Impact of Hypoglycemia on Glucose Variability over Time for Individuals with Open-Source Automated Insulin Delivery Systems
1
CeADAR—Ireland’s Centre for AI, University College Dublin, D04 V2N9 Dublin, Ireland
2
OpenAPS, Seattle, WA 98101, USA
*
Author to whom correspondence should be addressed.
Diabetology 2024, 5(5), 514-536; https://doi.org/10.3390/diabetology5050038
Submission received: 19 September 2024
/
Revised: 15 October 2024
/
Accepted: 16 October 2024
/
Published: 21 October 2024
Abstract
:This study investigates glucose conditions preceding and following various hypoglycemia levels in individuals with type 1 diabetes using open-source automated insulin delivery (AID) systems. It also seeks to evaluate relationships between hypoglycemia and subsequent glycemic variability. Methods: Analysis of continuous glucose monitor (CGM) data from 122 adults with type 1 diabetes using open-source AID from the OpenAPS Data Commons was conducted. This study comprehensively analyzed the effects of hypoglycemia on glycemic variability, covering time periods before and after hypoglycemia. Results: Glucose variability normalization post-hypoglycemia can take up to 48 h, with severe hypoglycemia (41–50 mg/dL) linked to prolonged normalization. A cyclical pattern was observed where hypoglycemia predisposes individuals to further hypoglycemia, even with AID system use. A rise in glucose levels often precedes hypoglycemia, followed by an elevated mean time above range (TAR) post-hypoglycemia, indicating a ‘rebound’ effect. The experimental results are further validated on T1DEXI data (n = 222), originating from commercial AID systems. Different hypoglycemia categorization approaches did not show significant differences in glycemic variability outcomes. The level of hypoglycemia does influence the pattern of subsequent glucose fluctuations. Conclusion: Hypoglycemia, especially at lower levels, significantly impacts subsequent glycemic variability, even with the use of open-source AID systems. This should be studied further with a broader set of commercial AID systems to understand if these patterns are true of all types of AID systems. If these patterns occur in all types of AID systems, it underscores potential opportunities for enhancements in AID algorithms and highlights the importance of educating healthcare providers and people with diabetes about post-hypoglycemia glucose variability.
1. Introduction
In 2021, approximately 8.4 million people were living with type 1 diabetes, a figure projected to nearly double by 2040 [1]. Managing type 1 diabetes (T1D) involves frequent glucose monitoring and insulin administration, either manually or through continuous subcutaneous insulin infusion (CSII) pumps. Recent advancements like continuous glucose monitors (CGM) and automated insulin delivery (AID) systems are enhancing glycemic outcomes and quality of life for those with diabetes [2]. However, the delay in insulin’s peak effect and the pharmacokinetic curve [3] present challenges in synchronizing dosage with variables such as food intake and physical activity [4], which can influence glucose-related fluctuations and result in hypoglycemia.
Over 25 glucose variability (GV) and glucose-related metrics are prevalent in diabetes research. These metrics, accessible via open-source tools like cgmquantify and CGM-GUIDE, help to evaluate glucose data and understand glycemic variability [5,6,7,8,9,10,11,12,13,14]. AID systems improve time in the target glucose range (TIR) as well as time above range (TAR) and time below range (TBR); however, hypoglycemia (time spent below range) still occurs for a variety of reasons [3]. The American Diabetes Association categorizes hypoglycemia into “Level 1” (<70 mg/dL, but >54 mg/dL) and “Level 2” (between 40 and <54 mg/dL) [15]. Understanding the correlation of hypoglycemia with glucose variability is a possible focus area for improving clinical outcomes and quality of life, as previous studies indicate a predictive relationship between glucose variability and severe hypoglycemia [16,17].
Using an open-source AID dataset [18], this study explores the interplay between hypoglycemia, hyperglycemia, and glycemic variability-related metrics of continuous glucose monitor (CGM) sensor data. This includes ADA’s Level 1 and Level 2 designations, as well as additional levels of 41–50 mg/dL, 51–60 mg/dL, and 61–70 mg/dL, to evaluate whether there are distinct pattern differences associated with different levels of hypoglycemia. We sought to evaluate glycemic variability and identify patterns in CGM sensor glucose levels before and after different levels of hypoglycemia in users of open-source AID systems in real-world data. We hypothesize that an understanding of patterns in glucose levels and GV-related metrics before and after hypoglycemia could provide potential insights for AID system enhancement, specifically related to improving post-hypoglycemic glycemic variability and reducing the occurrence of additional hypoglycemia in individuals with insulin-requiring diabetes.
2. Materials and Methods
2.1. Dataset Collection and Description
The primary dataset used in this paper originates from the OpenAPS Data Commons [19], which allows for anonymized data donation [20] from real-world users of open-source AID (OS-AID) systems (which include OpenAPS, AndroidAAPS/AAPS, and Loop). These systems incorporate various commercial insulin pumps (such as older Medtronic pumps, Omnipod, and Bluetooth-enabled pumps like DANA*R/S, with rapid-acting insulin) and continuous glucose monitoring (CGM) sensors (such as Dexcom G4, G5, G6, and Medtronic sensors). The dataset features over 46,070 days of data and more than 10 million CGM data points, along with insulin dosing and algorithmic decision data. Participants’ demographics were self-reported, and the data were fully anonymized for research purposes. Further details of the data collection, cleaning process, and more detailed characterization of this version (n = 122 adults with diabetes) of the dataset can be found in our previously published papers [21,22], and all analyses described below use open-source processing scripts [23]. In addition to OpenAPS Data Commons, we validated the experimental findings on commercial AID systems by analyzing the Type 1 Diabetes EXercise Initiative (T1DEXI) (https://www.jaeb.org/t1dexi/, accessed on 27 January 2023) dataset collected as part of a 4-week exercise study in adults using AID (n = 222).
2.2. Glucose Analysis Metrics and Statistical Tests
The statistical and variability metrics for glucose analysis calculated in this paper include mean, minimum, and maximum of CGM sensor data, the first, second, and third quartiles, and the interquartile range. We also measure the interday standard deviation (SD), along with time outside range for both hypoglycemia (TOR < 70) and hyperglycemia (TOR > 180), and time in range (TIR). Moreover, the analysis considers the J_index (which stresses both the importance of the mean level and variability of glycemic levels) [24] as well as low blood glucose index (LBGI), high blood glucose index (HBGI), Coefficient of Variation (CV), and Glucose Management Indicator (GMI) [25] to evaluate glycemic variability.
To probe the relationship between hypoglycemia and glucose variability, we utilized a series of statistical tests, including the Shapiro–Wilk (SW) test for testing the normality of data distribution; the Z-test, which compares the means of hypoglycemia and glucose variability, to investigate any significant differences; and the Kolmogorov–Smirnov (KS) test, a nonparametric test to determine if two datasets differ significantly.
2.3. Experimental Workflow
This study is a retrospective longitudinal analysis aimed at understanding glycemic variability patterns before and after hypoglycemic events in individuals using open-source AID systems. The primary outcome was the impact of hypoglycemia on glucose analysis metrics, including TIR, TAR, and TBR as well as the aforementioned GV metrics. Secondary outcomes included the evaluation of rebound hyperglycemia and the cyclical nature of hypoglycemia.
A three-phased structured experimental workflow (Figure A1 in Appendix A) has been designed to investigate the impact of hypoglycemia, at different levels, on glucose variability (GV).
- Phase 1—Design Decisions: We selected the GV metrics of TIR, TOR < 70, TOR > 180, LBGI, HBGI, SD, POR, J_Index, CV, and GMI. We defined hypoglycemia levels at the following five different ranges: 41–50 mg/dL, 51–60 mg/dL, 61–70 mg/dL, 40–54 mg/dL, and 55–70 mg/dL. This includes more granular, 10-point ranges, as well as ADA-defined level 1 and 2 ranges, to determine whether a tighter hypoglycemia range influenced resulting patterns.
- Phase 2—Data Preparation: We used identical pre-processing and preparation methods from previous work [22].
- Phase 3—Out-of-Whack Analyser (visualized in Figure 1): We computed GV-related metrics before and after each instance of hypoglycemia at various intervals: 3, 6, 12, 24, and 48 h. A hypoglycemic event, or instance of hypoglycemia, was defined as any instance where sensor glucose fell below or into the specified range. Successive data points of hypoglycemia after the first were also identified and the entire contiguous sequence is considered a hypoglycemia segment, with the GV calculated before and after this segment. In some cases, where sensor glucose returned above range and then returned into the defined hypoglycemia range, a second hypoglycemic segment was recorded and separately analyzed. This provided a comprehensive view of the impact of different levels of hypoglycemia on GV over different timeframes. Subsequently, we calculated mean statistics for each individual, providing a representative measure for comparison across subjects. Lastly, we conducted a distribution analysis to visualize and interpret the mean GV distributions before and after hypoglycemia.
Figure 1.
Process flow diagram for three-stage out-of-whack analyser.
Data are organized into tables (Appendix A) to provide a comprehensive view of glucose analysis and variability metrics including TBR, TIR, TOR, SD, and J_Index for various time periods surrounding the instances of hypoglycemia, respectively. To quantify these metrics, we incorporated several statistical measures: mean, standard deviation (SD), minimum (Min), first quartile (25%), median (50%), third quartile (75%), maximum (Max), Shapiro–Wilk test (SW-test, to test for normal distribution), Z-test (to evaluate significant differences in GV metrics at various time intervals around hypoglycemic events), and Kolmogorov–Smirnov test (KS-test, to evaluate differences in GV distributions before and after hypoglycemic events).
The organization of data spans from 48 h before (−48 h) to 48 h after (+48 h) hypoglycemia, with interval granularity increasing as the event approaches. These periods are further subdivided into five categories of hypoglycemia: 41–50 mg/dL, 51–60 mg/dL, 61–70 mg/dL, 40–54 mg/dL, and 55–70 mg/dL. Additional experimental validation on commercial AID users from the T1DEXI dataset was performed for ADA’s Level 1 hypoglycemia category (40–54 mg/dL).
To facilitate visual inspection and trend analysis, figures are designated to represent these data across the first three hypoglycemia categories (41–50 mg/dL, 51–60 mg/dL, and 61–70 mg/dL), including Figure 2A–C and Figure 3A,B (and Appendix A Figure A2A,B) for TBR, TIR, TAR, SD, J_Index, HBGI, and LBGI, respectively. Additional figures (Appendix A, Figure A3A–C, Figure A4A,B and Figure A5A,B), focus on the remaining hypoglycemia levels (40–54 mg/dL and 55–70 mg/dL).
3. Results
This analysis of instances of hypoglycemia in the OpenAPS Data Commons dataset revealed how different levels of hypoglycemia affect glycemic variability over time. The levels of hypoglycemia have an impact on the patterns of glucose levels both before and after such events, with notable variations in GV metrics across different categories or levels of hypoglycemia. Specifically, the data show that glucose levels often begin to fluctuate significantly as hypoglycemia approaches, with this variability extending up to 48 h post-event before normalizing.
We define GV ‘normalization’ as the point at which the distribution of glucose variability and analysis metrics returns to its overall distribution within the population dataset. Testing across different time horizons demonstrated that, on average, 48 h was sufficient for these metrics to stabilize after hypoglycemic events. Although individual variability and factors such as physical activity could influence glycemic patterns, extending the analysis beyond 48 h showed no significant differences in the recovery patterns, leading us to conclude that 48 h represents a reasonable period for normalization at the population level.
These patterns do not follow a normal distribution and there are statistically significant differences in glucose metrics between the various hypoglycemia severity categories. Although we calculated CV and GMI, we found no clear pattern connecting them to hypoglycemia at the studied timeframes.
Notably, the level of hypoglycemia has a significant impact on the patterns of glucose levels before and after hypoglycemia (Table A1 in Appendix A). In the periods leading up to the hypoglycemia (indicated by the negative time intervals), the average TBR increases as the time period approaches the event (Figure 2A and Figure 3A). In other words, a prolonged period of lower glucose levels often precedes other instances of hypoglycemia, regardless of the level of hypoglycemia.
However, the actual mean values for TBR differ between the categories. Lower hypoglycemia (41–50 mg/dL) is generally preceded by longer periods of low glucose levels than less severe lows (61–70 mg/dL). This indicates that the level of hypoglycemia may be influenced by the duration and severity of the preceding low-glucose period.
