Remote Monitoring of Diabetes: A Cloud-Connected Digital System for Individuals With Diabetes and Their Health Care Providers

Abstract

Benefits of telemedicine have been proven in the field of diabetes. Among a number of technical solutions, Diabeo® has been studied in both type 1 and type 2 diabetes with intensive insulin therapy. This digital therapeutic system contains a self-monitoring glucose logbook and offers automated insulin dose recommendations thanks to a fully customizable algorithm. In addition, the cloud-based dedicated software also has features to facilitate remote monitoring, including a platform for diabetes nurses who perform coaching and treatment adjustment. A detailed description of this telemedicine system is provided, as well as results of completed clinical studies. In particular, TeleDiab 1’s positive results on HbA1c in type 1 diabetes are detailed. We conclude with a discussion of the role of this telemedicine system within the landscape of mobile apps for diabetes.

Insulin is the primary treatment for type 1 diabetes (T1DM) and is delivered by multiple daily injections (MDI) or continuous subcutaneous insulin injection (CSII). In type 2 diabetes (T2DM) insulin therapy is often started as a daily injection of basal insulin, with the subsequent addition of bolus insulin, finally leading to MDI or CSII regimen. Despite these sophisticated therapeutic modalities, many patients with diabetes remain uncontrolled: approximately 3 out of 4 people with T2D and 2 out of 3 with T1D present with HbA1c above target, mainly related to nonadherence, concerns about management of dose titration and constraints in applying the complex calculations to determine bolus insulin.^1-3 To address this issue, use of telemedicine devices and reorganization of the process of care may improve glycemic control in people with T1D and T2D as shown in several trials.^4-6 The Diabeo® system (DS) (Voluntis, Paris, France), a medical device software, was developed by CERITD and Voluntis to help patients and health care professionals (HCPs) with diabetes monitoring and therapeutic decision-making.

Diabeo System

DS is a class IIb CE-marked medical device in Europe and is now being reimbursed by the French Authority for Health (Haute Autorité de Santé–HAS). DS consists of two key components: (1) a mobile application (MA) for patients available on Android or iOS operating systems and (2) a web portal (WP) accessed through any web browser. Data entered by the patient in the MA are automatically uploaded to the WP. Conversely, the parameters set by HCPs on the WP are directly implemented in the patient’s MA.

The MA is designed to be used several times a day, and provides the following key features:

• A connected diabetes logbook where the patient can enter current and historical glucose data, insulin, physical activity, carbohydrate intake, and hypoglycemic events (Figure 1). This logbook assists patients in data review, analysis, and evaluation through the computation of glucose statistics, in order to support diabetes self-management. Logbook data automatically uploads to a secure cloud database and is available on the WP for review by the HCP. A WP is also available for patients to view their data.

• A bolus calculator that provides bolus insulin titration guidance based on the logbook data entered by the patient (Figure 1). Bolus advice is based on current and past glucose values, glucose goals, hypo- and hyperglycemia, carbohydrate intake, planned physical activity (none, moderate or intense), insulin on board and reported ketones measured either by blood or urine. These data are processed by the MA, based on the embedded algorithmic parameters initially defined by HCPs through the Remote Treatment Configuration (RTC) module in the WP (Figure 2). The parameters include glucose goals, insulin brand, insulin plan (fixed meal and flexible insulin therapy), treatment type (MDI or pump), insulin/carb ratio, and insulin correction factor.

• Auto-adaptive dosing recommendations. This feature optimizes patient-specific algorithm parameters after automated analysis of their glucose history. Auto-adaptation of the insulin/carb ratio takes into account several postprandial blood glucose readings. For example, if the insulin/carb ratio for breakfast is initially set at 2 units per CHO portion, the system will suggest to the patient an increase of this ratio to 2.2 units per CHO portion if postbreakfast glucose values are regularly above the threshold. The patient can accept or decline the proposed dose. If the dose is accepted, then the relevant parameter will automatically reset and be used by the device for subsequent bolus calculations until a new state of auto-adaptation occurs. This adaptation can be repeated multiple times until the target glucose is reached, is controlled by several safety rules and is compatible with both fixed meal and flexible insulin therapy plans.

