Glycemic Outcomes with an Open-Source Automated Insulin Delivery System in Adults with Type 1 Diabetes: A 1-Year Real-World Observational Study

Abstract

Objective:

To evaluate the long-term effectiveness and safety of the AndroidAPS (AAPS) system in adults with type 1 diabetes (T1D) under real-world clinical conditions.

Methods:

This retrospective, single-center study included 27 adults with T1D (mean age 39 ± 9.6 years; 55.6% women; diabetes duration 21.5 ± 11.3 years) who initiated AAPS and were followed for 12 months. Glycemic metrics were obtained from real-time continuous glucose monitoring (rtCGM) data. Primary outcomes were changes in time in range (TIR; 70–180 mg/dL) and HbA1c. Secondary outcomes included time in tight range (TITR; 70–140 mg/dL), time below range (TBR), coefficient of variation (CV), and the glycemic risk index (GRI). Safety outcomes included severe hypoglycemia and diabetic ketoacidosis (DKA).

Results:

TIR increased from 67.8% at baseline to 84.1% at 6 months and 79.9% at 12 months (P < 0.001). HbA1c decreased from 6.6% to 6.1% and 6.0%, respectively. TITR improved by 14.8% at 6 months and 10.6% at 12 months, while GRI decreased from 40.2 to 21.3. Over 88% of participants achieved TIR > 70% and CV < 36% at both follow-ups, and up to 74% achieved TITR > 50% at 6 months, remaining at 63% at 12 months. No episodes of severe hypoglycemia or DKA were reported.

Conclusions:

AAPS demonstrated sustained long-term effectiveness and safety in adults with T1D under real-world clinical practice conditions. Glycemic outcomes were comparable to those reported with commercially available closed-loop automated insulin delivery systems, supporting AAPS as a viable and effective advanced management option for T1D.

Keywords

AndroidAPS open-source automated insulin delivery time in range time in tight range type 1 diabetes

Introduction

Automated insulin delivery (AID) systems represent a major advancement in the management of type 1 diabetes (T1D), improving glycemic outcomes and reducing treatment burden. However, cost, accessibility, and regulatory limitations have restricted their adoption.¹

Open-source AID (OS-AID) systems—also referred to as “do-it-yourself” (DIY) artificial pancreas systems—have emerged as community-driven, customizable alternatives integrating real-time continuous glucose monitoring (rtCGM) and insulin pump therapy, and are developed and maintained by online communities of people with T1D and their families under the #WeAreNotWaiting movement.² The most widely used platforms are AndroidAPS, Loop, and OpenAPS. AndroidAPS (also now called AAPS), designed for Android devices, uses the OpenAPS algorithm and allows different configurations including temporary basal rate automation only (oref0) or also small microboluses called “supermicroboluses” (SMB) (oref1), based on predicted glucose trends.³

Although OS-AID systems are not formally approved by regulatory agencies, the recent FDA authorization of Tidepool Loop, a commercial version of the Loop algorithm, illustrates growing recognition of these approaches.⁴ Clinical supervision remains essential to ensure appropriate follow-up and documentation of off-label use.

Evidence from clinical trials^5–11 and real-world studies^2,12–18 shows that OS-AID systems can lower HbA1c, increase time in range (TIR), and reduce time below range (TBR) and above range (TAR), improving quality of life. In recent years, the evidence base supporting OS-AID has expanded considerably, with emerging trials incorporating longer follow-up and more rigorous methodological designs. Within this evolving context, real-world data such as ours offer valuable insight into the sustained effectiveness and safety of these systems in routine clinical practice.

Recently, complementary metrics such as the glycemic risk index (GRI) and time in tight range (TITR; 70–140 mg/dL) have been proposed to refine glycemic assessment.^19,20 The GRI quantifies hypo and hyperglycemia, emphasizing extreme values, whereas TITR reflects time in euglycemia and may better discriminate control in well-managed patients.

In this context, this study aimed to evaluate, under real-world clinical conditions, the long-term effectiveness and safety of AAPS in adults with T1D, incorporating expanded metrics such as GRI and TITR alongside conventional glycemic parameters.