In the periods following the hypoglycemia (indicated by the positive time intervals), the average TBR decreases, indicating a recovery in glucose levels. This recovery pattern is observed across all severity categories, suggesting that glucose levels tend to stabilize after hypoglycemia, regardless of the event’s severity. The speed and extent of this recovery appear to vary between the categories. Following more severe hypoglycemia (41–50 mg/dL), the average TBR remains higher in the initial hours after the event, suggesting a slower recovery. In contrast, following less severe events (61–70 mg/dL), the average TBR decreases more rapidly, indicating a quicker recovery.
However, it can be observed from the box-plot distributions in Figure 2A and Figure A3A that it may take up to 48 h to completely stabilize TBR as per normal (or overall) TBR distribution.
When looking at hypoglycemia 41–50 mg/dL, we observe that the mean TIR tends to decrease with shorter time intervals around the hypoglycemia (Table A2 in Appendix A). For example, the mean TIR is 74.89% for the 48 h window but decreases to 62.25% for the 3 h window. This suggests that lower hypoglycemia has a substantial impact on TIR, particularly in the immediate hours surrounding the event. For less severe hypoglycemia (51–60 mg/dL and 61–70 mg/dL), a similar pattern can be observed, with TIR decreasing as the time interval around the hypoglycemia becomes shorter. The SW-test, KS-test, and Z-test show statistically significant differences, which suggests that the TIR distribution varies significantly depending on the severity of the hypoglycemia and the time interval around the event.
It can also be observed 48 h before the hypoglycemia that individuals tend to experience higher glucose levels in the days leading up to a hypoglycemia, irrespective of the event’s severity (Table A3 in Appendix A). In the period following the hypoglycemia (from +3 h to +48 h), the mean TAR first decreases and then increases across all categories. The severity of the hypoglycemia appears to influence these patterns. Following hypoglycemia in the lowest range (41–50 mg/dL), the mean TAR tends to be higher, indicating a greater tendency towards hyperglycemia, and what people with diabetes might colloquially describe as a “rebound”. Following less severe hypoglycemia (61–70 mg/dL), the mean TAR > 180 tends to be lower, indicating a lesser tendency towards rebound hyperglycemia.
In terms of standard deviation, as hypoglycemia approaches, the SD decreases, suggesting that glucose levels become less variable (Table A4 in Appendix A). For instance, the mean SD decreases from 48.04 mg/dL at −48 h to 34.43 mg/dL at −3 h for the 41–50 mg/dL category. This pattern is observed across all categories of hypoglycemia. Lower glucose levels are associated with smaller SD values both before and after the event.
The mean J_Index score tends to decrease as the time period gets closer to the hypoglycemia, both before and after the event (Table A5 in Appendix A). This further confirms that the glucose outcomes are less optimal closer to the hypoglycemia. There is also a general trend of decreasing J_Index score associated with lower glucose, indicating worse glucose outcomes associated with increased severity of hypoglycemia.
4. Discussion
Hypoglycemia remains a significant challenge in diabetes management, and understanding the patterns and after-effects of different levels of hypoglycemia is critical to improving therapeutic approaches and improving outcomes. In our study, in evaluating the relationship between different levels of hypoglycemia and resulting patterns in glucose outcomes using CGM data from PWD using OS-AID systems, we observed that glucose levels can take up to 48 h to stabilize post-hypoglycemia. More pronounced lows (e.g., 41–50 mg/dL) correlate with prolonged normalization times. This observation extends the common understanding that the after-effects of hypoglycemia are not immediate, and the ‘shadow’ of an episode of lower glucose can last for up to two days.
Furthermore, our data underscore the cyclic nature of hypoglycemia: an episode of low glucose is observed to be associated with subsequent hypoglycemia. This trend persists even with advanced diabetes therapies like automated insulin delivery (AID) systems. Particularly at levels of 41–50 mg/dL, in this dataset, we observed a greater risk of recurring hypoglycemia within 48 h than at slightly higher levels (61–70 mg/dL). A similarly noteworthy trend is also the heightened glucose levels observed in the days leading up to hypoglycemia. As hypoglycemia draws near, the duration spent above the optimal range decreases. Following severe hypoglycemia (41–50 mg/dL), we observed an elevated mean time above range (TAR), a phenomenon often termed a “rebound” in the diabetes community. This mirrors previous studies and understandings of hypoglycemia in T1D where higher fluctuations (and thus increased GV) increase the likelihood of subsequent instances of hypoglycemia; yet, it remains unclear whether this is a cause or a consequence of hypoglycemia itself [26].
The distribution of time in range (TIR) is influenced by the severity of the preceding hypoglycemia. Higher levels of hypoglycemia disturb TIR to a lesser degree than lower hypoglycemia, potentially due to their closer proximity to the optimal range (70–180 mg/dL). While ADA defines hypoglycemia based on two levels, with “Level 1” being <70 mg/dL and >54 mg/dL and “Level 2” < 54 mg/dL, we sought to explore the relationship in more narrow categories and evaluated glycemic variability and outcomes related to hypoglycemia at ranges of 41–50 mg/dL, 51–60 mg/dL, and 61–70 mg/dL. This categorization was in pursuit of a deeper understanding of how the intensity of lows impacts glycemic variability and improves the quantification of hypoglycemia associated with glycemic variability, which has not previously been analyzed.
While the relationship between hypoglycemia and glycemic variability is increasingly studied using CGM data, glycemic variability is often assessed for an overarching period (e.g., days to weeks or longer) [27] and/or blocks of time (e.g., day versus night), rather than specifically in relationship to instances of hypoglycemia [28]. The previous literature investigating the glycemic variability and hypoglycemia relationship often relies on computations from fingerstick blood glucose testing (SMBG) [16], although there are a few studies increasingly using CGM data [26], more closely mirroring our analyses, albeit in a much smaller dataset. We found that the level of hypoglycemia does indeed affect the intensity of glycemic variability distributions, but we found no difference in the outcomes between the three smaller ranges (41–50, 51–60, 61–70 mg/dL) and the broader Level 1 (<70, >54 mg/dL) and Level 2 (<54 mg/dL) categorization used historically. If these patterns are repeatedly observed (as planned in future studies) across other AID datasets, then future glycemic variability analyses in relation to hypoglycemia likely can use the more general Level 1 and Level 2 hypoglycemia categorizations. It is also worth highlighting that clinical readers may have a perception of hypoglycemia event categorizations related to historical definitions of hypoglycemia, validated by fingerstick blood glucose testing (often referred to as SMBG). In our study, we found that evaluating instances of hypoglycemia separated by a single data point (e.g., where sensor glucose returned above 70 mg/dL for a single data point and then dipped to <70 mg/dL again), rather than defining this time period collectively as a single instance of hypoglycemia, did not influence the patterns observed, especially as we focused on the broader time periods of 3 h or more following the segment(s) of hypoglycemia in sensor glucose levels.
This study provides insights into the patterns of glycemic variability among insulin-requiring individuals using OS-AID systems across different levels of hypoglycemia. The observed GV patterns pre- and post-hypoglycemia are essential for advancing diabetes therapy, as they can inform the development of predictive models and automated remediation strategies for hypoglycemia based on these patterns. Understanding these patterns is vital for refining AID algorithms to improve patient outcomes by reducing the incidence and impact of hypoglycemia, particularly when they may be influenced by user behavior or non-controllable variables such as temperature or hormone changes. Furthermore, the preliminary analysis from adult participants in the T1DEXI dataset confirms that these patterns also exist among users of multiple types of commercial AID systems (Figure A6 and Table A6).
While our initial analysis did not find any association between rebound hyperglycemia rates and severity with specific clinical features (e.g., gender, age), this may be due to the limited demographic and clinical data available in the donated real-world dataset. Future research will aim to include more comprehensive demographic and clinical information to better understand these associations. Additional analysis in the future will also evaluate potential determinants of these patterns (at an individual level), including assessment of insulin activity prior to hypoglycemia, glucose targets in the AID system, as well as patterns of carbohydrate intake related to treating hypoglycemia. These studies could possibly then collectively contribute to the development of strategies that minimize these patterns, focusing on behavioral responses to hypoglycemia, insulin activity levels preceding, and following hypoglycemia, and both user and system interventions that might perpetuate recurrent hypoglycemia.
Limitations
This work leverages retrospective data from insulin-requiring individuals using open-source automated insulin delivery (AID) systems. Further, the analysis to date focused primarily on establishing patterns related to glucose data, due to the complex, retrospective nature of the real-world dataset where variables such as IOB are not perfectly logged due to the intermittent upload nature of the dataset. As such, forthcoming work will replicate using different datasets that include commercial AID user data (already underway) to determine whether these glycemic variability patterns related to hypoglycemia also pertain to non-AID users. Target glucose level, which can be customized by users of OS-AID to different levels than those available in commercial AID systems, was not factored into the current analysis; but, the user-set target of OS-AID or targets of commercial AID systems should be considered in future analyses related to patterns of hypoglycemia. Similarly, carbohydrate corrections and carbohydrate intake overall should be analyzed in future studies. Some may hypothesize that large amounts of carb correction may drive subsequent swings in glucose levels, yet many individuals using AID systems correct with smaller amounts of carbohydrate intake because the AID system has reduced insulin levels, which also addresses hypoglycemia correction. A deeper look into the interplay in both carb intake and insulin reduction, associated with subsequent variability, would be beneficial.
For this work, we chose a 48 h window preceding and following hypoglycemia, with a resolution down to 3 h pre- and post-hypoglycemia. In some cases, there appears to be an increase in metrics around the 6 h mark, such as TBR < 70 mg/dL and TAR > 180 mg/dL. Future work should evaluate other time intervals at smaller time intervals to understand whether this increase is an artifact of the chosen time intervals or a significant increase at that particular time. This may also be influenced by the time of insulin action, which is often around 5–7 h when used in AID systems and is another factor to be explored in future work.
5. Conclusions
This paper investigated the dynamics of different levels of hypoglycemia and glycemic variability in people with type 1 diabetes using open-source automated insulin delivery (AID) systems. Our findings revealed that a single instance of hypoglycemia is related to subsequent hypoglycemia in a short timeframe and that hyperglycemia often precedes significant hypoglycemia. This study showed that glucose variability can take up to 48 h to stabilize post-hypoglycemia, with more pronounced hypoglycemia correlating with prolonged normalization times, even with the use of AID systems. Such insights suggest that future refinements in automated insulin delivery (AID) system algorithms could enhance glucose management following instances of hypoglycemia and periods of more extreme glucose fluctuations and that there are areas of opportunity for increasing education for diabetes care providers and people living with diabetes about post-hypoglycemia timeframes.
Author Contributions
Conceptualization, A.S. and D.M.L.; methodology, A.S. and D.M.L.; software, A.S. and D.M.L.; validation, A.S. and D.M.L.; formal analysis, A.S. and D.M.L.; investigation, A.S. and D.M.L.; resources, A.S. and D.M.L.; data curation, A.S. and D.M.L.; writing—original draft preparation, A.S. and D.M.L.; writing—review and editing, A.S. and D.M.L.; visualization, A.S. and D.M.L.; supervision, A.S. and D.M.L.; project administration, A.S. and D.M.L.; funding acquisition, A.S. and D.M.L. All authors have read and agreed to the published version of the manuscript.
Funding
This work was in part performed under a grant from the Leona M. and Harry B. Helmsley Charitable Trust (Grant #: 2407-07175).
Institutional Review Board Statement
Ethics approval was not required, as this work evaluated data from a retrospective, already collected, anonymized dataset.
Informed Consent Statement
Not applicable.
Data Availability Statement
All programming scripts and tools developed for the analysis in this paper are made public and online, with each source cited within the paper. Data accessed for this paper are part of the OpenAPS Data Commons.
Acknowledgments
Thank you to members of the diabetes community who have donated their data from a variety of open-source automated insulin delivery (AID) systems to the OpenAPS Data Commons.
Conflicts of Interest
The authors declare no financial conflicts of interest. DML is a volunteer developer of one of the open-source AID systems, OpenAPS. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.
Appendix A
Figure A1.
Experimental workflow.
Table A1.
Statistics of glucose analysis and variability metrics for time below range (TBR) or time outside range less than 70 mg/dL (TOR < 70%) for various time periods surrounding the instances of hypoglycemia, respectively. The metrics include mean, standard deviation (SD), minimum (Min), first quartile (25%), median (50%), third quartile (75%), maximum (Max), Shapiro–Wilk test (SW-test), Z-test, and Kolmogorov–Smirnov test (KS-test). The results of the SW-test, Z-test, and KS-test provide information about the normality of the data and the equality of distribution, respectively. In all cases for SW-test except 24 h ahead of hypoglycemia in the range of 41–50 mg/dL, p < 0.05 indicates that the data do not follow a normal distribution. Furthermore, Z-test and KS-test indicate that the distributions differ significantly from each other (p < 0.05).
