Improved Low-Glucose Predictive Alerts Based on Sustained Hypoglycemia: Model Development and Validation Study.
continuous glucose monitoring
diabetes
false alert rate
glucose monitoring
hypoglycemia
machine learning
model generalizability
quantile regression forests
sustained hypoglycemia
Journal
JMIR diabetes
ISSN: 2371-4379
Titre abrégé: JMIR Diabetes
Pays: Canada
ID NLM: 101719410
Informations de publication
Date de publication:
29 Apr 2021
29 Apr 2021
Historique:
received:
03
01
2021
accepted:
17
03
2021
revised:
09
03
2021
entrez:
29
4
2021
pubmed:
30
4
2021
medline:
30
4
2021
Statut:
epublish
Résumé
Predictive alerts for impending hypoglycemic events enable persons with type 1 diabetes to take preventive actions and avoid serious consequences. This study aimed to develop a prediction model for hypoglycemic events with a low false alert rate, high sensitivity and specificity, and good generalizability to new patients and time periods. Performance improvement by focusing on sustained hypoglycemic events, defined as glucose values less than 70 mg/dL for at least 15 minutes, was explored. Two different modeling approaches were considered: (1) a classification-based method to directly predict sustained hypoglycemic events, and (2) a regression-based prediction of glucose at multiple time points in the prediction horizon and subsequent inference of sustained hypoglycemia. To address the generalizability and robustness of the model, two different validation mechanisms were considered: (1) patient-based validation (model performance was evaluated on new patients), and (2) time-based validation (model performance was evaluated on new time periods). This study utilized data from 110 patients over 30-90 days comprising 1.6 million continuous glucose monitoring values under normal living conditions. The model accurately predicted sustained events with >97% sensitivity and specificity for both 30- and 60-minute prediction horizons. The false alert rate was kept to <25%. The results were consistent across patient- and time-based validation strategies. Providing alerts focused on sustained events instead of all hypoglycemic events reduces the false alert rate and improves sensitivity and specificity. It also results in models that have better generalizability to new patients and time periods.
Sections du résumé
BACKGROUND
BACKGROUND
Predictive alerts for impending hypoglycemic events enable persons with type 1 diabetes to take preventive actions and avoid serious consequences.
OBJECTIVE
OBJECTIVE
This study aimed to develop a prediction model for hypoglycemic events with a low false alert rate, high sensitivity and specificity, and good generalizability to new patients and time periods.
METHODS
METHODS
Performance improvement by focusing on sustained hypoglycemic events, defined as glucose values less than 70 mg/dL for at least 15 minutes, was explored. Two different modeling approaches were considered: (1) a classification-based method to directly predict sustained hypoglycemic events, and (2) a regression-based prediction of glucose at multiple time points in the prediction horizon and subsequent inference of sustained hypoglycemia. To address the generalizability and robustness of the model, two different validation mechanisms were considered: (1) patient-based validation (model performance was evaluated on new patients), and (2) time-based validation (model performance was evaluated on new time periods).
RESULTS
RESULTS
This study utilized data from 110 patients over 30-90 days comprising 1.6 million continuous glucose monitoring values under normal living conditions. The model accurately predicted sustained events with >97% sensitivity and specificity for both 30- and 60-minute prediction horizons. The false alert rate was kept to <25%. The results were consistent across patient- and time-based validation strategies.
CONCLUSIONS
CONCLUSIONS
Providing alerts focused on sustained events instead of all hypoglycemic events reduces the false alert rate and improves sensitivity and specificity. It also results in models that have better generalizability to new patients and time periods.
Identifiants
pubmed: 33913816
pii: v6i2e26909
doi: 10.2196/26909
pmc: PMC8120423
doi:
Types de publication
Journal Article
Langues
eng
Pagination
e26909Subventions
Organisme : FDA HHS
ID : P50 FD006428
Pays : United States
Informations de copyright
©Darpit Dave, Madhav Erraguntla, Mark Lawley, Daniel DeSalvo, Balakrishna Haridas, Siripoom McKay, Chester Koh. Originally published in JMIR Diabetes (https://diabetes.jmir.org), 29.04.2021.