Material and Methods

Study design and population

This was an observational, single-center, retrospective study with a 12-month follow-up period conducted under routine real-world clinical practice conditions including adults with T1D previously treated with multiple daily insulin injections (MDI) or continuous subcutaneous insulin infusion (CSII), all using rtCGM. Participants initiated AAPS between November 2018 and December 2020 as part of routine follow-up at the Endocrinology and Nutrition Department of Infanta Sofía University Hospital (Madrid, Spain). The study followed the Declaration of Helsinki, and all participants provided written informed consent.

All participants used the oref1 (SMB) version of AAPS on compatible smartphones connected to an Accu-Chek Insight insulin pump. System settings, including glycemic targets and algorithm parameters, were individualized by users without a standardized protocol. However, as part of routine care in our department, the clinical team provided ongoing support to individuals using OS-AID systems. This included education during follow-up visits, review of CGM and insulin delivery data, and collaborative adjustment of configuration parameters when appropriate. Such involvement forms part of our standard clinical practice to ensure safe and effective OS-AID use.

Outcomes

Primary outcomes were changes in TIR (70–180 mg/dL) and HbA1c at 6 and 12 months after system initiation. Secondary outcomes included changes in TAR (>180 and > 250 mg/dL), TBR (<70 and < 54 mg/dL), TITR (70–140 mg/dL), mean glucose, coefficient of variation (CV), glucose management indicator (GMI), and GRI with components hypoglycemia (CHypo) and hyperglycemia (CHyper). Safety endpoints included severe hypoglycemia, diabetic ketoacidosis (DKA), and diabetes-related hospitalizations.

Data collection and analysis

Glycemic data were retrospectively extracted from electronic medical records and the Nightscout platform. For each time point (baseline, 6 and 12 months), data from the last 14 days with >75% sensor use were analyzed. HbA1c was measured using the National Glycohemoglobin Standardization Program.

The GRI was calculated using the standard formula: GRI = (3.0 × %TBR < 54 mg/dL) + (2.4 × %TBR < 70 mg/dL) + (1.6 × %TAR > 250 mg/dL) + (0.8 × %TAR > 180 mg/dL). The theoretical maximum value was 100. Based on the GRI percentile, participants were classified into five zones representing progressively worse glycemic profiles: zone A (0–20 percentile), zone B (21–40), zone C (41–60), zone D (61–80), and zone E (81–100).

Statistical analyses were performed using SPSS version 25 (IBM Corp., Armonk, NY, USA). Continuous variables were compared using paired t or Wilcoxon tests, and categorical variables with chi-square. A P value of <0.05 was considered significant.

Results

A total of 27 adults with T1D (mean age 39 ± 9.6 years; 55.6% women) were included. Mean body mass index (BMI) was 25.4 ± 3.8 kg/m², diabetes duration 21.5 ± 11.3 years, and baseline HbA1c 6.6% ± 0.5%. Nineteen participants (70.4%) used CSII with sensor-augmented pump therapy and 8 (29.6%) MDI therapy, all with rtCGM.

Comorbidities included retinopathy (18.5%), neuropathy (7.4%), and prior severe hypoglycemia (3.7%). None had recent DKA nor diabetic nephropathy. Four participants used adjuvant agents (two SGLT2i and two GLP-1 receptor agonists).

All participants used AAPS in conjunction with an Accu-Chek Insight insulin pump. Regarding the CGM system, 18 participants (66.7%) used the Dexcom G6, and 9 (33.3%) used the FreeStyle Libre modified with NFC–Bluetooth transmission. Most participants (66.7%) had a university-level education, including four healthcare professionals and three engineers.

Glycemic outcomes

TIR (70–180 mg/dL) increased from 67.8% ± 12.8% at baseline to 84.1% ± 6.4% at 6 months and 79.9% ± 10.0% at 12 months (P < 0.001), accompanied by significant reductions in TAR > 180 mg/dL (19.2% ± 13.6% to 12.1% ± 6.0% and 12.2% ± 6.1%) and TAR > 250 mg/dL (6.5% [3–11] to 1.2% [0–3] and 2.2% [0–5]) at 6 and 12 months, respectively (P < 0.001 for all).