Table A1.
Statistics of glucose analysis and variability metrics for time below range (TBR) or time outside range less than 70 mg/dL (TOR < 70%) for various time periods surrounding the instances of hypoglycemia, respectively. The metrics include mean, standard deviation (SD), minimum (Min), first quartile (25%), median (50%), third quartile (75%), maximum (Max), Shapiro–Wilk test (SW-test), Z-test, and Kolmogorov–Smirnov test (KS-test). The results of the SW-test, Z-test, and KS-test provide information about the normality of the data and the equality of distribution, respectively. In all cases for SW-test except 24 h ahead of hypoglycemia in the range of 41–50 mg/dL, p < 0.05 indicates that the data do not follow a normal distribution. Furthermore, Z-test and KS-test indicate that the distributions differ significantly from each other (p < 0.05).
| Mean (%) | SD (%) | Min (%) | 25% | 50% | 75% | Max (%) | SW-Test | Z-Test | KS-Test | |
|---|---|---|---|---|---|---|---|---|---|---|
| Overall | 4.09 | 3.06 | 0.23 | 1.74 | 3.33 | 5.69 | 16.97 | <0.05 | ||
| −48 h (41–50) | 6.48 | 3.79 | 0 | 3.65 | 5.82 | 8.81 | 21.18 | <0.05 | <0.05 | <0.05 |
| −48 h (51–60) | 6.44 | 3.82 | 0.91 | 3.82 | 5.29 | 8.94 | 20.68 | <0.05 | <0.05 | <0.05 |
| −48 h (61–70) | 6.44 | 3.82 | 0.91 | 3.82 | 5.29 | 8.94 | 20.68 | <0.05 | <0.05 | <0.05 |
| −24 h (41–50) | 8.02 | 4.08 | 0 | 4.92 | 7.13 | 10.74 | 21.89 | <0.05 | <0.05 | <0.05 |
| −24 h (51–60) | 7.99 | 4.2 | 1.07 | 5.12 | 6.77 | 10.63 | 22.91 | <0.05 | <0.05 | <0.05 |
| −24 h (61–70) | 7.99 | 4.2 | 1.07 | 5.12 | 6.77 | 10.63 | 22.91 | <0.05 | <0.05 | <0.05 |
| −12 h (41–50) | 10.58 | 5.1 | 0 | 6.91 | 9.76 | 13.92 | 26.39 | <0.05 | <0.05 | <0.05 |
| −12 h (51–60) | 10.46 | 5.06 | 1.36 | 7.1 | 9.2 | 13.25 | 31.56 | <0.05 | <0.05 | <0.05 |
| −12 h (61–70) | 9.07 | 4.33 | 1.84 | 5.99 | 8.07 | 11.19 | 27.19 | <0.05 | <0.05 | <0.05 |
| −6 h (41–50) | 16.23 | 7.96 | 0 | 10.95 | 14.87 | 19.69 | 48.61 | <0.05 | <0.05 | <0.05 |
| −6 h (51–60) | 15.72 | 7.31 | 2.72 | 10.55 | 14.67 | 19.28 | 42.67 | <0.05 | <0.05 | <0.05 |
| −6 h (61–70) | 13.26 | 5.98 | 3.68 | 8.91 | 12.36 | 16.21 | 36.45 | <0.05 | <0.05 | <0.05 |
| −3 h (41–50) | 25.93 | 11.85 | 0 | 18.58 | 24.15 | 31.8 | 58.33 | <0.05 | <0.05 | <0.05 |
| −3 h (51–60) | 25.02 | 10.71 | 5.45 | 17.44 | 23.98 | 30.03 | 58.24 | <0.05 | <0.05 | <0.05 |
| −3 h (61–70) | 25.02 | 10.71 | 5.45 | 17.44 | 23.98 | 30.03 | 58.24 | <0.05 | <0.05 | <0.05 |
| +3 h (41–50) | 31.84 | 11.92 | 2.78 | 22.66 | 30.46 | 37.63 | 73.35 | <0.05 | <0.05 | <0.05 |
| +3 h (51–60) | 29.05 | 10.16 | 12.56 | 21.42 | 27.85 | 34.11 | 65.09 | <0.05 | <0.05 | <0.05 |
| +3 h (61–70) | 22.4 | 8.05 | 8.95 | 16.8 | 21.49 | 26.9 | 50.53 | <0.05 | <0.05 | <0.05 |
| +6 h (41–50) | 19.15 | 7.77 | 1.39 | 13.61 | 18.95 | 23.2 | 51.98 | <0.05 | <0.05 | <0.05 |
| +6 h (51–60) | 17.88 | 7.1 | 6.45 | 12.61 | 16.81 | 21.59 | 43.22 | <0.05 | <0.05 | <0.05 |
| +6 h (61–70) | 14.16 | 5.86 | 4.84 | 9.74 | 13.06 | 17.47 | 37.68 | <0.05 | <0.05 | <0.05 |
| +12 h (41–50) | 12.4 | 5.06 | 0.69 | 9.1 | 12.24 | 15.19 | 28.35 | <0.05 | <0.05 | <0.05 |
| +12 h (51–60) | 11.53 | 4.91 | 3.24 | 8.28 | 10.33 | 14.46 | 31.28 | <0.05 | <0.05 | <0.05 |
| +12 h (61–70) | 9.48 | 4.27 | 2.42 | 6.46 | 8.42 | 11.97 | 28.46 | <0.05 | <0.05 | <0.05 |
| +24 h (41–50) | 9.24 | 4.27 | 0.35 | 6.16 | 8.98 | 12.01 | 23.81 | 0.059 | <0.05 | <0.05 |
| +24 h (51–60) | 8.44 | 4 | 2.26 | 5.58 | 7.44 | 11.07 | 22.47 | <0.05 | <0.05 | <0.05 |
| +24 h (61–70) | 7.25 | 3.73 | 1.7 | 4.7 | 6.1 | 9.6 | 21.46 | <0.05 | <0.05 | <0.05 |
| +48 h (41–50) | 7.08 | 3.7 | 0.17 | 4.44 | 6.33 | 9.84 | 21.24 | <0.05 | <0.05 | <0.05 |
| +48 h (51–60) | 6.68 | 3.72 | 1.25 | 4.22 | 5.79 | 8.91 | 20.63 | <0.05 | <0.05 | <0.05 |
| +48 h (61–70) | 5.98 | 3.55 | 1.08 | 3.62 | 5.06 | 8.05 | 19.41 | <0.05 | <0.05 | <0.05 |
| −48 h (40–54) | 6.4 | 3.89 | 0.86 | 3.56 | 5.71 | 8.95 | 20.96 | <0.05 | <0.05 | <0.05 |
| −48 h (55–70) | 5.88 | 3.67 | 0.96 | 3.5 | 4.96 | 7.89 | 19.79 | <0.05 | <0.05 | <0.05 |
| −24 h (40–54) | 7.91 | 4.25 | 1.43 | 4.69 | 6.9 | 10.45 | 23.26 | <0.05 | <0.05 | <0.05 |
| −24 h (55–70) | 7.1 | 3.92 | 1.58 | 4.58 | 5.99 | 9.45 | 21.58 | <0.05 | <0.05 | <0.05 |
| −12 h (40–54) | 10.43 | 5.22 | 1.19 | 6.76 | 9.74 | 13.11 | 32.4 | <0.05 | <0.05 | <0.05 |
| −12 h (55–70) | 9.18 | 4.58 | 1.82 | 5.96 | 8.18 | 11.63 | 28.6 | <0.05 | <0.05 | <0.05 |
| −6 h (40–54) | 15.89 | 7.78 | 2.38 | 10.11 | 14.93 | 19.93 | 45.74 | <0.05 | <0.05 | <0.05 |
| −6 h (55–70) | 13.43 | 6.37 | 3.65 | 8.59 | 12.65 | 16.16 | 37.79 | <0.05 | <0.05 | <0.05 |
| −3 h (40–54) | 25.55 | 11.9 | 4.76 | 16.1 | 24.08 | 31.77 | 61.8 | <0.05 | <0.05 | <0.05 |
| −3 h (55–70) | 21.05 | 8.91 | 6.09 | 14.33 | 19.93 | 26.11 | 54.91 | <0.05 | <0.05 | <0.05 |
| +3 h (40–54) | 30.99 | 11.9 | 10.34 | 22.29 | 30.19 | 38.08 | 72.83 | <0.05 | <0.05 | <0.05 |
| +3 h (55–70) | 23.3 | 8.62 | 9.21 | 17.24 | 22.21 | 28.04 | 51.81 | <0.05 | <0.05 | <0.05 |
| +6 h (40–54) | 18.74 | 8.01 | 6.21 | 12.74 | 18.34 | 22.83 | 50.41 | <0.05 | <0.05 | <0.05 |
| +6 h (55–70) | 14.98 | 6.34 | 5.43 | 10.16 | 14.16 | 18.09 | 40.05 | <0.05 | <0.05 | <0.05 |
| +12 h (40–54) | 12.2 | 5.29 | 3.57 | 8.49 | 11.99 | 14.87 | 33.24 | <0.05 | <0.05 | <0.05 |
| +12 h (55–70) | 9.7 | 4.48 | 2.62 | 6.67 | 8.7 | 12.16 | 29.1 | <0.05 | <0.05 | <0.05 |
| +24 h (40–54) | 8.93 | 4.31 | 1.81 | 5.68 | 8.71 | 11.71 | 23.87 | <0.05 | <0.05 | <0.05 |
| +24 h (55–70) | 7.34 | 3.84 | 1.81 | 4.74 | 6.21 | 9.41 | 21.62 | <0.05 | <0.05 | <0.05 |
| +48 h (40–54) | 6.93 | 3.88 | 1.36 | 4.13 | 6.01 | 9.64 | 21.5 | <0.05 | <0.05 | <0.05 |
| +48 h (55–70) | 6.1 | 3.65 | 1.11 | 3.68 | 5.09 | 8.17 | 19.97 | <0.05 | <0.05 | <0.0 |
Table A2.
Statistics of glucose analysis and variability metrics for time in range (TIR) for various time periods surrounding the instances of hypoglycemia, respectively. The metrics include mean, standard deviation (SD), minimum (Min), first quartile (25%), median (50%), third quartile (75%), maximum (Max), Shapiro–Wilk test (SW-test), Z-test, and Kolmogorov–Smirnov test (KS-test). The SW-test results indicate that the TIR distribution is not normally distributed for most of the groups (p < 0.05). The Z-test results show significant differences (p < 0.05) between the mean TIR of many groups and the overall mean TIR. The KS-test results show significant differences (p < 0.05) in the TIR distribution between many groups and the overall TIR distribution.
Table A2.