Références
Diabetes Technol Ther. 2017 Jun;19(S3):S25-S37
pubmed: 28585879
IEEE J Biomed Health Inform. 2019 May;23(3):1251-1260
pubmed: 29993728
J Diabetes Sci Technol. 2008 Jul;2(4):612-21
pubmed: 19885237
Diabetes Technol Ther. 2013 Jul;15(7):538-43
pubmed: 23631608
IEEE Trans Neural Syst Rehabil Eng. 2020 Apr;28(4):961-969
pubmed: 32054581
Brief Bioinform. 2020 May 21;21(3):791-802
pubmed: 31220208
Diabetes Technol Ther. 2013 Oct;15(10):792-801
pubmed: 23883406
Diabetes Care. 2018 Oct;41(10):2155-2161
pubmed: 30089663
Crit Care Nurs Q. 2005 Oct-Dec;28(4):317-23
pubmed: 16239820
J Diabetes Sci Technol. 2020 Jun 1;:1932296820922622
pubmed: 32476492
Diabetes Technol Ther. 2020 Nov;22(11):787-793
pubmed: 32267773
J Diabetes Sci Technol. 2014 Jan 1;8(1):117-122
pubmed: 24876547
IEEE Trans Biomed Eng. 2012 Jun;59(6):1550-60
pubmed: 22374344
Diabetes Technol Ther. 2018 Jun;20(S2):S250-S253
pubmed: 29873525
Diabetes Technol Ther. 2020 Dec;22(12):883-891
pubmed: 32324062
IEEE Trans Cybern. 2016 Jul;46(7):1704-14
pubmed: 27244754
IEEE J Biomed Health Inform. 2019 Mar;23(2):650-659
pubmed: 29993992
J Diabetes Sci Technol. 2016 Aug 22;10(5):1149-60
pubmed: 27381030
J Diabetes Sci Technol. 2019 Sep;13(5):919-927
pubmed: 30650997
Sensors (Basel). 2020 Jul 13;20(14):
pubmed: 32668724
Int J Numer Method Biomed Eng. 2017 Jun;33(6):
pubmed: 27644067
Diabetes Technol Ther. 2009 Apr;11(4):243-53
pubmed: 19344199
J Thorac Dis. 2019 Mar;11(Suppl 4):S574-S584
pubmed: 31032076
J Diabetes Sci Technol. 2017 Jan;11(1):138-147
pubmed: 27530720
IEEE J Biomed Health Inform. 2020 Feb;24(2):414-423
pubmed: 31369390
J Diabetes Sci Technol. 2018 Jan;12(1):190-198
pubmed: 28741369
Physiol Meas. 2004 Aug;25(4):905-20
pubmed: 15382830
J Diabetes Sci Technol. 2010 Jan 01;4(1):25-33
pubmed: 20167164
J Diabetes Sci Technol. 2010 Mar 01;4(2):404-18
pubmed: 20307402
Diabetes Technol Ther. 2019 Feb;21(S1):S123-S137
pubmed: 30785328
J Diabetes Sci Technol. 2015 Apr 30;9(5):1126-37
pubmed: 25931581
Health Informatics J. 2019 Dec;25(4):1170-1187
pubmed: 29278956
PLoS One. 2019 Dec 11;14(12):e0219636
pubmed: 31826018
IEEE J Biomed Health Inform. 2013 Jan;17(1):71-81
pubmed: 23008265
J Diabetes Sci Technol. 2013 May 01;7(3):789-94
pubmed: 23759412
J Diabetes Sci Technol. 2014 Jul;8(4):731-7
pubmed: 24876412
Diabetes Technol Ther. 2013 Aug;15(8):634-43
pubmed: 23848178
Pediatr Diabetes. 2017 Aug;18(5):332-339
pubmed: 27125223
Ind Eng Chem Res. 2013 Sep 4;52(35):
pubmed: 24187436
IEEE Trans Inf Technol Biomed. 2010 Jan;14(1):157-65
pubmed: 19858035
PLoS One. 2017 Nov 7;12(11):e0187754
pubmed: 29112978
PLoS One. 2019 May 21;14(5):e0217199
pubmed: 31112566
J Diabetes Sci Technol. 2010 Sep 01;4(5):1146-55
pubmed: 20920434
Diabetes Care. 2017 Dec;40(12):1631-1640
pubmed: 29162583
J Diabetes Sci Technol. 2007 Sep;1(5):624-9
pubmed: 19885130
Sensors (Basel). 2021 Jan 12;21(2):
pubmed: 33445438
Diabetes Technol Ther. 2010 Jan;12(1):81-8
pubmed: 20082589
J Diabetes Sci Technol. 2012 Sep 01;6(5):1142-7
pubmed: 23063041
Diabetes. 1999 Mar;48(3):445-51
pubmed: 10078542
PLoS One. 2018 Mar 7;13(3):e0194025
pubmed: 29513751
J Med Internet Res. 2019 May 01;21(5):e11030
pubmed: 31042157
IEEE Trans Biomed Eng. 2007 May;54(5):931-7
pubmed: 17518291
Sensors (Basel). 2020 Mar 19;20(6):
pubmed: 32204318
J Diabetes Sci Technol. 2018 Mar;12(2):282-294
pubmed: 29493359
Comput Methods Programs Biomed. 2014;113(1):144-52
pubmed: 24192453
J Med Internet Res. 2018 May 30;20(5):e10775
pubmed: 29848472
Diabetes Care. 2019 Aug;42(8):1593-1603
pubmed: 31177185
IEEE J Biomed Health Inform. 2019 Apr 17;:
pubmed: 30998484
Diabetes Technol Ther. 2005 Feb;7(1):3-14
pubmed: 15738700
Diabetes Care. 2015 Mar;38(3):355-64
pubmed: 25205142
J Diabetes Sci Technol. 2018 Mar;12(2):265-272
pubmed: 29493356