Mean glucose decreased from 141 ± 21.2 mg/dL to 126.2 ± 14.6 mg/dL at 6 months and 129.3 ± 11.5 mg/dL at 12 months (P < 0.001), with a parallel reduction in GMI from 6.5% ± 0.7% to 6.2% ± 0.8% and 6.1% ± 0.8%. TBR < 70 mg/dL decreased from 4.2% ± 2.6% to 2.0% ± 1.4% and 3.2% ± 1.8%, and TBR < 54 mg/dL from 2.6% [0.5–5] to 0.4% [0–4] and 0.9% [0–4] at 6 and 12 months, respectively (P < 0.001 for all) (Table 1, Fig. 1).

Table 1.

Glycemic Outcomes and System Use at 6 and 12 Months Following AAPS Initiation

	Baseline	6 months (AAPS)	12 months (AAPS)	P
Glycemic outcomes
TIR 70–180 mg/dL (%) (mean ± SD)¹	67.8 ± 12.8^a	84.1 ± 6.4^b	79.9 ± 10^b	<0.001
TITR 70–140 mg/dL (%) (mean ± SD)¹	53 ± 15.6^a	67.8 ± 13.5^b	63.6 ± 15.8^b	<0.001
Time >180 mg/dL (%) (mean ± SD)¹	19.2 ± 13.6^a	12.1 ± 6^b	12.2 ± 6.1^b	<0.001
Time >250 mg/dL (%) (median [IQR])²	6.5 (3–10)^a	1.2 (0–3)^b	2.2 (0–5)^b	<0.001
Time <70 mg/dL (%) (mean ± SD)¹	4.2 ± 2.6^a	2 ± 1.4^b	3.2 ± 1.8^b	<0.001
Time <54 mg/dL (%) (median [IQR])²	2.6 (0.5–5)^a	0.4 (0–4)^b	0.9 (0–4)^b	<0.001
Mean glucose (mg/dL) (mean ± SD)¹	141 ± 21.2^a	126.2 ± 14.6^b	129.3 ± 11.5^b	<0.001
GMI (%) (mean ± SD)¹	6.5 ± 0.7^a	6.2 ± 0.8^a,b	6.1 ± 0.8^b	<0.001
CV of glucose (%) (mean ± SD)¹	36.6 ± 6.4^a	29.5 ± 5^b	30.4 ± 5.7^b	<0.001
GRI (%) (mean ± SD)¹	40.2 ± 15.3^a	19.7 ± 7.8^a,b	21.3 ± 8.8^b	<0.001
CHypo (median (IQR))²	4.2 (3–8)^a	1.6 (1–3)^b	2.7 (2–4)^b	<0.001
CHyper (median (IQR))²	15 (9–22)^a	6.8 (5–12)^a,b	7.5 (4–12)^b	<0.001
GVI (median (IQR))²	1.3 (1–2)^a	1.3 (1–2)^a	1.3 (1–2)^b	<0.001
HbA1c (%) (mean ± SD)¹	6.6 ± 0.5^a	6.1 ± 0.5^b	6 ± 0.4^b	<0.001
Parameters of use
Total insulin dose (U/day) (median [IQR])²	38.2 (28–45)	47 (30–50)	46.3 (26–45)	0.086
Total basal insulin (%) (mean ± SD)¹	56.8 ± 13	44.2 ± 12.5	47.8 ± 13.1	0.595
Total bolus insulin (%) (mean ± SD)¹	43.2 ± 12.3	55.8 ± 19.2	52.2 ± 14.8	0.597
Carbohydrate intake (g/day) (median [IQR])²	110 (93–138)	117 (87–132)	119 (80–130)	0.351
Objective (mg/dL) (mean ± SD)¹		90 ± 5	95 ± 10	0.961
BMI (kg/m²⁾ (mean ± SD)¹	25.4 ± 3.8	26.3 ± 5.6	26 ± 4.3	0.338

Bold values denote statistical significance at the p < 0.05 level.

Different superscripts indicate statistically significant pairwise differences between follow-up times (p < 0.05).

¹Friedman test.

²Repeated-measures ANOVA.

BMI, body mass index; CHyper, hyperglycemia component; CHypo, hypoglycemia component; CSII, continuous subcutaneous insulin infusion; CV, coefficient of variation; GMI, glucose management indicator; GRI, glycemic risk index; MDI, multiple daily injections; rtCGM, real-time continuous glucose monitoring.

IQR: interquartile range; SD: standard deviation.

FIG. 1.