Statistics of glucose analysis and variability metrics for time in range (TIR) for various time periods surrounding the instances of hypoglycemia, respectively. The metrics include mean, standard deviation (SD), minimum (Min), first quartile (25%), median (50%), third quartile (75%), maximum (Max), Shapiro–Wilk test (SW-test), Z-test, and Kolmogorov–Smirnov test (KS-test). The SW-test results indicate that the TIR distribution is not normally distributed for most of the groups (p < 0.05). The Z-test results show significant differences (p < 0.05) between the mean TIR of many groups and the overall mean TIR. The KS-test results show significant differences (p < 0.05) in the TIR distribution between many groups and the overall TIR distribution.
| Mean (%) | SD (%) | Min (%) | 25% | 50% | 75% | Max (%) | SW-Test | Z-Test | KS-Test | |
|---|---|---|---|---|---|---|---|---|---|---|
| Overall | 76.99 | 8.89 | 49.75 | 71.32 | 77.91 | 83.67 | 96.87 | 0.662 | ||
| −48 h (41–50) | 74.89 | 9.25 | 38.68 | 70.64 | 75.2 | 80.84 | 93.91 | <0.05 | 0.098 | 0.293 |
| −48 h (51–60) | 75.59 | 8.78 | 48.52 | 70.71 | 76.66 | 81.43 | 95.3 | 0.172 | 0.259 | 0.293 |
| −48 h (61–70) | 75.59 | 8.78 | 48.52 | 70.71 | 76.66 | 81.43 | 95.3 | 0.172 | 0.259 | 0.293 |
| −24 h (41–50) | 73.14 | 9.82 | 40.58 | 67.61 | 74.07 | 79.75 | 91.98 | <0.05 | <0.05 | 0.058 |
| −24 h (51–60) | 74.59 | 8.45 | 49.9 | 69.41 | 75.27 | 80.01 | 94.78 | 0.321 | <0.05 | 0.118 |
| −24 h (61–70) | 74.59 | 8.45 | 49.9 | 69.41 | 75.27 | 80.01 | 94.78 | 0.321 | <0.05 | 0.118 |
| −12 h (41–50) | 71.27 | 10.1 | 38.8 | 67.07 | 71.39 | 77.48 | 90.14 | <0.05 | <0.05 | <0.05 |
| −12 h (51–60) | 73.05 | 8.46 | 52.79 | 68.19 | 73.21 | 78.27 | 92.04 | 0.098 | <0.05 | <0.05 |
| −12 h (61–70) | 75.32 | 8.22 | 50.33 | 70.45 | 75.73 | 80.67 | 93.81 | 0.085 | 0.164 | 0.163 |
| −6 h (41–50) | 67.12 | 11.1 | 24.4 | 63.06 | 68.82 | 72.95 | 93.06 | <0.05 | <0.05 | <0.05 |
| −6 h (51–60) | 69.63 | 8.53 | 47.41 | 65.66 | 70.3 | 74.52 | 87.36 | <0.05 | <0.05 | <0.05 |
| −6 h (61–70) | 73.38 | 7.71 | 51.31 | 70.17 | 74.2 | 78.33 | 90.19 | <0.05 | <0.05 | <0.05 |
| −3 h (41–50) | 62.25 | 11.1 | 28.17 | 56.92 | 63.53 | 70.25 | 86.11 | <0.05 | <0.05 | <0.05 |
| −3 h (51–60) | 64.89 | 9.47 | 35.44 | 60.6 | 65.69 | 71.18 | 85.11 | <0.05 | <0.05 | <0.05 |
| −3 h (61–70) | 64.89 | 9.47 | 35.44 | 60.6 | 65.69 | 71.18 | 85.11 | <0.05 | <0.05 | <0.05 |
| +3 h (41–50) | 59.14 | 10.4 | 26.52 | 52.65 | 60.7 | 66.4 | 94.44 | <0.05 | <0.05 | <0.05 |
| +3 h (51–60) | 63.26 | 8.71 | 34.74 | 58.85 | 64.44 | 68.3 | 78.98 | <0.05 | <0.05 | <0.05 |
| +3 h (61–70) | 71.1 | 6.99 | 47.54 | 67.32 | 71.66 | 75.66 | 83.74 | <0.05 | <0.05 | <0.05 |
| +6 h (41–50) | 64.81 | 9.98 | 39.93 | 58.95 | 65.08 | 70.45 | 97.22 | 0.22 | <0.05 | <0.05 |
| +6 h (51–60) | 67.8 | 8.19 | 43.32 | 64.42 | 69.16 | 72.51 | 84.82 | <0.05 | <0.05 | <0.05 |
| +6 h (61–70) | 72.94 | 6.97 | 56.41 | 69.31 | 73.28 | 77.42 | 87.17 | <0.05 | <0.05 | <0.05 |
| +12 h (41–50) | 69.68 | 9.59 | 45.29 | 64.62 | 70.33 | 75.24 | 98.61 | 0.25 | <0.05 | <0.05 |
| +12 h (51–60) | 71.64 | 8.31 | 47.63 | 66.85 | 72.36 | 76.8 | 91.04 | 0.355 | <0.05 | <0.05 |
| +12 h (61–70) | 74.85 | 7.73 | 52.15 | 69.38 | 75.76 | 79.98 | 92.52 | 0.187 | 0.067 | <0.05 |
| +24 h (41–50) | 73.02 | 8.96 | 51.39 | 67.97 | 73.45 | 77.5 | 99.31 | 0.335 | <0.05 | <0.05 |
| +24 h (51–60) | 74.38 | 8.24 | 52.65 | 70.02 | 74.55 | 80.02 | 94.56 | 0.56 | <0.05 | 0.058 |
| +24 h (61–70) | 76.38 | 7.84 | 55.7 | 72.11 | 77.26 | 81.8 | 94.96 | 0.268 | 0.604 | 0.38 |
| +48 h (41–50) | 74.95 | 8.13 | 54.63 | 70.41 | 75.24 | 79.6 | 94.7 | 0.764 | 0.087 | 0.163 |
| +48 h (51–60) | 75.78 | 8.07 | 53.19 | 71.14 | 76.11 | 81.11 | 94.8 | 0.389 | 0.311 | 0.38 |
| +48 h (61–70) | 77.09 | 7.91 | 54.91 | 72.1 | 77.84 | 82.21 | 95.87 | 0.572 | 0.931 | 0.914 |
| −48 h (40–54) | 75.16 | 9.45 | 39.04 | 69.56 | 75.87 | 81.39 | 94.38 | <0.05 | 0.159 | 0.298 |
| −48 h (55–70) | 76.75 | 8.54 | 49.23 | 71.64 | 77.93 | 82.69 | 95.45 | 0.122 | 0.865 | 0.995 |
| −24 h (40–54) | 73.91 | 9.41 | 44.09 | 68.73 | 74.29 | 80.76 | 93.03 | 0.169 | <0.05 | 0.086 |
| −24 h (55–70) | 76.19 | 8.28 | 50.84 | 70.48 | 77.59 | 81.97 | 94.76 | 0.119 | 0.523 | 0.829 |
| −12 h (40–54) | 72.23 | 9.83 | 40.93 | 66.14 | 72.22 | 78.03 | 95.14 | 0.718 | <0.05 | <0.05 |
| −12 h (55–70) | 75.2 | 8.21 | 51.67 | 71.13 | 76.19 | 80.68 | 93.56 | 0.065 | 0.14 | 0.167 |
| −6 h (40–54) | 68.46 | 10.3 | 32.1 | 63.14 | 68.98 | 74.23 | 91.32 | 0.223 | <0.05 | <0.05 |
| −6 h (55–70) | 72.95 | 7.87 | 50.36 | 69.31 | 73.4 | 78.34 | 89.74 | <0.05 | <0.05 | <0.05 |
| −3 h (40–54) | 63.74 | 11.1 | 35.46 | 58.02 | 63.74 | 71.34 | 90.47 | 0.07 | <0.05 | <0.05 |
| −3 h (55–70) | 69.95 | 7.82 | 39.71 | 65.98 | 69.83 | 75.32 | 86.1 | <0.05 | <0.05 | <0.05 |
| +3 h (40–54) | 60.03 | 10.2 | 26.99 | 53.26 | 61.35 | 66.99 | 86.51 | 0.134 | <0.05 | <0.05 |
| +3 h (55–70) | 69.92 | 7.48 | 46.97 | 65.6 | 70.15 | 75.19 | 83.59 | <0.05 | <0.05 | <0.05 |
| +6 h (40–54) | 65.56 | 9.33 | 44.99 | 59.53 | 66.49 | 71.55 | 86.7 | 0.169 | <0.05 | <0.05 |
| +6 h (55–70) | 71.7 | 7.3 | 53.65 | 68.11 | 72.1 | 76.47 | 86.51 | 0.16 | <0.05 | <0.05 |
| +12 h (40–54) | 70.53 | 9.01 | 44.94 | 65.46 | 71.21 | 76.8 | 87.6 | 0.158 | <0.05 | <0.05 |
| +12 h (55–70) | 74.54 | 7.86 | 50.98 | 69.19 | 75.73 | 79.63 | 92.5 | 0.162 | <0.05 | <0.05 |
| +24 h (40–54) | 73.55 | 8.69 | 50.21 | 67.61 | 73.83 | 79.34 | 93.12 | 0.852 | <0.05 | <0.05 |
| +24 h (55–70) | 76.19 | 7.94 | 55.53 | 71.8 | 76.73 | 81.79 | 95.05 | 0.227 | 0.517 | 0.386 |
| +48 h (40–54) | 75.27 | 8.35 | 54.08 | 69.75 | 75.7 | 80.88 | 94.35 | 0.873 | 0.16 | 0.225 |
| +48 h (55–70) | 76.91 | 8.05 | 54.29 | 72.16 | 77.57 | 82.56 | 95.64 | 0.432 | 0.969 | 0.972 |
Table A3.
Statistics of glucose analysis and variability metrics for time above range (TAR) or time outside range greater than 180 mg/dL (TOR > 180%) for various time periods surrounding the instances of hypoglycemia, respectively. The metrics include mean, standard deviation (SD), minimum (Min), first quartile (25%), median (50%), third quartile (75%), maximum (Max), Shapiro–Wilk test (SW-test), Z-test, and Kolmogorov–Smirnov test (KS-test).
Table A3.
Statistics of glucose analysis and variability metrics for time above range (TAR) or time outside range greater than 180 mg/dL (TOR > 180%) for various time periods surrounding the instances of hypoglycemia, respectively. The metrics include mean, standard deviation (SD), minimum (Min), first quartile (25%), median (50%), third quartile (75%), maximum (Max), Shapiro–Wilk test (SW-test), Z-test, and Kolmogorov–Smirnov test (KS-test).
