For patients with type 2 diabetes (T2D), achieving recommended glycemic targets remains difficult, especially in people treated with basal insulin. One of the reasons for this difficulty is related to the challenge of titrating insulin doses. Insulia is a digital solution combining a smartphone app for basal insulin dose suggestions and a web portal accessible to professionals to personalize and manage patients’ treatments remotely. Beyond remote monitoring of basal insulin therapy, the app uses the data entered by the patients to calculate the recommended basal insulin dose according to the objectives set by the patient’s physician. This dose calculation is triggered by the patient’s request.

A predecessor device to Insulia called Diabeo-Basal was evaluated in the TeleDiab-2 study [1]. This randomized controlled trial evaluated the efficacy and safety of two remote monitoring systems to optimize basal insulin initiation in patients with poorly controlled T2D (hemoglobin A_1c [HbA_1c] 7.5%-10%). A total of 191 participants (mean age 58.7 years, mean HbA_1c 8.9%) were randomized into three groups: group 1 (standard care, n=63), group 2 (interactive voice response system, n=64), and group 3 (Insulia app software, n=64). After 4 months of follow-up, HbA_1c reduction was significantly higher in the remote monitoring groups (group 2: –1.44% and group 3: –1.48% vs group 1: –0.92%; P=.002). In addition, twice as many patients in the telemonitoring groups achieved their target fasting blood glucose (FBG) level as in the control group, and insulin doses were adjusted to higher levels. No severe hypoglycemia was observed in the remote monitoring groups, and the frequency of mild hypoglycemia was similar in all groups.

Consequently, Insulia was available by prescription and used in France as part of a nationwide program financing new health remote monitoring systems (Expérimentations de télémédecine pour l'amélioration des parcours en santé [ETAPES] program: National Experiments on Remote Diabetes Monitoring [2]) since the end of 2020. Despite the potential benefits for patients suggested by the TeleDiab-2 study, Insulia, as with other apps offering insulin dose calculation, carries a risk of incorrect dose recommendations, which could lead to suboptimal disease control [3]. In this field, perhaps even more than for other health products, it is necessary to complement experimental results with the analysis of monitoring data on both the clinical efficacy and safety of the apps during their use in real life.

The data entered by the patients and their physicians in the Insulia app are collected on a dedicated computer platform. This study aims to analyze this database to determine how the glycemic control of Insulia users has evolved when using the app in a real-life setting in France.

The Insulia database included 484 patients enrolled as users of the app for 6 months or more. Among them, 111 patients did not conduct any dose calculation with the device or did not indicate any basal insulin dose injected since their registration on the app. Consequently, 373 patients were considered in the main analysis. Among them, 91 (24.4%) patients were identified as regular and compliant users over a 6-month period.

The characteristics of the patients are described in Table 1. On average, they were aged 55.8 (SD 11.9) years, and 59.2% (n=221) were men. At the time of their first use of the Insulia device, 48.6% (n=181) of them had a BMI ≥30 kg/m² (average BMI 30.6 kg/m²). The mean HbA_1c level was 10.1% (SD 2%). The individual FBG target ranged from 70-100 to 100-150 mg/dL. The target ranges were 80-130 mg/dL for 33.5% (n=125) of the patients and 80-120 mg/dL for 30% (n=112) of the patients. The first calculated basal insulin dose averaged 25.5 IU with significant variability (between 4 IU and 92 IU according to the patients). Among compliant and regular users (n=91, 24.4% of the patients over the 6 months of observation), the HbA_1c level at baseline was slightly lower compared to other patients (9.6% vs 10.3%; P=.002), and the FBG target was slightly more stringent, with a higher proportion of patients having an FBG target in the range of 80-120 mg/dL and a lower proportion in the range of 100-150 mg/dL.

The percentage of patients defined as regular and compliant users evolved during the 6-month period of the study with a progressive disaffection of the patients from the second month of use. A similar evolution was observed considering only regular use of the device (at least 5 dose calculations per week; Figure 1).

Figure 2 shows the percentage of Insulia users with an average FBG in their individualized target range over 15-day periods according to whether they are regular and compliant users. The percentages are calculated on the number of patients still using the device over the 15-day period. We observed an increase in the proportion of patients achieving their target FBG for regular and compliant users (+1.86% every 2 weeks). No clear improvement was observed in other patients (irregular or not compliant users of the app).

After 6 months of use (5.5 to 6 months), the FBG target was achieved in 60% (55/91) of the regular and compliant users versus 51.5% (145/282) of the other patients (P=.15), although the FBG target was slightly more stringent for the regular and compliant users.

Variables available at baseline (age, gender, BMI, HbA_1c level, and insulin dose at Insulia initiation) were tested in a logistic model to explain potential factors associated with achieving the individualized FBG target at 6 months among regular and compliant users (n=91; Figure 3). None of these variables were significantly associated with this outcome.

Figure 4 presents the evolution of the HbA_1c level over time for patients having at least 2 measurements regardless of the time elapsed between these two measurements (all patients, n=182, and all regular and compliant users, n=69). In both cases, a slight but significant decrease in HbA_1c values of –0.155% and –0.161% per month, respectively, was observed.

Data were available at baseline and 6 months for only 82 patients. The mean HbA_1c level decreased from 9.9% to 7.2% after 6 months (±1.5 months) of app use, with no significant difference according to the degree of Insulia use.

Table 2 shows the numbers and proportions of patients who reported in the app that they had at least one change in treatment because of hypoglycemia, defined as a blood glucose measurement <70 mg/dL and whether this hypoglycemia was symptomatic or not. A favorable trend (–0.16% per month) was observed but not statistically significant due to the low number of patients.

Finally, to identify a possible center effect, a subgroup analysis was conducted on the sample of patients enrolled in the principal investigating center (Centre Hospitalier Sud Francilien [CHSF], Corbeil-Essonnes) where 68 (18.3%) patients of the overall population considered in the analysis (N=373) were enrolled. The baseline characteristics of those patients were similar to those of other centers, excluding a higher average HbA_1c level at the time of the first basal insulin dose calculation (10.9%, SD 2.3% vs 10.0%, SD 1.9%). Interestingly, individualized FBG targets were also often less strict in this center than in other centers (patients with a target between 100-150 mg/dL: 20/68, 29.4% vs 12/305, 3.9%), and over the first 6 months, patients at the CHSF were less often regular and compliant users than at the other centers (9/68, 13.2% vs 82/305, 26.9%; P=.02). Despite these discrepancies, the percentage of Insulia app users achieving their FBG target after 6 months was not different considering the overall population (P=.77) or only regular and compliant users (P=.75), excluding a center effect on the results.