Column charts showing percentage time below range (TBR < 54 and < 70 mg/dL), time in range (TIR 70–180 mg/dL), and time above range (TAR > 180 and > 250 mg/dL) at baseline, 6 months, and 12 months of AndroidAPS (AAPS) use.

CV improved from 36.6% ± 6.4% to ≈30% (P < 0.001). HbA1c decreased by 0.5% at 6 months and 0.6% at 12 months (P < 0.001). GRI decreased from 40.2 ± 15.3 at baseline to 19.7 ± 7.8 and 21.3 ± 8.8 at 6 and 12 months, respectively (P < 0.001), with parallel reductions in CHypo (4.2 [3–8] to 1.6 [1–3] and 2.7 [2–4]) and CHyper (15 [9–22] to 6.8 [5–12] and 7.5 [4–12]) (Fig. 2).

FIG. 2.

Glycemic risk index (GRI) grid displaying the hyperglycemia component (CHyper) versus the hypoglycemia component (CHypo) at baseline and after 12 months of AAPS use.

The proportion of participants achieving TIR > 70% increased from 44.4% to 100% at 6 months and 88.9% at 12 months, while those with TBR < 4% rose from 51.8% to 88.9% and 92.6%, respectively. Participants meeting both targets increased from 25.9% to 88.9% and 81.5%. TITR > 50% rose from 22.2%–74.1% and 63.0%, and CV < 36% from 48.1% to 88.9%. At baseline, 74.1% of participants had HbA1c <7.0%, increasing to 92.6% at 6 months and 96.3% at 12 months. The proportion achieving HbA1c ≤6.5% rose from 44.4% to 85.2% at 6 months and 81.5% at 12 months.

GRI showed strong negative correlations with TIR (r = −0.92) and TITR (r = −0.82) and positive correlations with HbA1c (r = 0.82), GMI (r = 0.77), TAR > 180 mg/dL (r = 0.89), TAR > 250 mg/dL (r = 0.85), and mean glucose (r = 0.69). Moderate positive correlations were also found with TBR < 70 mg/dL (r = 0.39) and TBR < 54 mg/dL (r = 0.47) (P < 0.001 for all). No significant changes were observed in BMI, total daily insulin dose, basal/bolus ratio, or carbohydrate intake during follow-up.

Safety

No severe hypoglycemia, DKA, or diabetes-related hospitalizations occurred during follow-up, and all participants continued using AAPS. Reported satisfaction was high, although not formally assessed.

Discussion

Despite major advances with AID systems, up to 28% of users still fail to meet recommended glycemic targets, highlighting persistent unmet clinical needs.²¹ In this context, open-source AID (OS-AID) systems, developed and maintained by individuals with T1D and their families, have emerged as accessible and customizable alternatives outside conventional regulatory frameworks.

Real-world evidence has shown that OS-AID systems improve glycemic outcomes, safety, and quality of life in both pediatric^13,22,23 and adult^{2,8,12,14–17} populations, although data on their long-term sustainability remain limited.

This study demonstrates sustained real-world use of AAPS over 12 months in adults with T1D, with clinically relevant improvements in TIR (+16.2% at 6 months and +12.0% at 12 months), HbA1c (−0.6%), TITR, and GRI, without increases in TBR or serious adverse events.

These findings are consistent with previous short-term studies using AAPS^8,15 and other OS-AID systems such as Loop,¹² as well as longer-term evaluations by Patel¹⁶ and Liarakos.¹⁷ Wu et al.¹⁵ reported a 9.3% increase in TIR (final 84.3%) and a 0.8% reduction in HbA1c after 3 months, while Gawrecki et al.⁸ observed an 11.3% increase in TIR (final 79.3%) and a 0.5% decrease in HbA1c. In a multicenter study of 558 Loop users, Lum et al.¹² found a 6.6% rise in TIR (final 73%) and a 0.3% HbA1c reduction, without increased TBR.

To date, only two studies have evaluated OS-AID systems beyond 1 year of use: Patel et al.¹⁶ (1.4 years; TIR 73%, TBR < 70 mg/dL 2.4%, HbA1c −0.9%) and Liarakos et al.¹⁷ (1.7 years; HbA1c −0.8%, TIR 78.6%, TBR 3.2%).