| Mean (%) | SD (%) | Min (%) | 25% | 50% | 75% | Max (%) | SW-Test | Z-Test | KS-Test | |
|---|---|---|---|---|---|---|---|---|---|---|
| Overall | 18.92 | 9.78 | 0.8 | 12.98 | 17.99 | 25.52 | 49.67 | 0.22 | ||
| −48 h (41–50) | 18.63 | 10.15 | 0 | 11.24 | 18.36 | 23.77 | 53.05 | <0.05 | 0.836 | 0.971 |
| −48 h (51–60) | 17.97 | 9.63 | 1.47 | 11.31 | 17.02 | 24.35 | 49.65 | 0.062 | 0.482 | 0.825 |
| −48 h (61–70) | 17.97 | 9.63 | 1.47 | 11.31 | 17.02 | 24.35 | 49.65 | 0.062 | 0.482 | 0.825 |
| −24 h (41–50) | 18.84 | 10.64 | 0 | 10.79 | 18.18 | 24.27 | 50.64 | <0.05 | 0.953 | 0.825 |
| −24 h (51–60) | 17.42 | 9.28 | 1.09 | 10.92 | 15.99 | 22.92 | 47.62 | 0.061 | 0.261 | 0.596 |
| −24 h (61–70) | 17.42 | 9.28 | 1.09 | 10.92 | 15.99 | 22.92 | 47.62 | 0.061 | 0.261 | 0.596 |
| −12 h (41–50) | 18.15 | 10.59 | 0 | 9.98 | 17.6 | 23.89 | 55.28 | <0.05 | 0.588 | 0.596 |
| −12 h (51–60) | 16.49 | 8.91 | 1.39 | 10.56 | 15.19 | 21.68 | 42.45 | <0.05 | 0.063 | 0.118 |
| −12 h (61–70) | 15.61 | 9 | 1.12 | 9.74 | 14.1 | 20.52 | 45.76 | <0.05 | <0.05 | <0.05 |
| −6 h (41–50) | 16.65 | 11.43 | 0 | 8.37 | 16 | 22.37 | 69.25 | <0.05 | 0.128 | 0.058 |
| −6 h (51–60) | 14.64 | 8.84 | 1.14 | 8.52 | 14.6 | 18.69 | 46.01 | <0.05 | <0.05 | <0.05 |
| −6 h (61–70) | 13.36 | 8.17 | 1.08 | 7.9 | 12.53 | 17.51 | 40.8 | <0.05 | <0.05 | <0.05 |
| −3 h (41–50) | 11.83 | 9.96 | 0 | 5.4 | 9.94 | 15.22 | 59.26 | <0.05 | <0.05 | <0.05 |
| −3 h (51–60) | 10.09 | 7.53 | 0.61 | 5.06 | 9.22 | 12.94 | 42.95 | <0.05 | <0.05 | <0.05 |
| −3 h (61–70) | 10.09 | 7.53 | 0.61 | 5.06 | 9.22 | 12.94 | 42.95 | <0.05 | <0.05 | <0.05 |
| +3 h (41–50) | 9.02 | 7.66 | 0 | 3.38 | 7.24 | 12.66 | 43.06 | <0.05 | <0.05 | <0.05 |
| +3 h (51–60) | 7.69 | 6.04 | 0.16 | 3.72 | 6.28 | 10.2 | 37.73 | <0.05 | <0.05 | <0.05 |
| +3 h (61–70) | 6.5 | 4.51 | 0.24 | 3.23 | 5.31 | 9.11 | 19.77 | <0.05 | <0.05 | <0.05 |
| +6 h (41–50) | 16.04 | 10.15 | 0 | 8.4 | 14.74 | 22.91 | 52.48 | <0.05 | <0.05 | 0.083 |
| +6 h (51–60) | 14.32 | 8.72 | 1.42 | 8.41 | 12.82 | 19.25 | 49.83 | <0.05 | <0.05 | <0.05 |
| +6 h (61–70) | 12.9 | 7.78 | 0.85 | 7.16 | 11.19 | 17.92 | 33.2 | <0.05 | <0.05 | <0.05 |
| +12 h (41–50) | 17.92 | 9.77 | 0.04 | 10.94 | 16.36 | 24.31 | 45.22 | 0.2 | 0.465 | 0.596 |
| +12 h (51–60) | 16.83 | 8.94 | 0.77 | 11.05 | 15.54 | 22.25 | 43.52 | <0.05 | 0.111 | 0.163 |
| +12 h (61–70) | 15.67 | 8.55 | 0.57 | 10.31 | 13.59 | 21.36 | 41.69 | <0.05 | <0.05 | <0.05 |
| +24 h (41–50) | 17.74 | 9.26 | 0.14 | 11.54 | 17.15 | 23.79 | 42.44 | 0.295 | 0.374 | 0.714 |
| +24 h (51–60) | 17.17 | 8.96 | 0.89 | 11.7 | 15.86 | 22.61 | 43.24 | 0.083 | 0.183 | 0.38 |
| +24 h (61–70) | 16.37 | 8.7 | 0.42 | 11.09 | 14.74 | 21.28 | 41.94 | 0.116 | <0.05 | 0.058 |
| +48 h (41–50) | 17.98 | 8.87 | 0.07 | 11.67 | 18.47 | 23.61 | 40.96 | 0.477 | 0.469 | 0.825 |
| +48 h (51–60) | 17.53 | 8.92 | 0.79 | 11.54 | 17.03 | 22.6 | 44.52 | 0.192 | 0.289 | 0.38 |
| +48 h (61–70) | 16.93 | 8.81 | 0.48 | 11.72 | 16.43 | 22.3 | 43.39 | 0.236 | 0.126 | 0.293 |
| −48 h (40–54) | 18.44 | 10.4 | 0.13 | 10.88 | 17.68 | 24.51 | 51.61 | <0.05 | 0.707 | 0.829 |
| −48 h (55–70) | 17.37 | 9.65 | 0.03 | 10.61 | 15.88 | 23.38 | 48.98 | 0.071 | 0.237 | 0.489 |
| −24 h (40–54) | 18.19 | 10.17 | 0.26 | 9.77 | 17.34 | 23.65 | 49.54 | <0.05 | 0.572 | 0.719 |
| −24 h (55–70) | 16.71 | 9.35 | 0.01 | 9.91 | 15.15 | 22.81 | 46.63 | 0.056 | 0.09 | 0.225 |
| −12 h (40–54) | 17.34 | 10 | 0 | 9.83 | 17.21 | 23.21 | 49.33 | <0.05 | 0.236 | 0.225 |
| −12 h (55–70) | 15.62 | 9.04 | 0.01 | 9.22 | 14.12 | 20.59 | 44.42 | <0.05 | <0.05 | <0.05 |
| −6 h (40–54) | 15.65 | 10.43 | 0 | 8.11 | 14.84 | 20.49 | 62.52 | <0.05 | <0.05 | <0.05 |
| −6 h (55–70) | 13.61 | 8.41 | 0 | 7.99 | 12.93 | 17.96 | 40.11 | <0.05 | <0.05 | <0.05 |
| −3 h (40–54) | 10.72 | 8.95 | 0 | 4.83 | 9.21 | 12.6 | 53 | <0.05 | <0.05 | <0.05 |
| −3 h (55–70) | 9 | 6.33 | 0.01 | 4.61 | 8.26 | 10.96 | 31.55 | <0.05 | <0.05 | <0.05 |
| +3 h (40–54) | 8.97 | 7.64 | 0 | 3.3 | 6.98 | 13.29 | 36.67 | <0.05 | <0.05 | <0.05 |
| +3 h (55–70) | 6.78 | 4.92 | 0.01 | 3.19 | 5.56 | 9.49 | 20.37 | <0.05 | <0.05 | <0.05 |
| +6 h (40–54) | 15.7 | 9.59 | 0.59 | 8.1 | 14.28 | 22.64 | 48.06 | <0.05 | <0.05 | <0.05 |
| +6 h (55–70) | 13.32 | 8.04 | 0 | 7.46 | 11.77 | 18.95 | 34.81 | <0.05 | <0.05 | <0.05 |
| +12 h (40–54) | 17.27 | 9.31 | 0.66 | 10.79 | 16.08 | 22.81 | 44.89 | 0.148 | 0.2 | 0.489 |
| +12 h (55–70) | 15.77 | 8.74 | 0.01 | 9.99 | 13.82 | 21.58 | 42.18 | 0.051 | <0.05 | <0.05 |
| +24 h (40–54) | 17.52 | 9.43 | 0.26 | 10.78 | 16.77 | 24.32 | 43.78 | 0.161 | 0.28 | 0.602 |
| +24 h (55–70) | 16.47 | 8.89 | 0.03 | 11.25 | 14.81 | 21.67 | 41.72 | 0.171 | 0.054 | 0.167 |
| +48 h (40–54) | 17.81 | 9.31 | 0.17 | 11.35 | 17.85 | 24.48 | 42.65 | 0.369 | 0.382 | 0.602 |
| +48 h (55–70) | 17 | 9.04 | 0.04 | 10.75 | 15.59 | 22.61 | 43.82 | 0.324 | 0.132 | 0.489 |
Table A4.
Statistics of glucose analysis and variability metrics for standard deviation (SD) in mg/dL for various time periods surrounding the instances of hypoglycemia, respectively. The metrics include mean, standard deviation (SD), minimum (Min), first quartile (25%), median (50%), third quartile (75%), maximum (Max), Shapiro–Wilk test (SW-test), Z-test, and Kolmogorov–Smirnov test (KS-test). The SD is smallest at the time of the hypoglycemia, as this is when glucose levels reach their lowest. Following the hypoglycemia, the SD begins to increase again, indicating that glucose levels become more variable. For example, the mean SD for the 41–50 mg/dL category increases from 34.69 mg/dL at +3 h to 48.03 mg/dL at +48 h. More severe hypoglycemia (lower glucose levels) is associated with smaller SD values both before and after the event. This suggests that more severe hypoglycemia is associated with less variable glucose levels. For SD, the data do not follow a normal distribution and have a statistically significant difference (generally p < 0.05) in distribution when compared with normal (or overall) distribution.
Table A4.
Statistics of glucose analysis and variability metrics for standard deviation (SD) in mg/dL for various time periods surrounding the instances of hypoglycemia, respectively. The metrics include mean, standard deviation (SD), minimum (Min), first quartile (25%), median (50%), third quartile (75%), maximum (Max), Shapiro–Wilk test (SW-test), Z-test, and Kolmogorov–Smirnov test (KS-test). The SD is smallest at the time of the hypoglycemia, as this is when glucose levels reach their lowest. Following the hypoglycemia, the SD begins to increase again, indicating that glucose levels become more variable. For example, the mean SD for the 41–50 mg/dL category increases from 34.69 mg/dL at +3 h to 48.03 mg/dL at +48 h. More severe hypoglycemia (lower glucose levels) is associated with smaller SD values both before and after the event. This suggests that more severe hypoglycemia is associated with less variable glucose levels. For SD, the data do not follow a normal distribution and have a statistically significant difference (generally p < 0.05) in distribution when compared with normal (or overall) distribution.
| Mean (%) | SD (%) | Min (%) | 25% | 50% | 75% | Max (%) | SW-Test | Z-Test | KS-Test | |
|---|---|---|---|---|---|---|---|---|---|---|
| Overall | 50.16 | 10.8 | 21.47 | 43.22 | 50.23 | 58.47 | 77.33 | 0.831 | ||
| −48 h (41–50) | 48.04 | 10.39 | 26.31 | 40.74 | 47.36 | 54.26 | 76.52 | 0.53 | 0.152 | 0.293 |
| −48 h (51–60) | 47.03 | 10.29 | 23.23 | 40.32 | 46.67 | 54.06 | 74.72 | 0.91 | <0.05 | 0.083 |
| −48 h (61–70) | 47.03 | 10.29 | 23.23 | 40.32 | 46.67 | 54.06 | 74.72 | 0.91 | <0.05 | 0.083 |
| −24 h (41–50) | 47.07 | 10.27 | 26.79 | 40.28 | 46.08 | 53.88 | 76.52 | 0.113 | <0.05 | <0.05 |
| −24 h (51–60) | 45.3 | 9.77 | 23.04 | 39.25 | 44.41 | 52.21 | 71.14 | 0.903 | <0.05 | <0.05 |
| −24 h (61–70) | 45.3 | 9.77 | 23.04 | 39.25 | 44.41 | 52.21 | 71.14 | 0.903 | <0.05 | <0.05 |
| −12 h (41–50) | 44.5 | 10.53 | 26.13 | 37.82 | 43.92 | 50.5 | 84.25 | <0.05 | <0.05 | <0.05 |
| −12 h (51–60) | 42.18 | 9.49 | 22.13 | 35.63 | 41.85 | 47.81 | 70.43 | 0.855 | <0.05 | <0.05 |
| −12 h (61–70) | 40.32 | 9.57 | 20.57 | 33.98 | 40.49 | 47.14 | 69.7 | 0.61 | <0.05 | <0.05 |
| −6 h (41–50) | 39.53 | 9.86 | 12.36 | 33.02 | 39.17 | 44.45 | 80.8 | <0.05 | <0.05 | <0.05 |
| −6 h (51–60) | 37.29 | 8.45 | 20.98 | 32.05 | 37.45 | 42.26 | 58.62 | 0.299 | <0.05 | <0.05 |
| −6 h (61–70) | 34.89 | 8.64 | 17.95 | 29.52 | 34.31 | 40.9 | 57.46 | 0.414 | <0.05 | <0.05 |
| −3 h (41–50) | 34.43 | 9.48 | 7.72 | 28.26 | 33.88 | 39.06 | 73.79 | <0.05 | <0.05 | <0.05 |
| −3 h (51–60) | 31.39 | 7.82 | 14.33 | 25.57 | 30.45 | 36.02 | 55.23 | <0.05 | <0.05 | <0.05 |
| −3 h (61–70) | 31.39 | 7.82 | 14.33 | 25.57 | 30.45 | 36.02 | 55.23 | <0.05 | <0.05 | <0.05 |
| +3 h (41–50) | 34.69 | 9.41 | 11.73 | 29.56 | 33.2 | 41.08 | 65.49 | 0.19 | <0.05 | <0.05 |
| +3 h (51–60) | 30.91 | 8.14 | 11.72 | 26.26 | 30.07 | 36.08 | 57.16 | 0.275 | <0.05 | <0.05 |
| +3 h (61–70) | 26.95 | 7.15 | 11.61 | 22.2 | 26.99 | 32.2 | 45.09 | 0.773 | <0.05 | <0.05 |
| +6 h (41–50) | 43.23 | 10.47 | 21.69 | 35.84 | 43.14 | 49.39 | 64.35 | 0.168 | <0.05 | <0.05 |
| +6 h (51–60) | 39.66 | 9.35 | 19.78 | 34.05 | 39.21 | 44.97 | 68.05 | 0.331 | <0.05 | <0.05 |
| +6 h (61–70) | 36.07 | 8.95 | 17.97 | 30.2 | 35.23 | 42.02 | 58.45 | 0.617 | <0.05 | <0.05 |
