Novel Indicators for Adverse Glycemic Events Detection Analysis Based on Continuous Glucose Monitoring Neural Network Predictive Models

doi:10.1007/s12204-022-2439-0

Abstract

Abstract: This paper proposes five indicators to evaluate the effectiveness and viability for adverse glycemic events detection based on predicted blood glucose (BG) values. False negative rate (FNR) and false positive rate (FPR) are defined to evaluate whether it can detect adverse glycemic events (AGEs) based on the predicted value. The temporal overlap (TO) and time difference (TD) are proposed to evaluate whether the predicted model can capture the accurate time duration of AGEs. The sum of squared percent (SSP) measures comprehensive similarity between prediction values and true values. We examined 328 patients with type 2 diabetes, containing real continuous glucose monitoring data with 5-minute time intervals. Autoregressive integrated moving average model has lower FNR and FPR. The gated recurrent unit has better temporal behavior where the mean TO with standard deviation is calculated as 0.84±0.18, and the mean TD with standard deviation is (4.39±4.01) min. Neural models have better effects on SSP (for hypoglycemia, long-short tern memory possesses 0.149 and 0.246). These five indicators are able to evaluate whether we can detect abnormal BG levels and reveal the temporal behavior of AGEs effectively. The proposed neural predictive models have more promising application in AGE detection.

Key words: adverse glycemic events detection, glucose prediction, neural network, evaluation, indicators

CLC Number:

TP399

LU Guannan1 (卢冠男), WANG Mengling1∗ (王梦灵), FOX Tamara2, JIANG Peng3 (蒋鹏), JIANG Fusong3 (蒋伏松). Novel Indicators for Adverse Glycemic Events Detection Analysis Based on Continuous Glucose Monitoring Neural Network Predictive Models[J]. J Shanghai Jiaotong Univ Sci, 2022, 27(4): 498-504.