Our findings are also comparable to those of the CREATE clinical trial, in which AAPS was used for 24 weeks in 22 adults with T1D, achieving a TIR of 74.5%, TBR of 1.6%, and an HbA1c reduction of −0.5%, with sustained improvements during the extension phase.^9,10

Furthermore, comparative studies suggest that OS-AID systems can achieve glycemic outcomes equivalent to, or even superior to, those of commercial systems.²⁴ The CODIAC study showed similar TIR values between AAPS (84.2%) and Control-IQ (85.7%) after 3 months of use.¹¹ Likewise, Wu et al. reported higher TIR with OS-AID (78.3%) compared with commercial systems (71.2%) in real-world settings over 12 weeks.²⁵

In our study, 88.9% of participants achieved TIR > 70% and TBR < 70 mg/dL < 4% at 6 months, and 81.5% maintained this at 12 months, exceeding the results from CREATE⁹ (64%) and Liarakos et al.¹⁷ (45.3%). These differences are more likely explained by the characteristics of real-world use in our cohort, where users were able to adjust settings freely and without protocol-imposed limitations, together with behavioral factors such as lower carbohydrate intake. Collectively, these elements may have contributed to achieving higher TIR values under routine clinical conditions. Higher proportions also reached HbA1c <7% (92.6% and 96.3% at 6 and 12 months) and ≤6.5% (85.2% and 81.5%), compared with Lum et al.¹² (75%) and Liarakos et al. (75.7% and 51.4%).

Unlike previous studies by Patel and Liarakos,^16,17 our analysis included additional metrics such as CV, TITR, and GRI. Glycemic stability improved notably, with 88.9% maintaining CV < 36% at 12 months. To our knowledge, this is the first study to assess the long-term effects of AAPS on TITR (70–140 mg/dL) and GRI.

Although no consensus exists on optimal TITR thresholds, values >50% are considered clinically relevant. In our cohort, 67% (n = 18) and 78% (n = 21) achieved TITR > 50% at 6 and 12 months, supporting its value as a complementary metric in AID users. TITR has emerged as a sensitive metric of euglycemia, complementing traditional indicators such as TIR, HbA1c, and CV. In our cohort, TITR increased by 14.8% at 6 months and 10.6% at 12 months (mean 67.8% and 63.6%), while GRI decreased from 40.2 to 21.3. These results exceed those of commercial systems, with TITR increases of only 5% and GRI above 28% in Control-IQ, and a +11.6% TITR (final 48.8%) with MiniMed™ 780G.²⁶

Despite good baseline control (HbA1c 6.6%), AAPS use further improved TIR, TITR, HbA1c, and CV, without increases in TBR or adverse events. Given the baseline CV (36.6%) above the recommended <36%, its reduction is clinically relevant and may lower long-term complication risk.²⁷ Although the margin for improvement was narrower in this well-controlled cohort, individuals with higher baseline HbA1c showed greater reductions, while those already within target still achieved relevant gains in complementary metrics such as glycemic stability. These findings suggest that AAPS may benefit both users with suboptimal control and those already meeting recommended goals.

Most users had higher education but no advanced technical expertise, indicating that AAPS can be adopted by individuals with diverse backgrounds. While education may ease initial adoption, sustained use appeared more related to user motivation and engagement than to technical proficiency. These aspects should be considered when interpreting the generalizability of our findings.

Strengths include the 12-month follow-up, exclusive evaluation of AAPS in adults, availability of rtCGM data before initiation, and use of emerging metrics such as TITR and GRI. Importantly, unlike most observational studies, all data were extracted from objective medical records during routine care, enhancing the robustness of the findings. Limitations include the retrospective design, lack of a control group, and possible self-selection bias. No structured patient-reported outcomes were collected. Although users frequently expressed high satisfaction during routine visits, validated measures of quality of life, treatment burden, or user experience were not obtained, preventing a formal assessment of the broader impact of AAPS. Although improvements were sustained at 12 months, small declines in TIR and TITR and a rise in TBR underscore the need for longer-term evaluation. Finally, the individualized configuration of AAPS parameters, despite routine clinical support, was not standardized and reflects real-world use, but may limit reproducibility and generalizability to other settings.