| +12 h (41–50) | 46.3 | 10.57 | 18.14 | 39.27 | 46.25 | 54.5 | 69.86 | 0.734 | <0.05 | <0.05 |
| +12 h (51–60) | 43.95 | 9.54 | 20.15 | 37.89 | 43.98 | 50.05 | 71.75 | 0.992 | <0.05 | <0.05 |
| +12 h (61–70) | 41.32 | 9.38 | 17.83 | 35.67 | 41.41 | 46.89 | 67.7 | 0.948 | <0.05 | <0.05 |
| +24 h (41–50) | 47.4 | 10.51 | 14.69 | 40.72 | 46.83 | 54.8 | 69.63 | 0.387 | 0.064 | 0.293 |
| +24 h (51–60) | 45.83 | 9.63 | 21.31 | 38.97 | 45.55 | 52.53 | 70.2 | 0.991 | <0.05 | <0.05 |
| +24 h (61–70) | 44.07 | 9.55 | 19.08 | 38.34 | 43.7 | 50.58 | 67.51 | 0.929 | <0.05 | <0.05 |
| +48 h (41–50) | 48.03 | 9.96 | 19.88 | 41.93 | 47.63 | 56.16 | 69.39 | 0.372 | 0.142 | 0.596 |
| +48 h (51–60) | 47.04 | 9.7 | 20.5 | 40.72 | 46.97 | 53.55 | 69.03 | 0.841 | <0.05 | 0.083 |
| +48 h (61–70) | 45.83 | 9.71 | 19.89 | 40.49 | 45.44 | 52.69 | 68.33 | 0.709 | <0.05 | <0.05 |
| −48 h (40–54) | 47.43 | 10.99 | 16.16 | 40.63 | 46.81 | 55.05 | 76.87 | 0.978 | 0.064 | 0.167 |
| −48 h (55–70) | 45.73 | 10.76 | 12.81 | 38.74 | 44.78 | 52.43 | 75.34 | 0.924 | <0.05 | <0.05 |
| −24 h (40–54) | 46.1 | 10.62 | 17.35 | 39.4 | 46.1 | 52.37 | 71.85 | 0.582 | <0.05 | <0.05 |
| −24 h (55–70) | 43.77 | 10.35 | 12.57 | 37.24 | 42.44 | 50.52 | 68.49 | 0.879 | <0.05 | <0.05 |
| −12 h (40–54) | 43.26 | 10.5 | 21.34 | 35.88 | 42.71 | 49.85 | 77.83 | 0.296 | <0.05 | <0.05 |
| −12 h (55–70) | 40.2 | 10.02 | 11.73 | 34.06 | 38.89 | 46.4 | 65.97 | 0.718 | <0.05 | <0.05 |
| −6 h (40–54) | 38.64 | 9.45 | 19.16 | 31.9 | 38.55 | 45.09 | 67.15 | 0.233 | <0.05 | <0.05 |
| −6 h (55–70) | 35.07 | 8.96 | 10.83 | 29.74 | 34.66 | 40.54 | 56.56 | 0.475 | <0.05 | <0.05 |
| −3 h (40–54) | 33.33 | 8.71 | 16.34 | 27.71 | 32.33 | 37.18 | 69.22 | <0.05 | <0.05 | <0.05 |
| −3 h (55–70) | 28.78 | 7.8 | 8.31 | 23.5 | 28.65 | 32.07 | 50.74 | 0.518 | <0.05 | <0.05 |
| +3 h (40–54) | 34.25 | 9.59 | 11.74 | 28.67 | 32.42 | 40.32 | 60.83 | 0.384 | <0.05 | <0.05 |
| +3 h (55–70) | 27.8 | 7.93 | 6.86 | 22.57 | 27.3 | 33.57 | 46.93 | 0.831 | <0.05 | <0.05 |
| +6 h (40–54) | 42.39 | 10.37 | 20.95 | 35.31 | 41.09 | 48.61 | 65.66 | 0.202 | <0.05 | <0.05 |
| +6 h (55–70) | 37 | 9.5 | 10.16 | 31.09 | 36.06 | 43.4 | 61.53 | 0.964 | <0.05 | <0.05 |
| +12 h (40–54) | 45.39 | 10.26 | 22.13 | 38.14 | 45.18 | 53.19 | 70.76 | 0.684 | <0.05 | <0.05 |
| +12 h (55–70) | 41.55 | 9.97 | 12.18 | 35.29 | 41.63 | 47.13 | 68.9 | 0.95 | <0.05 | <0.05 |
| +24 h (40–54) | 46.82 | 10.57 | 17.24 | 39.53 | 46.53 | 54.39 | 70.29 | 0.882 | <0.05 | 0.086 |
| +24 h (55–70) | 44.11 | 10.24 | 12.85 | 37.73 | 43.35 | 50.82 | 68.02 | 0.844 | <0.05 | <0.05 |
| +48 h (40–54) | 47.51 | 10.47 | 17.19 | 40.46 | 47.12 | 55.32 | 69.61 | 0.489 | 0.065 | 0.167 |
| +48 h (55–70) | 45.81 | 10.38 | 13.42 | 40.07 | 45.69 | 52.96 | 67.79 | 0.605 | <0.05 | <0.05 |
Table A5.
Statistics of glucose analysis and variability metrics for J_Index in mg/dL for various time periods surrounding the instances of hypoglycemia, respectively. The metrics include mean, standard deviation (SD), minimum (Min), first quartile (25%), median (50%), third quartile (75%), maximum (Max), Shapiro–Wilk test (SW-test), Z-test, and Kolmogorov–Smirnov test (KS-test). Overall, for the J_index, the mean score is 36.48 mg/dL, with a standard deviation (SD) of 10.62 mg/dL. The minimum score is 15.23 mg/dL, and the maximum score is 73.93 mg/dL. There is a non-normal (p < 0.05) distribution of scores for most time periods and levels of hypoglycemia and most show a significant (p < 0.005) difference from the overall distribution.
Table A5.
Statistics of glucose analysis and variability metrics for J_Index in mg/dL for various time periods surrounding the instances of hypoglycemia, respectively. The metrics include mean, standard deviation (SD), minimum (Min), first quartile (25%), median (50%), third quartile (75%), maximum (Max), Shapiro–Wilk test (SW-test), Z-test, and Kolmogorov–Smirnov test (KS-test). Overall, for the J_index, the mean score is 36.48 mg/dL, with a standard deviation (SD) of 10.62 mg/dL. The minimum score is 15.23 mg/dL, and the maximum score is 73.93 mg/dL. There is a non-normal (p < 0.05) distribution of scores for most time periods and levels of hypoglycemia and most show a significant (p < 0.005) difference from the overall distribution.
| Mean (%) | SD (%) | Min (%) | 25% | 50% | 75% | Max (%) | SW-Test | Z-Test | KS-Test | |
|---|---|---|---|---|---|---|---|---|---|---|
| Overall | 36.48 | 10.62 | 15.23 | 29.87 | 35.49 | 43.84 | 73.93 | 0.185 | ||
| −48 h (41–50) | 35.76 | 11 | 15.88 | 27.95 | 35.6 | 41.15 | 74.26 | <0.05 | 0.631 | 0.596 |
| −48 h (51–60) | 34.89 | 10.45 | 15.67 | 27.83 | 33.87 | 41.23 | 73.45 | <0.05 | 0.28 | 0.293 |
| −48 h (61–70) | 34.89 | 10.45 | 15.67 | 27.83 | 33.87 | 41.23 | 73.45 | <0.05 | 0.28 | 0.293 |
| −24 h (41–50) | 35.43 | 11.16 | 15.8 | 27.19 | 34.45 | 40.08 | 69.88 | <0.05 | 0.49 | 0.293 |
| −24 h (51–60) | 33.83 | 9.98 | 15.1 | 27.27 | 32.39 | 39.91 | 70.4 | <0.05 | 0.066 | 0.118 |
| −24 h (61–70) | 33.83 | 9.98 | 15.1 | 27.27 | 32.39 | 39.91 | 70.4 | <0.05 | 0.066 | 0.118 |
| −12 h (41–50) | 34.2 | 12.01 | 14.66 | 25.71 | 33.07 | 39.5 | 87.75 | <0.05 | 0.151 | 0.163 |
| −12 h (51–60) | 32.16 | 9.77 | 15 | 25.88 | 30.55 | 38.01 | 61.53 | <0.05 | <0.05 | <0.05 |
| −12 h (61–70) | 31.32 | 10.12 | 14.33 | 25.39 | 29.45 | 36.05 | 69.37 | <0.05 | <0.05 | <0.05 |
| −6 h (41–50) | 30.96 | 12.47 | 6.81 | 22.99 | 30.18 | 34.8 | 98.55 | <0.05 | <0.05 | <0.05 |
| −6 h (51–60) | 28.86 | 9.42 | 13.21 | 22.65 | 28.15 | 33.72 | 59.69 | <0.05 | <0.05 | <0.05 |
| −6 h (61–70) | 27.69 | 9.13 | 12.53 | 21.93 | 26.92 | 31.2 | 57.25 | <0.05 | <0.05 | <0.05 |
| −3 h (41–50) | 24.92 | 10.19 | 5.27 | 18.74 | 23.29 | 28.83 | 72.29 | <0.05 | <0.05 | <0.05 |
| −3 h (51–60) | 22.94 | 8.34 | 9.2 | 17.23 | 21.84 | 26.14 | 55.78 | <0.05 | <0.05 | <0.05 |
| −3 h (61–70) | 22.94 | 8.34 | 9.2 | 17.23 | 21.84 | 26.14 | 55.78 | <0.05 | <0.05 | <0.05 |
| +3 h (41–50) | 21.74 | 8.07 | 5.95 | 16.68 | 20.11 | 27.2 | 53.37 | <0.05 | <0.05 | <0.05 |
| +3 h (51–60) | 20.27 | 6.77 | 7.22 | 16.21 | 19.4 | 24.61 | 48.49 | <0.05 | <0.05 | <0.05 |
| +3 h (61–70) | 19.08 | 5.57 | 8.7 | 15.36 | 18.54 | 23.1 | 35.01 | 0.288 | <0.05 | <0.05 |
| +6 h (41–50) | 30.82 | 10.64 | 12.08 | 23.61 | 29.59 | 37.14 | 70.51 | <0.05 | <0.05 | <0.05 |
| +6 h (51–60) | 28.72 | 9.26 | 12.8 | 22.73 | 27.48 | 33.47 | 64.19 | <0.05 | <0.05 | <0.05 |
| +6 h (61–70) | 27.02 | 8.32 | 12.03 | 21.41 | 25.51 | 32.72 | 50.88 | 0.07 | <0.05 | <0.05 |
| +12 h (41–50) | 33.91 | 10.36 | 13.13 | 26.86 | 33.03 | 40.56 | 70.34 | 0.097 | 0.079 | <0.05 |
| +12 h (51–60) | 32.55 | 9.52 | 14.37 | 26.52 | 31.29 | 38.36 | 62.32 | <0.05 | <0.05 | <0.05 |
| +12 h (61–70) | 31.1 | 9.01 | 13.77 | 25.68 | 29.42 | 37 | 60.21 | 0.089 | <0.05 | <0.05 |
| +24 h (41–50) | 34.37 | 9.95 | 13.9 | 27.47 | 33.72 | 40.98 | 64.7 | 0.354 | 0.143 | 0.221 |
| +24 h (51–60) | 33.61 | 9.43 | 15.1 | 27.28 | 33.21 | 38.71 | 59.59 | 0.107 | <0.05 | 0.083 |
| +24 h (61–70) | 32.67 | 9.14 | 14.53 | 26.83 | 31.34 | 38.3 | 58.26 | 0.267 | <0.05 | <0.05 |
| +48 h (41–50) | 34.92 | 9.43 | 14.34 | 28.26 | 34.67 | 41.28 | 59.33 | 0.887 | 0.265 | 0.482 |
| +48 h (51–60) | 34.38 | 9.41 | 14.74 | 27.91 | 34.46 | 39.83 | 60.89 | 0.568 | 0.135 | 0.221 |
| +48 h (61–70) | 33.71 | 9.31 | 14.87 | 27.79 | 32.87 | 39.78 | 61.72 | 0.473 | <0.05 | 0.083 |
| −48 h (40–54) | 35.42 | 11.26 | 11.12 | 27.59 | 34.66 | 41.81 | 73.37 | <0.05 | 0.484 | 0.719 |
| −48 h (55–70) | 34.17 | 10.53 | 9.71 | 27.34 | 32.8 | 40.72 | 73.18 | 0.121 | 0.116 | 0.298 |
| −24 h (40–54) | 34.64 | 10.83 | 11.88 | 26.42 | 33.75 | 40.61 | 67.7 | <0.05 | 0.215 | 0.298 |
| −24 h (55–70) | 33.06 | 10.13 | 9.6 | 26.35 | 31.49 | 38.53 | 66.99 | 0.212 | <0.05 | <0.05 |
| −12 h (40–54) | 33.25 | 11.01 | 13.82 | 24.77 | 32.6 | 40.01 | 75.92 | <0.05 | <0.05 | 0.086 |
| −12 h (55–70) | 31.22 | 10.13 | 9.22 | 25.19 | 29.32 | 35.51 | 62.78 | <0.05 | <0.05 | <0.05 |
| −6 h (40–54) | 30.01 | 10.77 | 11.96 | 22.53 | 29.31 | 34.73 | 78.49 | <0.05 | <0.05 | <0.05 |
| −6 h (55–70) | 27.85 | 9.35 | 8.61 | 22.09 | 27.16 | 31.69 | 58.47 | <0.05 | <0.05 | <0.05 |
| −3 h (40–54) | 23.95 | 9.55 | 9.22 | 18.31 | 22.54 | 26.64 | 71.4 | <0.05 | <0.05 | <0.05 |
| −3 h (55–70) | 21.81 | 7.33 | 7.22 | 16.68 | 21.83 | 24.59 | 48.66 | <0.05 | <0.05 | <0.05 |
| +3 h (40–54) | 21.73 | 8.35 | 6.52 | 16.48 | 19.84 | 27.71 | 50.17 | <0.05 | <0.05 | <0.05 |
| +3 h (55–70) | 19.47 | 6.13 | 6.62 | 15.64 | 18.82 | 23.45 | 34.66 | 0.157 | <0.05 | <0.05 |
| +6 h (40–54) | 30.3 | 10.27 | 12.33 | 22.87 | 28.81 | 36.88 | 62.8 | <0.05 | <0.05 | <0.05 |
| +6 h (55–70) | 27.56 | 8.76 | 8.17 | 21.58 | 26.61 | 33.64 | 50.99 | 0.257 | <0.05 | <0.05 |
| +12 h (40–54) | 33.17 | 10 | 13 | 26.9 | 31.96 | 40.43 | 61.35 | 0.201 | <0.05 | <0.05 |
| +12 h (55–70) | 31.24 | 9.35 | 9.21 | 24.6 | 29.76 | 37.15 | 60.81 | 0.351 | <0.05 | <0.05 |
| +24 h (40–54) | 34.08 | 10.12 | 10.87 | 27.66 | 33.45 | 40.57 | 60.79 | 0.373 | 0.095 | 0.167 |
| +24 h (55–70) | 32.72 | 9.52 | 9.65 | 26.97 | 31.57 | 38.57 | 58.34 | 0.62 | <0.05 | <0.05 |
| +48 h (40–54) | 34.7 | 9.94 | 11.21 | 28.1 | 33.81 | 41.28 | 59.99 | 0.955 | 0.212 | 0.602 |
| +48 h (55–70) | 33.75 | 9.72 | 9.93 | 27.36 | 32.75 | 40.06 | 61.47 | 0.825 | 0.053 | 0.167 |
Figure A2.
