when it suits them/is considered necessary. The PwDs do not have to come to the doctor’s office for a glucose measurement, something that was taken for granted before the introduction of SMBG.

• The immediacy of the measurement results: no blood samples have to be sent to a laboratory, with corresponding delays in the availability of the results.

• Pre-analytical handling: no (suitable) test tubes need to be kept on hand with the need to handle, store, and ship the blood samples accordingly.

• These are also positive aspects from an environmental point of view.

• Cost: less cost/time for the PwD, they do not have to make an appointment at a practice and visit them to initiate the laboratory measurement. The CGM can be more cost-effective than fingerstick blood glucose measurements in PwD who are utilizing a significant number of test strips per day (albeit this may vary by country). In the United States, CGM can be a cost savings compared to, eg, checking fingerstick blood glucose measurements four times daily.

• Quality of life: CGM improves the quality of life for the PwD in regard to reduced fear of hypoglycemia with self-monitoring glycemic trends. ⁶

• Improved care: remote monitoring of CGM is a benefit for patient-provider improved care. The CGM can be reviewed by the provider as a billable service up to once monthly and greatly aid in overcoming the clinical inertia of improving glycemic care. No longer waiting three months for HbA_1c if learning to incorporate glycemic targets such as GMI (glucose management indicator) and TIR (time in range) as targets for the PwD.

• Frequency of measurements: the measurement frequency of a parameter (or a combination of different parameters) can be adapted by the PwD directly to their current needs situation. Responding to a current need is an option that otherwise requires potentially time-consuming interaction with the diabetes team. In a hectic practice, it can be difficult to reach a suitable member of staff in the event of an acute need. It is necessary to describe/explain the current situation to them, which then leads to a trip to the practice, etc. This can be time-consuming and inappropriate for the specific situation of the PwD.

Cons

• The overall situation of the user: the measurement of a single laboratory value without taking into account the overall status of the user and their current medical situation can lead to incorrect decisions by the PwD.

• The user may make treatment decisions without fully understanding their medical implications.

• They come to the practice (or then to the hospital) too late because they are too confident in their ability to assess the situation.

• Accuracy: the accuracy of home tests is likely to be slightly below that of lab tests but acceptable in most situations.

• Costs: who bears the costs for the actual measurement with an appropriate product? The user himself or his health insurance company? The costs for the individual measurement are probably (significantly) higher than if hundreds of measurements are carried out on a large laboratory machine, whereby this must be clarified, ie, how do the costs for the measurement compare with the total costs charged to the health insurance company by the laboratory/doctor?

• Environmental considerations: from such aspects, the production, distribution, use, and disposal of the amount of “product” for a single measurement (or a small series of measurements) is probably significantly more problematic than the use of a single test tube for measurement in a central laboratory. The ecological footprint of DTCT measurements should be compared to those in a central laboratory, whereby some aspects that are also of considerable relevance from an ecological point of view, such as less travel, must be taken into account in such complex considerations.^{7 -9}

• Measurement frequency: this can be adapted by the user to their acute needs (which is a clear pro), but the availability of such an option can also lead to measurements being taken much more frequently than is indicated/necessary. In cases of measurement overuse, the measurement frequency also influences the costs. As with other options for monitoring body options through wearables, this can lead to a kind of hypervigilance or obsession with having knowledge of glycemic data at all times. ¹⁰

• Quality management: a critical question is the quality of the measurement and how the resulting data are recorded and analyzed. If the products used in DTCT are manufactured elsewhere in the world and transported and stored over long distances under unclear conditions, how reliable is the quality of the measurement? Is there any kind of monitoring of the measurement quality, also in terms of the guidelines of global and country-specific medical associations and regulatory agencies? It makes no sense if the requirements for the measurement quality of laboratory parameters in the professional environment are constantly being increased to provide the best possible care for the PwD from a medical point of view, if, on the contrary, a gray area develops that can evade monitoring because users order such products online and have them delivered to their doorstep.

• Mental health implications: for some patients having access to, eg, glucose data are anxiety-provoking. More so if they have not had appropriate education as to expectations for targets (TIR, % of hyperglycemia that is expected in time above range, etc).

• The question is whether these risks are not outweighed by other advantages, such as the immediate availability of the information associated with a measurement result (eg, increased ketone body concentrations).

Information and Training

If in the future more “laboratory measurements” are carried out by the users themselves (which is to be assumed), then the provision of adequate information on the measured parameter is essential. This requires that the information is prepared in such a way, also using modern media, that even laypersons can interpret the measured value correctly, which may also mean that they consult a doctor. As can be seen in the diabetes area, PwD with a chronic disease can very well deal adequately with self-measured readings and optimize their diabetes therapy. Good training is essential for this.

Artificial Intelligence

A whole series of the aspects discussed so far and their evaluation are based on how diabetes therapy is and has been carried out up to now; the increasing use of artificial intelligence (AI) will probably lead to significant changes sooner rather than later. If the PwD receives a measurement result through a DTCT approach—which is probably available in digital form in many, if not all cases—then this can be further used/processed directly by AI, through so-called “Clinical Decision Support Systems” (CDSS) by the diabetes team or through “Patient Decision Support Systems” (PDSS) by the PwD themselves.

The AI does not view the measured value in isolation but can see it in a much larger context, ie, together with various other measured values of the respective user and other relevant information, such as their movement behavior and the ambient temperature. Furthermore, it can be linked to the user’s entire medical history and current medical knowledge by taking into account the relevant databases. At the same time, wearables can provide the AI with information about the user’s vital signs, which are recorded non-invasively and continuously. Based on this wealth of information, the AI can provide the user with specific instructions for action within seconds; it can also initiate other activities if necessary, such as emergency medical services.

In terms of improving the quality management of the individual user, an AI can achieve a level of quality in the long-term monitoring of PwD that a diabetes team can hardly offer, even between visits. The option of remote monitoring of CGM can help to overcome therapeutic inertia for optimizing diabetes therapy. ¹¹ However, whether it can always adequately recognize unusual situations and constellations and act appropriately is not guaranteed. Even if this is not always the case, in many situations, AI will act as well, if not better, than a human can, as is the case with autonomous driving. Here, media coverage (= our perception) focuses on such “mistakes,” although humans also make such mistakes, probably even more often than AI.

Given the speed of development of AI and the extent to which it is used in our world, including in a professional sense in laboratory medicine, ie, to recognize patterns and critical situations, these considerations may fall short. Even 10 or 20 years ago, we could not have imagined the extent to which we interact with smartphones today, the extent to which they have already taken control of our lives in a way that we could not have imagined in the past and which we would only have laughed at. In our modern digital world, many tasks are and will increasingly be managed by the user without us even thinking twice about it. Each of us takes the digital handling of our bank account by ourselves for granted; in the past, this was associated with a visit to “sacred banking halls,” where the bankers graciously entered figures into small savings books and received or paid out money. These are all things that we can hardly imagine today and that have massively changed the world of banking.

Summary

Ultimately, the decision to use DTCT will be made by the PwD themselves, as there have been several misjudgments by “experts” in recent decades where users have “voted with their dollars” for or against a certain option, eg, with a certain CGM system. In this case, the practical advantages (no more finger pricking, easy handling) were so predominant that this product is now the market leader. If users see a relevant advantage for themselves in DTCT, this option will become established in the longer term, presumably relatively unaffected by what medical professionals think and feel about it. It will require an open approach to the topic of DTCT by health care professional (HCP), otherwise, there is a risk of missing out on a relevant development.