Conclusions

This study provides evidence supporting the long-term effectiveness and safety of the AAPS system in adults with T1D under real-world conditions. Sustained improvements in TIR, TITR, GRI, and HbA1c were observed without increased hypoglycemia, demonstrating AAPS as a viable and effective option for advanced management of T1D.

Authors’ Contributions

S.Amuedo, M.A.-G., and S.Azriel wrote the initial research proposal, reviewed the proposal and the study results, and wrote the article. S.Amuedo., M.A.-G, and S.Azriel reviewed and edited the research proposal and article, and contributed to the discussion. All authors read and approved the final version of this article.

Footnotes

Funding Information

No funding was received for this article.

Disclosure Statement

The authors declare that they have no known competing financial interests that could have appeared to influence the work reported in this article.

References

Weisman

, Bai

J-W

, Cardinez

, et al. Effect of artificial pancreas systems on glycaemic control in patients with type 1 diabetes: A systematic review and meta-analysis of outpatient randomised controlled trials. Lancet Diabetes Endocrinol, 2017; 5(7):501–512; doi: 10.1016/S2213-8587(17)30167-5

Lewis

, Leibrand

, #OpenAPS Community. Real-world use of open source artificial pancreas systems. J Diabetes Sci Technol, 2016; 10(6):1411; doi: 10.1177/1932296816665635

Toffanin

, Kozak

, Sumnik

, et al. In silico trials of an open-source android-based artificial pancreas: A new paradigm to test safety and efficacy of do-it-yourself systems. Diabetes Technol Ther, 2020; 22(2):112–120; doi: 10.1089/dia.2019.0375

Braune

, Hussain

, Lal

. The first regulatory clearance of an open-source automated insulin delivery algorithm. J Diabetes Sci Technol, 2023; 17(5):1139–1141; doi: 10.1177/19322968231164166

Nanayakkara

, Sharifi

, Burren

, et al. Hybrid closed loop using a do-it-yourself artificial pancreas system in adults with type 1 diabetes. J Diabetes Sci Technol, 2024; 18(4):889–896; doi: 10.1177/19322968231153882

Wilkinson

, Tomic

, Boyle

, et al. Study protocol for a randomised open-label clinical trial examining the safety and efficacy of the Android Artificial Pancreas System (AAPS) with advanced bolus-free features in adults with type 1 diabetes: The “CLOSE IT” (Closed Loop Open SourcE In Type 1 diabetes) trial. BMJ Open, 2024; 14(2):e078171; doi: 10.1136/bmjopen-2023-078171

Lei

, Lin

, Ling

, et al. Efficacy and safety of Android artificial pancreas system use at home among adults with type 1 diabetes mellitus in China: Protocol of a 26-week, free-living, randomised, open-label, two-arm, two-phase, crossover trial. BMJ Open, 2023; 13(8):e073263; doi: 10.1136/bmjopen-2023-073263

Gawrecki

, Zozulinska-Ziolkiewicz

, Michalak

, et al. Safety and glycemic outcomes of do-it-yourself AndroidAPS hybrid closed-loop system in adults with type 1 diabetes. PLoS One, 2021; 16(4):e0248965; doi: 10.1371/journal.pone.0248965

Burnside

, Lewis

, Crocket

, et al. Open-source automated insulin delivery in type 1 diabetes. N Engl J Med, 2022; 387(10):869–881; doi: 10.1056/NEJMoa2203913

10.

Burnside

, Lewis

, Crocket

, et al. Extended use of an open-source automated insulin delivery system in children and adults with type 1 diabetes: The 24-week continuation phase following the CREATE randomized controlled trial. Diabetes Technol Ther, 2023; 25(4):250–259; doi: 10.1089/dia.2022.0484

11.

, Hásková

, Radovnická

, et al. Comparison of Control-IQ and open-source AndroidAPS automated insulin delivery systems in adults with type 1 diabetes: The CODIAC study. Diabetes Obes Metab, 2024; 26(1):78–84; doi: 10.1111/dom.15289

12.

Lum

, Bailey

, Barnes-Lomen

, et al. A Real-World prospective study of the safety and effectiveness of the loop open source automated insulin delivery system. Diabetes Technol Ther, 2021; 23(5):367–375; doi: 10.1089/dia.2020.0535

13.