Comparison of distributions (box plots) for (A) LBGI and (B) HBGI before and after hypoglycemia at levels of 41–50 mg/dL, 51–60 mg/dL, and 61–70 mg/dL.
Figure A2.
Comparison of distributions (box plots) for (A) LBGI and (B) HBGI before and after hypoglycemia at levels of 41–50 mg/dL, 51–60 mg/dL, and 61–70 mg/dL.
Figure A3.
Comparison of distributions (box plots) for percentage of (A) time below range (TBR or TOR < 70), (B) time in range (TIR), and (C) time above range (TAR), before and after hypoglycemia at levels of 40–54 mg/dL and 55–70 mg/dL.
Figure A3.
Comparison of distributions (box plots) for percentage of (A) time below range (TBR or TOR < 70), (B) time in range (TIR), and (C) time above range (TAR), before and after hypoglycemia at levels of 40–54 mg/dL and 55–70 mg/dL.
Figure A4.
Comparison of distributions (box plots) for (A) standard deviation (SD) and (B) J_Index, before and after hypoglycemia at levels of 40–54 mg/dL and 55–70 mg/dL.
Figure A4.
Comparison of distributions (box plots) for (A) standard deviation (SD) and (B) J_Index, before and after hypoglycemia at levels of 40–54 mg/dL and 55–70 mg/dL.
Figure A5.
Comparison of distributions (box plots) for (A) LBGI and (B) HBGI before and after hypoglycemia at levels of 40–54 mg/dL and 55–70 mg/dL.
Figure A5.
Comparison of distributions (box plots) for (A) LBGI and (B) HBGI before and after hypoglycemia at levels of 40–54 mg/dL and 55–70 mg/dL.
Figure A6.
Comparison of distributions (box plots) for percentage of (A) time below range (TBR or TOR < 70), (B) time in range (TIR), (C) time above range (TAR), and (D) standard deviation, before and after hypoglycemia at level of 40–54 mg/dL. This analysis is extracted utilizing T1DEXI dataset containing large-scale diabetes data from adults (n = 222) using commercial AID systems.
Figure A6.
Comparison of distributions (box plots) for percentage of (A) time below range (TBR or TOR < 70), (B) time in range (TIR), (C) time above range (TAR), and (D) standard deviation, before and after hypoglycemia at level of 40–54 mg/dL. This analysis is extracted utilizing T1DEXI dataset containing large-scale diabetes data from adults (n = 222) using commercial AID systems.
Table A6.
Statistics of glucose analysis and variability metrics for TIR, TAR, TBR, and standard deviation for various time periods surrounding the instances of hypoglycemia, respectively. The metrics include mean, standard deviation (SD), minimum (Min), first quartile (25%), median (50%), third quartile (75%), maximum (Max), Shapiro–Wilk test (SW-test), Z-test, and Kolmogorov–Smirnov test (KS-test). These statistics are extracted utilizing T1DEXI dataset containing large-scale diabetes data from adults (n = 222) using a commercial AID system.
Table A6.
Statistics of glucose analysis and variability metrics for TIR, TAR, TBR, and standard deviation for various time periods surrounding the instances of hypoglycemia, respectively. The metrics include mean, standard deviation (SD), minimum (Min), first quartile (25%), median (50%), third quartile (75%), maximum (Max), Shapiro–Wilk test (SW-test), Z-test, and Kolmogorov–Smirnov test (KS-test). These statistics are extracted utilizing T1DEXI dataset containing large-scale diabetes data from adults (n = 222) using a commercial AID system.
| Mean (%) | SD (%) | Min (%) | 25% | 50% | 75% | Max (%) | SW-Test | Z-Test | KS-Test | |
|---|---|---|---|---|---|---|---|---|---|---|
| Time in Range (TIR) | ||||||||||
| Overall | 79.06 | 10.77 | 38.94 | 72.51 | 81.31 | 86.29 | 98.26 | <0.05 | 0.447 | <0.05 |
| −48 h (40–54) | 78.44 | 11.11 | 24.88 | 72.74 | 80.17 | 85.89 | 99.22 | <0.05 | 0.574 | <0.05 |
| −24 h (40–54) | 77.13 | 11.7 | 31.99 | 71.28 | 79.31 | 85.39 | 98.78 | <0.05 | 0.09 | <0.05 |
| −12 h (40–54) | 74.6 | 12.78 | 34.72 | 68.65 | 76.77 | 83.79 | 99.65 | <0.05 | 0.08 | <0.05 |
| −6 h (40–54) | 70.16 | 13.83 | 29.44 | 65.01 | 70.93 | 79.37 | 99.31 | <0.05 | <0.05 | <0.05 |
| −3 h (40–54) | 67.85 | 14.39 | 18.87 | 59.47 | 67.45 | 77.72 | 98.61 | <0.05 | <0.05 | <0.05 |
| +3 h (40–54) | 66.97 | 14.94 | 11.23 | 49.57 | 67.15 | 77.22 | 92.22 | <0.05 | <0.05 | <0.05 |
| +6 h (40–54) | 69.95 | 15.96 | 8.33 | 49.63 | 71.94 | 84.9 | 96.11 | <0.05 | <0.05 | <0.05 |
| +12 h (40–54) | 74.92 | 13.71 | 4.17 | 63.65 | 76.36 | 83.57 | 97.92 | <0.05 | 0.05 | <0.05 |
| +24 h (40–54) | 76.93 | 11.86 | 10.11 | 77.9 | 80.11 | 85 | 98.96 | <0.05 | 0.064 | <0.05 |
| +48 h (40–54) | 77.98 | 11.78 | 27.26 | 72.64 | 80.62 | 85.44 | 99.03 | <0.05 | 0.343 | 0.609 |
| Time After Range (TAR) | ||||||||||
| Overall | 18.47 | 10.98 | 1.48 | 10.72 | 16.05 | 23.88 | 59.9 | <0.05 | 0.778 | <0.05 |
| −48 h (40–54) | 17.71 | 11.18 | 0 | 10.22 | 15.66 | 22.49 | 71.67 | <0.05 | 0.498 | 0.447 |
| −24 h (40–54) | 17.83 | 11.65 | 0 | 9.25 | 15.58 | 23.08 | 51.11 | <0.05 | 0.581 | 0.525 |
| −12 h (40–54) | 18.5 | 12.38 | 0 | 9.27 | 16.16 | 24.37 | 55.5 | <0.05 | 0.978 | 0.311 |
| −6 h (40–54) | 19.15 | 13.62 | 0 | 8.69 | 16.92 | 27.37 | 64.44 | <0.05 | 0.59 | 0.165 |
| −3 h (40–54) | 13.95 | 13.12 | 0 | 4.43 | 11.6 | 20.47 | 73.61 | <0.05 | <0.05 | <0.05 |
| +3 h (40–54) | 13.03 | 13.11 | 0 | 3.88 | 9.57 | 18.7 | 80.85 | <0.05 | <0.05 | <0.05 |
| +6 h (40–54) | 18.36 | 15.5 | 0 | 5.25 | 13.35 | 23.55 | 87.5 | <0.05 | 0.937 | 0.08 |
| +12 h (40–54) | 17.71 | 13.03 | 0 | 9.15 | 15.23 | 23.66 | 65.97 | <0.05 | 0.08 | 0.054 |
| +24 h (40–54) | 17.78 | 11.71 | 0 | 9.85 | 15.72 | 23.11 | 66.98 | <0.05 | 0.552 | 0.853 |
| +48 h (40–54) | 18.08 | 11.87 | 0.35 | 9.81 | 15.76 | 23.44 | 72.14 | <0.05 | 0.737 | 0.915 |
| Time Below Range (TBR) | ||||||||||
| Overall | 2.47 | 1.9 | 0.18 | 1.16 | 2 | 3.21 | 10.15 | <0.05 | <0.05 | <0.05 |
| −48 h (40–54) | 3.86 | 2.96 | 0.09 | 1.87 | 3.36 | 5 | 20.89 | <0.05 | <0.05 | <0.05 |
| −24 h (40–54) | 5.03 | 3.67 | 0.17 | 2.73 | 4.28 | 6.27 | 23.23 | <0.05 | <0.05 | <0.05 |
| −12 h (40–54) | 6.89 | 4.65 | 0.35 | 4.03 | 5.88 | 8.71 | 30.21 | <0.05 | <0.05 | <0.05 |
| −6 h (40–54) | 10.7 | 6.88 | 0.69 | 8.94 | 9.42 | 12.94 | 44.66 | <0.05 | <0.05 | <0.05 |
| −3 h (40–54) | 18.2 | 10.58 | 5.34 | 13.69 | 19.92 | 25.94 | 77.62 | <0.05 | <0.05 | <0.05 |
| +3 h (40–54) | 20 | 11.17 | 5.56 | 13.77 | 17.87 | 23.99 | 87.5 | <0.05 | <0.05 | <0.05 |
| +6 h (40–54) | 11.69 | 8.62 | 6.79 | 6.79 | 9.98 | 14.94 | 68.74 | <0.05 | <0.05 | <0.05 |
| +12 h (40–54) | 7.37 | 6.51 | 1.48 | 6.89 | 6.29 | 8.72 | 36.08 | <0.05 | <0.05 | <0.05 |
| +24 h (40–54) | 5.29 | 4.69 | 0.69 | 2.76 | 4.99 | 6.31 | 36.08 | <0.05 | <0.05 | <0.05 |
| +48 h (40–54) | 3.94 | 3.12 | 0.1 | 2.1 | 3.07 | 5.06 | 18.5 | <0.05 | <0.05 | <0.05 |
| Standard Deviation (mg/dL) | ||||||||||
| Overall | 45.86 | 10.66 | 23.28 | 38.35 | 44.59 | 51.53 | 81.46 | <0.05 | 0.254 | <0.05 |
| −48 h (40–54) | 44.61 | 11.36 | 21.36 | 36.06 | 42.91 | 50.26 | 98.42 | <0.05 | 0.267 | <0.05 |
| −24 h (40–54) | 44.42 | 12.06 | 18.65 | 34.65 | 42.06 | 50.69 | 94.54 | <0.05 | 0.215 | 0.08 |
| −12 h (40–54) | 44.13 | 12.25 | 19.08 | 35.31 | 41.82 | 50.86 | 96.36 | <0.05 | 0.138 | <0.05 |
| −6 h (40–54) | 43.39 | 10.56 | 19.06 | 36.4 | 41.56 | 48.69 | 88.69 | <0.05 | <0.05 | <0.05 |
| −3 h (40–54) | 38.68 | 13.96 | 7.93 | 34.05 | 37.19 | 44.51 | 86.72 | <0.05 | <0.05 | <0.05 |
| +3 h (40–54) | 43 | 14.09 | 15.06 | 34.59 | 41.54 | 45.49 | 119.2 | <0.05 | <0.05 | <0.05 |
| +6 h (40–54) | 43.12 | 13.15 | 9.93 | 35.42 | 40.49 | 49.54 | 115.05 | <0.05 | <0.05 | <0.05 |
| +12 h (40–54) | 44.95 | 10.56 | 18.89 | 36.5 | 43.23 | 50.3 | 102.62 | <0.05 | 0.443 | 0.254 |
| +24 h (40–54) | 44.95 | 9.56 | 18.89 | 36.5 | 43.23 | 50.3 | 102.62 | <0.05 | 0.443 | 0.254 |
| +48 h (40–54) | 45.13 | 11.62 | 19.55 | 43.88 | 43.88 | 51.16 | 91.42 | <0.05 | 0.521 | 0.853 |
References
- Gregory, G.A.; Robinson, T.I.; Linklater, S.E.; Wang, F.; Colagiuri, S.; de Beaufort, C.; Donaghue, K.C.; Harding, J.L.; Wander, P.L.; Zhang, X.; et al. Global incidence, prevalence, and mortality of type 1 diabetes in 2021 with projection to 2040: A modelling study. Lancet Diabetes Endocrinol. 2022, 10, 741–760. [Google Scholar] [CrossRef] [PubMed]
- Knoll, C.; Schipp, J.; O’Donnell, S.; Wäldchen, M.; Ballhausen, H.; Cleal, B.; Gajewska, K.A.; Raile, K.; Skinner, T.; Braune, K. Quality of life and psychological well-being among children and adolescents with diabetes and their caregivers using open-source automated insulin delivery systems: Findings from a multinational survey. Diabetes Res. Clin. Pract. 2023, 196, 110153. [Google Scholar] [CrossRef] [PubMed]
