In total, 83 patients with type 2 diabetes were included (Supplementary Fig. 1), of whom 38.6% were new Gla-300 users. Of the 83 patients, 55.4% were male, mean (SD) age at index date was 49.2 (13.1) years and 31.3% had a diabetes duration of > 10 years (Table 1). New Gla-300 users were younger on average (mean [SD] age 45.5 [14.9] versus 51.6 [11.2] years, p = 0.03) and had a shorter duration of diabetes compared with previous users (9.4% versus 45.1% with duration > 10 years). The mean (SD) HbA1c at baseline was 8.4 (2.2) % (68.3 [24.0] mmol/mol) overall, and significantly higher for new Gla-300 users (9.6 [2.3] %, 81.4 [25.1] mmol/mol) versus previous Gla-300 users (7.7 [1.8] %, 60.6 [19.7] mmol/mol, p < 0.01). The mean (SD) total daily BI dose in the baseline period was 18.8 (12.1) units and 19.9 (13.2) units in the new and previous Gla-300 user groups, respectively. In the baseline period, 10 patients (12.0%) experienced hypoglycemia.

Mean (SD) HbA1c change from baseline to follow-up was − 1.2 (2.3) % (− 13.1 [25.1] mmol/mol) for the total population, − 2.4 (2.7) % (− 26.2 [29.5] mmol/mol) in new Gla-300 users, and − 0.5 (1.6) % (− 5.5 [17.5] mmol/mol) in previous Gla-300 users (Fig. 1a); this change was statistically significant in all groups (p < 0.05). Overall, users with HbA1c ≥ 10.0% (86.0 mmol/mol) at baseline experienced the largest change in HbA1c level in the follow-up period. Similar trends were observed for both new and previous Gla-300 users with larger reductions in HbA1c as baseline HbA1c increased (Supplementary Table 1). For the total population, mean change in HbA1c was greatest with 3 months’ duration of Mallya cap use (− 1.3 [2.1] %, − 14.2 [23.0] mmol/mol) compared with < 2 months (− 1.1 [2.5] %, − 12.0 [27.3] mmol/mol) or 2 to < 3 months (− 1.2 [2.3] %, − 13.1 [25.1] mmol/mol) (Fig. 1b).

Mean change in HbA1c from baseline to follow-up stratified by a prior Gla-300 use and b duration of Mallya cap use. *P < 0.05. Gla-300 insulin glargine 300 units/mL, SD standard deviation

Overall, the proportion of users achieving HbA1c < 7.0% (53.0 mmol/mol) increased from 32.5% at baseline to 44.6% at follow-up (Fig. 2a); this increase was mainly in new Gla-300 users, of whom 12.5% had HbA1c < 7.0% (53.0 mmol/mol) at baseline versus 40.6% at follow-up. Stratification by baseline HbA1c showed that the goal of HbA1c < 7.0% (53.0 mmol/mol) at follow-up was achieved by 77.8% of users who had a baseline HbA1c of < 7.0% (53.0 mmol/mol). Among higher-baseline HbA1c groups, target achievement ranged from 18.2% for users with baseline HbA1c of 9.0–9.9% (75.0–85.0 mmol/mol) to 37.5% of users with baseline HbA1c of ≥ 10.0% (86.0 mmol/mol, Supplementary Table 1).

HbA1c goal achievement of HbA1c < 7.0% (53.0 mmol/mol) at baseline and follow-up stratified by a prior Gla-300 use and b duration of Mallya cap use. Gla-300 insulin glargine 300 units/mL

Stratification by Mallya cap use showed that HbA1c < 7.0% (53.0 mmol/mol) at follow-up was achieved by 48.0%, 45.0%, and 42.1% of users with Mallya cap use duration of < 2 months, 2 to < 3 months, and 3 months, respectively (Fig. 2b). No obvious trend in Mallya cap use was observed for new (50.0%, 14.3% and 46.7% at < 2, 2 to < 3 months and 3 months, respectively) or previous Gla-300 users (46.7%, 61.5% and 39.1% at < 2, 2 to < 3 months, and 3 months, respectively).

HbA1c reduction > 0.5% (5.5 mmol/mol) was achieved by 65.6% of new Gla-300 users and 29.4% of previous Gla-300 users (Fig. 3a). The proportion of users with a HbA1c reduction from baseline to follow-up of > 0.5% (5.5 mmol/mol) increased steadily with baseline HbA1c, from 7.4% of users with a baseline HbA1c of < 7.0% (53.0 mmol/mol) achieving this reduction up to 100% of users with a baseline HbA1c ≥ 10.0% (86.0 mmol/mol) achieving this goal (Supplementary Table 1). An HbA1c reduction of > 0.5% (5.5 mmol/mol) was achieved by 36.0%, 40.0%, and 50.0% of users with Mallya cap use duration of < 2 months, 2 to < 3 months, and 3 months, respectively (Fig. 3b). Similar trends were observed for new (60.0%, 57.1%, and 73.3% at < 2, 2 to < 3, and 3 months, respectively) and previous Gla-300 users (20.0%, 30.8%, and 34.8% at < 2, 2 to < 3, and 3 months, respectively).

HbA1c reduction > 0.5% (5.5 mmol/mol) at follow-up stratified by a prior Gla-300 use and b duration of Mallya cap use. Gla-300 insulin glargine 300 units/m

Mean (SD) duration of Mallya cap use was 71.4 (25.4) days and median (range) was 87 (13–91) days. During this period, the mean (SD) number of days that Mallya cap was used to capture insulin dose values was 57.8 (25.2) days (median [range] 62 [13–90] days). The mean (SD) length of gaps between Mallya cap use was 2.4 (6.0) days (median [range] 1.3 [0–54] days) and 4 (4.8%) participants discontinued Mallya cap use during the follow-up period.

At follow-up, the mean (SD) total daily Gla-300 dose recorded using Mallya cap and H2S App was 22.2 (14.7) units and 23.4 (13.4) units in the new and previous Gla-300 user groups, respectively. The mean change in Gla-300 dose from baseline to follow-up was similar between the two groups (mean [SD], 3.3 [16.5] units in new user group and 3.4 [9.2] units in previous user group).

In total, 27.7% of the 83 patients experienced any hypoglycemia during follow-up, with an event rate of 3 per person per year (PPPY; Supplementary Table 2). When a standard time window was used to assess nocturnal hypoglycemia, six program users (7.2%) experienced nocturnal hypoglycemia ≤ 70 mg/dL (≤ 3.9 mmol/L), while one user (1.2%) experienced nocturnal hypoglycemia < 54 mg/dL (3.0 mmol/L), with an event rate of 0.6 PPPY and 0.1 PPPY, respectively. When an expanded window was used to assess nocturnal hypoglycemia, the proportions of patients experiencing nocturnal hypoglycemia ≤ 70 mg/dL (3.9 mmol/L) and < 54 mg/dL (3.0 mmol/L) increased to 15.7% and 4.8%, respectively. The event rates were 1.5 and 0.2 PPPY for ≤ 70 mg/dL (3.9 mmol/L) and < 54 mg/dL (3.0 mmol/L), respectively.