References 32

[1]	HAAK T, HANAIRE H, AJJAN R, et al. Flash glucose-sensing technology as a replacement for blood glucose monitoring for the management of insulin-treated type 2 diabetes: A multicenter, open-label randomized controlled trial [J]. Diabetes Therapy, 2017,8(1): 55-73.
[2]	CHARLEER S, DE BLOCK C, V AN HUFFEL L, etal. Quality of life and glucose control after 1 year of nationwide reimbursement of intermittently scanned continuous glucose monitoring in adults living with type 1 diabetes (FUTURE): A prospective observational real-world cohort study [J]. Diabetes Care, 2020,43(2): 389-397.
[3]	MONNIER L, COLETTE C, OWENS D R. The ap-plication of simple metrics in the assessment of glycaemic variability [J]. Diabetes & Metabolism, 2018,44(4): 313-319.
[4]	LISZKA-HACKZELL J J. Prediction of blood glucoselevels in diabetic patients using a hybrid AI technique[J]. Computers and Biomedical Research, 1999, 32(2):132-144.
[5]	CHUAH Z M, PARAMESRAN R, THAMBIRAT-NAM K, et al. A two-level partial least squares system for non-invasive blood glucose concentration prediction[J]. Chemometrics and Intelligent Laboratory Systems,2010, 104(2): 347-351.
[6]	GYUK P, V ASS áNYI I, K óSA I. Blood glucose level prediction for diabetics based on nutrition and insulinad ministration logs using personalized mathematical models [J]. Journal of Healthcare Engineering, 2019,2019: 8605206.
[7]	CONTRERAS I, OVIEDO S, VETTORETTI M, etal. Personalized blood glucose prediction: A hybrid ap-proach using grammatical evolution and physiological models [J]. PLoS ONE, 2017, 12(11): e0187754.
[8]	ZHAO H, ZHAO C H, YU C X, et al. Multiple order model migration and optimal model selection for on-line glucose prediction in Type 1 diabetes [J]. AIChEJournal, 2018, 64(3): 822-834.
[9]	ZHAO C H, YU C X. Rapid model identification for online subcutaneous glucose concentration prediction for new subjects with type I diabetes [J]. IEEETransactions on Bio-Medical Engineering, 2015, 62(5):1333-1344.
[10]	HAYES C, KRISKA A. Role of physical activity indiabetes management and prevention [J]. Journal of the American Dietetic Association, 2008, 108(4): S19-S23.
[11]	RODBARD D. Continuous glucose monitoring: Areview of recent studies demonstrating improved glycemic outcomes [J]. Diabetes Technology & Therapeutics, 2017, 19(S3): S25-S37.
[12]	F ACCHINETTI A. Continuous glucose monitoringsensors: Past, present and future algorithmic challenges [J]. Sensors, 2016, 16(12): 2093.
[13]	ZECCHIN C, F ACCHINETTI A, SPARACINO G,et al. Reduction of number and duration of hypo-glycemic events by glucose prediction methods: Aproof-of-concept in silico study [J]. Diabetes Technol-ogy & Therapeutics, 2013, 15(1): 66-77.
[14]	ZHONG A, CHOUDHARY P, MCMAHON C, et al.Effectiveness of automated insulin management fea-tures of the MiniMed ? 640G sensor-augmented in-sulin pump [J]. Diabetes Technology & Therapeutics,2016, 18(10): 657-663.
[15]	BERGENSTAL R M, GARG S, WEINZIMER S A,et al. Safety of a hybrid closed-loop insulin delivery system in patients with type 1 diabetes [J]. JAMA,2016, 316(13): 1407.
[16]	HAMDI T, BEN ALI J, DI COSTANZO V, et al. Ac-curate prediction of continuous blood glucose based on support vector regression and differential evolution algorithm [J]. Biocybernetics and Biomedical Engineering, 2018, 38(2): 362-372.
[17]	SISODIA D, SISODIA D S. Prediction of diabetesusing classification algorithms [J]. Procedia Computer Science, 2018, 132: 1578-1585.
[18]	JUNG M, LEE Y B, JIN S M, et al. Prediction of daytime hypoglycemic events using continuous glucose monitoring data and classification technique [EB/OL].(2017-04-27). https://arxiv.org/abs/1704.08769.
[19]	YEH H C, BROWN T T, MARUTHUR N, et al. Comparative effectiveness and safety of methods of insulindelivery and glucose monitoring for diabetes mellitus:A systematic review and meta-analysis [J]. Annals of Internal Medicine, 2012, 157(5): 336-347.
[20]	POLONSKY W H, HESSLER D, RUEDY K J, et al.The impact of continuous glucose monitoring on markers of quality of life in adults with type 1 diabetes: Fur-ther findings from the DIAMOND randomized clinicaltrial [J]. Diabetes Care, 2017, 40(6): 736-741.
[21]	LIU C Y, VEHí J, A V ARI P, et al. Long-term glucose forecasting using a physiological model and deconvolution of the continuous glucose monitoring signal [J].Sensors, 2019, 19(19): 4338.
[22]	GADALETA M, F ACCHINETTI A, GRISAN E, etal. Prediction of adverse glycemic events from continuous glucose monitoring signal [J]. IEEE Journal of Biomedical and Health Informatics, 2019, 23(2): 650-659.
[23]	SAKURAI K, KA W AI Y, YAMAZAKI M, et al. Pre-diction of lowest nocturnal blood glucose level based on self-monitoring of blood glucose in Japanese patients with type 2 diabetes [J]. Journal of Diabetes and its Complications, 2018, 32(12): 1118-1123.
[24]	GOTO A, ARAH O A, GOTO M, et al. Severe hypo-glycaemia and cardiovascular disease: Systematic re-view and meta-analysis with bias analysis [J]. BMJ,2013, 347: f4533.
[25]	INVESTIGATORS N S S, FINFER S, LIU B, et al.Hypoglycemia and risk of death in critically ill patients [J]. The New England Journal of Medicine, 2012,367(12): 1108-1118.
[26]	LU J Y, MA X J, ZHOU J, et al. Association of timein range, as assessed by continuous glucose monitoring, with diabetic retinopathy in type 2 diabetes [J].Diabetes Care, 2018, 41(11): 2370-2376.
[27]	NGUYEN M, HAN J L, SPANAKIS E K, et al. A re-view of continuous glucose monitoring-based composite metrics for glycemic control [J]. Diabetes Technology & Therapeutics, 2020, 22(8): 613-622.
[28]	YANG J, LI L, SHI Y M, et al. An ARIMA model with adaptive orders for predicting blood glucose concentrations and hypoglycemia [J]. IEEE Journal of Biomedical and Health Informatics, 2019, 23(3): 1251-1260.
[29]	HOCHREITER S, SCHMIDHUBER J. Long short-term memory [J]. Neural Computation, 1997, 9(8):1735-1780.
[30]	SIAMI-NAMINI S, TA V AKOLI N, SIAMI NAMIN A.A comparison of ARIMA and LSTM in forecasting time series [C]//2018 17th IEEE International Conference on Machine Learning and Applications. Orlando,FL, USA: IEEE, 2018: 1394-1401.
[31]	CHO K, V AN MERRIENBOER B, GUL-CEHRE C, et al. Learning phrase representations using RNN encoder-decoder for statistical machine translation [EB/OL]. (2014-06-03).https://arxiv.org/abs/1406.1078.
[32]	S H I H S Y , S U N F K , L E E H Y . T e m p o r a l p a t t e r nattention for multivariate time series forecasting [J].Machine Learning, 2019, 108(8/9): 1421-1441.