Fagerberg

, Borch

, Kristensen

, et al. Prevalence, safety, and metabolic control among danish children and adolescents with type 1 diabetes using open-source automated insulin delivery systems. Diabetes Technol Ther, 2024; 26(5):287–297; doi: 10.1089/dia.2023.0412

14.

Lewis

, Swain

, Donner

. Improvements in A1C and time-in-range in DIY closed-loop (OpenAPS) users. Diabetes, 2018; 67(Supplement_1):352-OR; doi: 10.2337/db18-352-OR

15.

, Luo

, Zheng

, et al. Use of a do-it-yourself artificial pancreas system is associated with better glucose management and higher quality of life among adults with type 1 diabetes. Ther Adv Endocrinol Metab, 2020; 11:2042018820950146; doi: 10.1177/2042018820950146

16.

Patel

, Crabtree

TSJ

, Taylor

, et al. Safety and effectiveness of do-it-yourself artificial pancreas system compared with continuous subcutaneous insulin infusions in combination with free style libre in people with type 1 diabetes. Diabet Med, 2022; 39(5):e14793; doi: 10.1111/dme.14793

17.

Liarakos

, Crabtree

TSJ

, Hussain

, et al. Long-term improvements in glycemia and user-reported outcomes associated with open-source automated insulin delivery systems in adults with type 1 diabetes in the United Kingdom: A real-world observational study. Diabetes Technol Ther, 2025; 27(4):283–291; doi: 10.1089/dia.2024.0556

18.

Melmer

, Züger

, Lewis

, et al. Glycaemic control in individuals with type 1 diabetes using an open source artificial pancreas system (OpenAPS). Diabetes Obes Metab, 2019; 21(10):2333–2337; doi: 10.1111/dom.13810

19.

Klonoff

, Wang

, Rodbard

, et al. A glycemia risk index (GRI) of Hypoglycemia and Hyperglycemia for continuous glucose monitoring validated by clinician ratings. J Diabetes Sci Technol, 2023; 17(5):1226–1242; doi: 10.1177/19322968221085273

20.

Rodbard

. Metrics to evaluate quality of glycemic control: Comparison of time in target, hypoglycemic, and hyperglycemic ranges with “Risk Indices. Diabetes Technol Ther, 2018; 20(5):325–334; doi: 10.1089/dia.2017.0416

21.

Laffel

, Sherr

, Liu

, et al. Limitations in achieving glycemic targets from CGM data and persistence of severe hypoglycemia in adults with type 1 diabetes regardless of insulin delivery method. Diabetes Care, 2025; 48(2):273–278; doi: 10.2337/dc24-1474

22.

Braune

, O’Donnell

, Cleal

, et al. Real-world use of do-it-yourself artificial pancreas systems in children and adolescents with type 1 diabetes: Online survey and analysis of self-reported clinical outcomes. JMIR Mhealth Uhealth, 2019; 7(7):e14087; doi: 10.2196/14087

23.

Petruzelkova

, Jiranova

, Soupal

, et al. Pre-school and school-aged children benefit from the switch from a sensor-augmented pump to an AndroidAPS hybrid closed loop: A retrospective analysis. Pediatr Diabetes, 2021; 22(4):594–604; doi: 10.1111/pedi.13190

24.

Jeyaventhan

, Gallen

, Choudhary

, et al. A real-world study of user characteristics, safety and efficacy of open-source closed-loop systems and Medtronic 670G. Diabetes Obes Metab, 2021; 23(8):1989–1994; doi: 10.1111/dom.14439

25.

, Lebbar

, Bonhoure

, et al. Open-source versus commercial automated insulin delivery system for type 1 diabetes management: A prospective observational comparative study from Canada. Diabetes Technol Ther, 2025; 27(7):517–526; doi: 10.1089/dia.2024.0561

26.

Rizzi

, Tartaglione

, Lucaccini Paoli

, et al. Evaluation of time in tight range and the glycaemia risk index in adults with type 1 diabetes using an advanced hybrid closed loop system: A 1-year real-world assessment. Diabetes Obes Metab, 2024; 26(9):4078–4086; doi: 10.1111/dom.15766

27.

Vigersky

, McMahon

. The relationship of hemoglobin A1C to time-in-range in patients with diabetes. Diabetes Technol Ther, 2019; 21(2):81–85; doi: 10.1089/dia.2018.0310