Dataset Information

Automated meal detection from continuous glucose monitor data through simulation and explanation.

ABSTRACT: BACKGROUND:Artificial pancreas systems aim to reduce the burden of type 1 diabetes by automating insulin dosing. These systems link a continuous glucose monitor (CGM) and insulin pump with a control algorithm, but require users to announce meals, without which the system can only react to the rise in blood glucose. OBJECTIVE:We investigate whether CGM data can be used to automatically infer meals in daily life even in the presence of physical activity, which can raise or lower blood glucose. MATERIALS AND METHODS:We propose a novel meal detection algorithm that combines simulations with CGM, insulin pump, and heart rate monitor data. When observed and predicted glucose differ, our algorithm uses simulations to test whether a meal may explain this difference. We evaluated our method on simulated data and real-world data from individuals with type 1 diabetes. RESULTS:In simulated data, we detected meals earlier and with higher accuracy than was found in prior work (25.7 minutes, 1.2 g error; compared with 48.3 minutes, 17.2 g error). In real-world data, we discovered a larger number of plausible meals than was found in prior work (30 meals, 76.7% accepted; compared with 33 meals, 39.4% accepted). DISCUSSION:Prior research attempted meal detection from CGM, but had delays and lower accuracy in real data or did not allow for physical activity. Our approach can be used to improve insulin dosing in an artificial pancreas and trigger reminders for missed meal boluses. CONCLUSIONS:We demonstrate that meal information can be robustly inferred from CGM and body-worn sensor data, even in challenging environments of daily life.

SUBMITTER: Zheng M

PROVIDER: S-EPMC6857509 | biostudies-literature | 2019 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Automated meal detection from continuous glucose monitor data through simulation and explanation.

Zheng Min M Ni Baohua B Kleinberg Samantha S

Journal of the American Medical Informatics Association : JAMIA 20191201 12

<h4>Background</h4>Artificial pancreas systems aim to reduce the burden of type 1 diabetes by automating insulin dosing. These systems link a continuous glucose monitor (CGM) and insulin pump with a control algorithm, but require users to announce meals, without which the system can only react to the rise in blood glucose.<h4>Objective</h4>We investigate whether CGM data can be used to automatically infer meals in daily life even in the presence of physical activity, which can raise or lower blo ...[more]

PMID: 31562509

Similar Datasets

Project description:BackgroundIn the Diabetes domain, events such as meals and exercises play an important role in the disease management. For that, many studies focus on automatic meal detection, specially as part of the so-called artificial [Formula: see text]-cell systems. Meals are associated to blood glucose (BG) variations, however such variations are not peculiar to meals, it mostly comes as a combination of external factors. Thus, general approaches such as the ones focused on glucose signal rate of change are not enough to detect personalized influence of such factors. By using a data-driven individualized approach for meal detection, our method is able to fit real data, detecting personalized meal responses even when such external factors are implicitly present.MethodsThe method is split into model training and selection. In the training phase, we start observing meal responses for each individual, and identifying personalized patterns. Occurrences of such patterns are searched over the BG signal, evaluating the similarity of each pattern to each possible signal subsequence. The most similar occurrences are then selected as possible meal event candidates. For that, we include steps for excluding less relevant neighbors per pattern, and grouping close occurrences in time globally. Each candidate is represented by a set of time and response signal related qualitative variables. These variables are used as input features for different binary classifiers in order to learn to classify a candidate as MEAL or NON-MEAL. In the model selection phase, we compare all trained classifiers to select the one that performs better with the data of each individual.ResultsThe results show that the method is able to detect daily meals, providing a result with a balanced proportion between detected meals and false alarms. The analysis on multiple patients indicate that the approach achieves good outcomes when there is enough reliable training data, as this is reflected on the testing results.ConclusionsThe approach aims at personalizing the meal detection task by relying solely on data. The premise is that a model trained with data that contains the implicit influence of external factors is able to recognize the nuances of the individual that generated the data. Besides, the approach can also be used to improve data quality by detecting meals, opening opportunities to possible applications such as detecting and reminding users of missing or wrongly informed meal events.