- Phillip, M.; Nimri, R.; Bergenstal, R.M.; Barnard-Kelly, K.; Danne, T.; Hovorka, R.; Kovatchev, B.P.; Messer, L.H.; Parkin, C.G.; Ambler-Osborn, L.; et al. Consensus Recommendations for the Use of Automated Insulin Delivery Technologies in Clinical Practice. Endocr. Rev. 2023, 44, 254–280. [Google Scholar] [CrossRef] [PubMed]
- Moon, S.J.; Jung, I.; Park, C.Y. Current Advances of Artificial Pancreas Systems: A Comprehensive Review of the Clinical Evidence. Diabetes Metab. J. 2021, 45, 813–839. [Google Scholar] [CrossRef]
- Kovatchev, B.P.; Otto, E.; Cox, D.; Gonder-Frederick, L.; Clarke, W. Evaluation of a new measure of blood glucose variability in diabetes. Diabetes Care 2006, 29, 2433–2438. [Google Scholar] [CrossRef]
- McDonnell, C.M.; Donath, S.M.; Vidmar, S.I.; Werther, G.A.; Cameron, F.J. A novel approach to continuous glucose analysis utilizing glycemic variation. Diabetes Technol. Ther. 2005, 7, 253–263. [Google Scholar] [CrossRef]
- Service, F.J.; Molnar, G.D.; Rosevear, J.W.; Ackerman, E.; Gatewood, L.C.; Taylor, W.F. Mean amplitude of glycemic excursions, a measure of diabetic instability. Diabetes 1970, 19, 644–655. [Google Scholar] [CrossRef]
- Bergenstal, R.M.; Beck, R.W.; Close, K.L.; Grunberger, G.; Sacks, D.B.; Kowalski, A.; Brown, A.S.; Heinemann, L.; Aleppo, G.; Ryan, D.B.; et al. Glucose management indicator (GMI): A new term for estimating A1C from continuous glucose monitoring. Diabetes Care 2018, 41, 2275–2280. [Google Scholar] [CrossRef]
- Bent, B.; Henriquez, M.; Dunn, J. Cgmquantify: Python and R Software Packages for Comprehensive Analysis of Interstitial Glucose and Glycemic Variability from Continuous Glucose Monitor Data. IEEE Open J. Eng. Med. Biol. 2021, 2, 263–266. [Google Scholar] [CrossRef]
- Rawlings, R.A.; Shi, H.; Yuan, L.H.; Brehm, W.; Pop-Busui, R.; Nelson, P.W. Translating glucose variability metrics into the clinic via continuous glucose monitoring: A graphical user interface for diabetes evaluation (CGM-GUIDE ©). Diabetes Technol. Ther. 2011, 13, 1241–1248. [Google Scholar] [CrossRef]
- Attaye, I.; van der Vossen, E.W.; Mendes Bastos, D.N.; Nieuwdorp, M.; Levin, E. Introducing the Continuous Glucose Data Analysis (CGDA) R Package: An Intuitive Package to Analyze Continuous Glucose Monitoring Data. J. Diabetes Sci. Technol. 2022, 16, 783–785. [Google Scholar] [CrossRef] [PubMed]
- Moscardó, V.; Giménez, M.; Oliver, N.; Hill, N.R. Updated Software for Automated Assessment of Glucose Variability and Quality of Glycemic Control in Diabetes. Diabetes Technol. Ther. 2020, 22, 701–708. [Google Scholar] [CrossRef] [PubMed]
- Vigers, T.; Chan, C.L.; Snell-Bergeon, J.; Bjornstad, P.; Zeitler, P.S.; Forlenza, G.; Pyle, L. Cgmanalysis: An R package for descriptive analysis of continuous glucose monitor data. PLoS ONE 2019, 14, e0216851. [Google Scholar] [CrossRef] [PubMed]
- Czerwoniuk, D.; Fendler, W.; Walenciak, L.; Mlynarski, W. GlyCulator: A glycemic variability calculation tool for continuous glucose monitoring data. J. Diabetes Sci. Technol. 2011, 5, 447–451. [Google Scholar] [CrossRef] [PubMed]
- ElSayed, N.A.; Aleppo, G.; Aroda, V.R.; Bannuru, R.R.; Brown, F.M.; Bruemmer, D.; Collins, B.S.; Hilliard, M.E.; Isaacs, D.; Johnson, E.L.; et al. 6. Glycemic Targets: Standards of Care in Diabetes—2023. Diabetes Care 2023, 46, S97–S110. [Google Scholar] [CrossRef]
- Kovatchev, B.; Cobelli, C. Glucose variability: Timing, risk analysis, and relationship to hypoglycemia in diabetes. Diabetes Care 2016, 39, 502–510. [Google Scholar] [CrossRef]
- DeVries, J.H. Glucose variability: Where it is important and how to measure it. Diabetes 2013, 62, 1405. [Google Scholar] [CrossRef]
- OpenAPS Data Commons. Available online: https://openaps.org/outcomes/data-commons/ (accessed on 22 January 2023).
- Ball, M.; Lewis, D. The Open APS Data Commons (OpenAPS + Open Humans). In Sage Bionetworks Global Assembly; Sage Bionetworks: Seattle, WA, USA, 2017. [Google Scholar]
- Greshake Tzovaras, B.; Angrist, M.; Arvai, K.; Dulaney, M.; Estrada-Galiñanes, V.; Gunderson, B.; Head, T.; Lewis, D.; Nov, O.; Shaer, O.; et al. Open Humans: A platform for participant-centered research and personal data exploration. GigaScience 2019, 8, giz076. [Google Scholar] [CrossRef]
- Zafar, A.; Lewis, D.M.; Shahid, A. Long-Term Glucose Forecasting for Open-Source Automated Insulin Delivery Systems: A Machine Learning Study with Real-World Variability Analysis. Healthcare 2023, 11, 779. [Google Scholar] [CrossRef]
- Shahid, A.; Lewis, D.M. Large-Scale Data Analysis for Glucose Variability Outcomes with Open-Source Automated Insulin Delivery Systems. Nutrients 2022, 14, 1906. [Google Scholar] [CrossRef]
- Shahid, A. Programming Scripts for Glucose Variability Analysis Related to Hypoglycemia (HypoGV) 2023. Available online: https://github.com/danamlewis/OpenHumansDataTools/tree/master/bin/GV-Hypoglycemia (accessed on 22 February 2024).
- Wojcicki, J.M. J-index. A new proposition of the assessment of current glucose control in diabetic patients. Horm. Metab. Res. 1995, 27, 41–42. [Google Scholar] [CrossRef] [PubMed]
- Kovatchev, B. Glycemic Variability: Risk Factors, Assessment, and Control. J. Diabetes Sci. Technol. 2019, 13, 627–635. [Google Scholar] [CrossRef] [PubMed]
- Monnier, L.; Wojtusciszyn, A.; Molinari, N.; Colette, C.; Renard, E.; Owens, D. Respective Contributions of Glycemic Variability and Mean Daily Glucose as Predictors of Hypoglycemia in Type 1 Diabetes: Are They Equivalent? Diabetes Care 2020, 43, 821–827. [Google Scholar] [CrossRef] [PubMed]
- Cho, S.H.; Kim, S.; Lee, Y.B.; Jin, S.M.; Hur, K.Y.; Kim, G.; Kim, J.H. Impact of continuous glucose monitoring on glycemic control and its derived metrics in type 1 diabetes: A longitudinal study. Front. Endocrinol. 2023, 14, 1165471. [Google Scholar] [CrossRef]
- Olafsdottir, A.F.; Polonsky, W.; Bolinder, J.; Hirsch, I.B.; Dahlqvist, S.; Wedel, H.; Nyström, T.; Wijkman, M.; Schwarcz, E.; Hellman, J.; et al. A Randomized Clinical Trial of the Effect of Continuous Glucose Monitoring on Nocturnal Hypoglycemia, Daytime Hypoglycemia, Glycemic Variability, and Hypoglycemia Confidence in Persons with Type 1 Diabetes Treated with Multiple Daily Insulin Injections (GOLD-3). Diabetes Technol. Ther. 2018, 20, 274–284. [Google Scholar] [CrossRef]
Figure 2.
Comparison of distributions (box plots) for percentage of (A) time below range (TBR or TOR < 70), (B) time in range (TIR), and (C) time above range (TAR > 180 mg/dL), before and after hypoglycemia at levels of 41–50 mg/dL, 51–60 mg/dL, and 61–70 mg/dL.
Figure 2.
Comparison of distributions (box plots) for percentage of (A) time below range (TBR or TOR < 70), (B) time in range (TIR), and (C) time above range (TAR > 180 mg/dL), before and after hypoglycemia at levels of 41–50 mg/dL, 51–60 mg/dL, and 61–70 mg/dL.
Figure 3.
Comparison of distributions (box plots) for (A) standard deviation (SD) and (B) J_Index, before and after hypoglycemia at levels of 41–50 mg/dL, 51–60 mg/dL, and 61–70 mg/dL.
Figure 3.
Comparison of distributions (box plots) for (A) standard deviation (SD) and (B) J_Index, before and after hypoglycemia at levels of 41–50 mg/dL, 51–60 mg/dL, and 61–70 mg/dL.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
MDPI and ACS Style
Shahid, A.; Lewis, D.M. Impact of Hypoglycemia on Glucose Variability over Time for Individuals with Open-Source Automated Insulin Delivery Systems. Diabetology 2024, 5, 514-536. https://doi.org/10.3390/diabetology5050038
AMA Style
Shahid A, Lewis DM. Impact of Hypoglycemia on Glucose Variability over Time for Individuals with Open-Source Automated Insulin Delivery Systems. Diabetology. 2024; 5(5):514-536. https://doi.org/10.3390/diabetology5050038
Chicago/Turabian StyleShahid, Arsalan, and Dana M. Lewis. 2024. "Impact of Hypoglycemia on Glucose Variability over Time for Individuals with Open-Source Automated Insulin Delivery Systems" Diabetology 5, no. 5: 514-536. https://doi.org/10.3390/diabetology5050038
APA StyleShahid, A., & Lewis, D. M. (2024). Impact of Hypoglycemia on Glucose Variability over Time for Individuals with Open-Source Automated Insulin Delivery Systems. Diabetology, 5(5), 514-536. https://doi.org/10.3390/diabetology5050038