• Ketone integration into the algorithm. If the presence of ketones is entered, the bolus calculator will increase the dose as customized by the HCP through the RTC.

• A basal adjustment algorithm to provide guidance for long acting insulin doses (1 or 2/d) or 4 insulin pump basal rates based on mean fasting and/or premeal glucose values, hypoglycemia events and past basal insulin doses.

• Coaching messages focused on glucose data are sent regularly to provide educational information. Messages may appear on the MA screen when specific glucose patterns are detected by the device such as insufficient or missing glucose or CHO data.

Figure 1.

Example of the glucose logbook on the mobile application (a) and example of the bolus dose recommendation provided to the patient (b).

Figure 2.

Example of treatment plan settings configured by the HCP in the web portal.

The web portal is used by HCPs for remote monitoring and remote treatment plan configuration. Key features are:

• A multipatient dashboard that allows sorting of patients according to statistical criteria such as average BG levels or hypoglycemic events (Figure 3). From this dashboard, each patient’s data can be viewed including the treatment plan, logbook and graphs for statistical analysis.

• Remote configuration capability. The RTC module provides the ability to remotely adjust the bolus calculator, basal adjustments and auto-adaptation which are sent to the patient’s MA to provide titration guidance. HCPs can also specify which parts of the configuration the patient can modify, if any. Treatment adjustments set remotely can be accepted or declined by the patient.

• Coaching and feedback based on glucose patterns are available through secure messaging initiated by the HCP with automatic messages generated by the device. These are fully customizable in the RTC on a per-patient and per-physician basis.

• An organizational tool for telemedicine is embedded in the WP including telemedicine visit forms, a management module for the physician and diabetes team including nurses, technical support resources and the automated alert messages configuration module described above.

Figure 3.

Example of the HCP web portal dashboard (patients’ names and data are fictional). (a) Patients’ list; (b) analysis messages; (c): weight evolution of a patient.

Connection With Meters

DS is compatible with the My Star Plus glucose meter (Sanofi) and now has a Bluetooth interface that connects with compatible meters including Accu-Chek Connect (Roche) and Contour Next One (Ascencia) allowing direct transfer of data from the meter to the MA. Glucose data can also be manually entered in the MA by the patients.

Who Is DS For?

DS is designed for T1D and T2D adult patients, treated with an intensive insulin regimen (MDI or CSII), who are considered by their physician to be sufficiently experienced in their diabetes management. DS is available by prescription only and must be supported by a trained diabetes team. It is not intended for the management of life-threatening situations and is not suitable for users with nonstandard lifestyle patterns (eg, working night shifts, staggered meal times).

Safety Concerns

The functionalities offered to patients and HCPs are clearly different, for safety reasons. Only patients can enter data in their MA logbook. They may also adjust some algorithm parameters as authorized by their HCP. Therapeutic adaptations submitted by the system are proposals that the patients can accept or refuse, Thus, when using the system patients remain in control of their own therapeutic decisions at all times. This applies to all decision support functions, that is, bolus calculator, basal adaptation algorithm, ketone algorithm, and coaching messages. When ketones are reported, an extra amount of rapid-acting insulin is suggested only if the glucose level entered is both very high and confirmed by the patient.

Clinical Evidence

DS was studied in the TeleDiab 1 study, with 180 adult T1D patients using a MDI or CSII.³ Patients were randomized between three groups: Control, DS, and DS + Teleconsultations (G1, G2, and G3, respectively). Participants in the control group (G1) had no electronic logbook but kept their paper logbook and were asked to attend two follow-up visits at the hospital, after 3 and 6 months. A difference in HbA1c reduction was observed between G1 and G2 (0.67% [0.35-0.99], P < .001) and G1 and G3 (0.91% [0.60-1.21], P < .001), but not between G2 and G3. A post hoc analysis showed that the metabolic benefit tended to be higher in G3 compared to G2 among patients using DS MA less assiduously (−0.93 ± 0.97 vs −0.46 ± 1.05, respectively; P = .084), suggesting that the human intervention associated with the mobile application is of paramount importance.⁷ Symptomatic severe and nonsevere hypoglycemia episodes were not increased from baseline and did not differ between groups at endpoint. Total physician time spent consulting during this 6-month study was 70 ± 31 minutes in G1 and 70 ± 22 minutes in G2. Total time spent tele-consulting during the all study period in G3 was similar (72 ± 30 minutes) with a mean number of teleconsultations of 8.7 ± 4.9 and a mean duration for one teleconsultation of 7.4 ± 3 minutes. However, such results will need to be confirmed by routine practice surveys, in order to assess medical time investment with DS in real-life setting.

Another version of DS, a basal calculator for T2D patients initiating basal insulin therapy, was assessed in the multicenter TeleDiab 2 trial.⁸ T2D patients (N = 191) who failed to reach HbA1c target with oral antidiabetic drugs were included. They were randomized in three groups: usual care, daily use of an interactive voice response system (IVRS), or daily use of DS-basal to support basal insulin titration. After 4 months, a −0.56% HbA1c improvement was demonstrated with DS-basal, compared to usual care. This benefit was maintained during the 9-month extension phase with more than twofold proportion of patients with HbA1c <7% in the DS-basal group compared to usual care (30.2% vs 13.8%, respectively; P = .023).

Next, the Telesage study (NCT02287532) was designed to demonstrate that DS could improve metabolic control in a larger population of both T1D and T2D patients treated with basal bolus insulin over a longer period of time.⁹ This 700 patient, 2-year, multicenter trial recently concluded and results have not been yet communicated.

Reimbursement

Clinical evidence from TeleDiab 1 has supported the first diabetes digital therapeutics reimbursement by the French Authority for Health for patients with T1DM treated with intensive insulin therapy. DS is now available in France in the national pilot plan (ETAPES) that provides reimbursement for the three components of the solution: (1) the technical solution that supports telemonitoring activity; (2) therapeutic support; and (3) medical telemonitoring, both of which can be delegated to a nurse.

Landscape of Mobile Medical Applications for Diabetes

DS provides a self-monitoring glucose logbook, a bolus calculator, and an auto-adaptive function of insulin therapy parameters in order to optimize patient-specific algorithm parameters after automated analysis of their glucose history. This last feature makes this device unique in the digital diabetes landscape, which includes more than 1000 diabetes apps, as listed by Drincic et al in 2016.¹⁰ Among these apps only 14 systems, including DS, have been scientifically assessed with a published trial and/or have received FDA clearance or CE Mark. Recently, DS has been identified as one of a small number of apps associated with clinically significant improvement.^11,12 In comparison, Blue Star® (WellDoc) for T2D combines a mobile coaching application and patient/provider web portals, with automated educational messages returned to the patient in response to SMBG data entered into the application. In a study by Quinn et al, Blue Star showed a 1.2% HbA1c reduction in the group of patients using this application, compared to the control group, but it should be noted that the mean baseline HbA1c was slightly higher in the this study compared to TeleDiab 1 study.¹³ Assessment of DS through clinical trials is of paramount importance as some authors have raised the question of accuracy and safety of medical applications that provide a bolus calculator.^14,15 Clinical evidence and clearance by regulatory agencies, as for DS, are two key elements of this software that HCPs should take into account when choosing an application for their patients, to avoid the risk of severe incidents potentially associated with inaccurate applications.

Conclusion

The DS combines (1) a smartphone application with a self-monitoring glucose logbook and an on-board automated insulin dose adjustment system and (2) a connected cloud-based software for telemedicine. This digital therapeutic system is dedicated to type 1 or type 2 diabetes patients with intensive insulin therapy (MDI or CSII). DS has been validated in two randomized clinical trial that showed significant HbA1c improvement with this system in both T1D and T2D patients. Results of a larger ongoing study will provide additional scientific data about this system.

For the first time in France, thanks to the national pilot of remote monitoring for patients with diabetes, telemonitoring is being reimbursed in a real world setting. Expectations of the French health authorities are to lower costs of outpatient and inpatient treatment and to improve metabolic control. DS will likely be one of the main systems utilized in the diabetes telemedicine pilots in France and can serve as a model for other countries as well.