Project description:BackgroundHigh postprandial glucose excursions may increase risk for disease. Individuals have widely varying glucose responses to different meals, and precision nutrition approaches often seek to personalize diets to minimize postprandial glycemic responses as measured by continuous glucose monitors (CGMs). However, it is unknown whether different CGM devices result in concordant meal rankings according to postprandial glycemic excursions.ObjectiveWe explored whether meal rankings according to postprandial glycemic excursions differ between 2 simultaneously worn CGMs.MethodsWe collected 27,489 simultaneous measurements from Dexcom G4 Platinum and Abbott Freestyle Libre Pro CGMs during 28 inpatient days in 16 adults without diabetes. Simultaneous glucose measurements obtained for 2 h following 760 ad libitum meals were used to compare within-subject meal rankings between the CGM devices according to their incremental glucose response.ResultsPostprandial responses to ad libitum meals were highly variable, with the Abbott and Dexcom systems resulting in within-subject incremental mean ± SD glucose CVs of 91.7 ± 1.9% and 94.2 ± 2.7%, respectively. Within-subject meal rankings for incremental glycemic responses were relatively discordant between CGMs, with a mean Kendall rank correlation coefficient of 0.43 ± 0.05. Meals in the bottom compared with those in the top half of incremental glycemic responses ranked by Abbott resulted in 50 ± 10% (P = 0.0002) less glycemic reduction as measured by Dexcom, and vice versa. The missing glycemic reduction by eating meals ranked according to the discordant CGM was inversely correlated with each subject's Kendall rank correlation coefficient (r = -0.95; P < 0.0001).ConclusionsPrecision nutrition approaches that use CGMs to personalize meal recommendations for minimizing glycemic excursions may be premature given the discordance of within-subject meal rankings between simultaneous CGM devices. More research is needed to clarify the source of this imprecision. This trial was registered at clinicaltrials.gov as NCT03407053.

Project description:BackgroundGestational Diabetes Mellitus (GDM) incidence and adverse outcomes have increased globally. The validity of the oral glucose tolerance test (OGTT) for GDM diagnosis has long been questioned, with no suitable substitute reported yet. Continuous Glucose Monitoring (CGM) is potentially a more acceptable and comprehensive test. The aim of this study was to assess the Freestyle Libre Pro 2 acceptability as a diagnostic test for GDM, then triangulating its results with OGTT results as well as risk factors and sonographic features of GDM.MethodsWomen wore the CGM device for 7 days at 24-28 weeks, undergoing the OGTT before CGM removal. CGM/OGTT acceptability as well as GDM risk factors evaluation occurred via three online surveys. CGM distribution/variability/time in range parameters, combined in a CGM Score of Variability (CGMSV), were triangulated with OGTT results and a risk-factor-based Total Risk Score (TRS). In a subgroup, GDM ultrasound features (as modified Ultrasound Gestational Diabetes Score - m-UGDS) were also incorporated.ResultsOf 107 women recruited, 87 (81%) were included: 74 (85%) with negative OGTT (NGT) and 13 (15%) positive (GDM). No significant difference was found between NGT and GDM in terms of demographics (apart from family history of diabetes mellitus), CGM parameters and perinatal outcomes. Women considered CGM significantly more acceptable than OGTT (81% versus 27% rating 5/5, p < 0.001). Of the 55 NGT with triangulation data, 28 were considered 'true negative' (TRS concordant with OGTT and CGMSV): of these 4/5 evaluated at ultrasound had m-UGDS below the cut-off. Five women were considered 'false negative' (negative OGTT with both TRS and CGMSV above the respective cut-offs). Triangulation identified also six 'false positive' women (positive OGTT but TRS and CGM both below the cut-offs). Only one woman for each of the last two categories had m-UGDS evaluated, with discordant results.ConclusionsCGM represents a more acceptable alternative for GDM diagnosis to the OGTT. CGM triangulation analysis suggests OGTT screening may result in both false positives and negatives. Further research including larger cohorts of patients, and additional triangulation elements (such as GDM biomarkers/outcomes and expanded m-UGDS) is needed to explore CGM potential for GDM diagnosis.

Dataset Information

Automated meal detection from continuous glucose monitor data through simulation and explanation.

Publications

Automated meal detection from continuous glucose monitor data through simulation and explanation.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets