To the best of our knowledge, this is the first direct comparison between the glycemic parameters of the open-source AID and the 780G systems among users with T1D younger than 21 years of age. Interestingly, our investigation revealed pre-existing differences between the two groups. Specifically, those who opted for the open-source AID system had a lower initial HbA1c levels but spent more time in the hypoglycemic range. At the end of follow-up, both groups improved their TIR70 − 180 levels to a similar extent.
Open-source systems are free, and anyone can access the instructions and codes via open-source platforms. However, since users are responsible for building their individual system use is limited to those who are capable of meeting the technical requirements. Indeed, most adult users (83%) and caregivers (87%) reportedly had a university degree or higher, one quarter had a professional background in information technology, and one-fifth had a professional background in biomedicine or healthcare.12
Our data show that HbA1c levels were lower in the open-source AID group prior to the initiation of this technology, although the time from diagnosis was similar, suggesting the pursuit of near-normal glucose levels as being principle to its application. This is in accordance with a study on the motivations to commence open-source AID, with improvement of glycemic outcomes having been reported as the primary motivation for 93.5% of adults and 95% of caregivers [11].
Individuals in both of our study groups benefited from the switch to AHCL technologies with both having improved TIR70-180mg/dL. The open-source AID group improved its HbA1c level by 0.25% and its TIR70-180 mg/dL level by 6%. This is consistent with the result of an online survey of 209 caregivers of children from 21 countries.[12] While using the AHCLs the reduction of HbA1c was ~ 0.3% in both groups, while the median TIR70-180 mg/dL level increased from 68.3–78.0% in the open-source AID and from 64.0–75.0% in the 780G group. These findings are in line with the 780G clinical trial results [13].
The improvement in TIR70-180 mg/dL in the open-source AID group was associated with relatively prolonged time spent in the hypoglycemic ranges. On the other hand, the 780G group spent significantly less time in the TBR54-70mg/dL and TBR< 54mg/dL, which may reflect the more conservative glucose target range in the 780G and the fact that these targets cannot be adjusted by users. Furthermore, the median TBR< 54mg/dL in the open-source AID group at the end of the follow-up (1.1%) was higher than the recommended target of less than 1%,[14] and the recommended target in the 780G group (0.0%). Similarly, the median TBR54-70mg/dL was higher than the recommended target of less than 4% in the open-source AID group and the recommended target in the 780G group (4.2% and 2.0%, respectively).
The open-source AID users spent approximately 16 minutes more per day below the 54 mg/dL range than the 780G users. This difference has clinical significance given that severe hypoglycemia is a major barrier to achieve optimal glycemic control and that it may affect the patients' quality of life[15]. It is well documented that intensive glycemic control increases the frequency of hypoglycemia [16]. These findings are important for caregivers, since treatment with the 780G system may be safer and more appropriate for those who are prone to recurrent hypoglycemic events. Of note, no severe hypoglycemic episodes were reported among either the open-source AID or the 780G users in the current study.
At the end of the follow-up, the open-source AID group had better TIR70 − 180 mg/dL compared with the 780G group, but this encouraging result could be associated with the well-controlled TIR70 − 180 mg/dL prior to open-source AID use. The change and the difference in all parameters were of the same magnitude in both groups.
Open-source AID offers several key benefits, including the flexibility to utilize a wide range of sensors and pumps, the ability to select preferred glucose targets, and the convenience of remote insulin delivery. This last feature is particularly valuable for parents of young children who require close monitoring and support. Nevertheless, this system requires advanced diabetes self-management abilities. In our study, the main advantages of the 780G system were the significantly decreased time spent in the hypoglycemic ranges (i.e., by one third for TBR< 54mg/dL and by one-half for TBR54 − 70mg/dL).