Hypoglycemia prediction using machine learning models for patients with type 2 diabetes

Bharath Sudharsan, Malinda Peeples, Mansur Shomali

Research output: Contribution to journalArticle

Abstract

Background: Minimizing the occurrence of hypoglycemia in patients with type 2 diabetes is a challenging task since these patients typically check only 1 to 2 self-monitored blood glucose (SMBG) readings per day. Method: We trained a probabilistic model using machine learning algorithms and SMBG values from real patients. Hypoglycemia was defined as a SMBG value < 70 mg/dL. We validated our model using multiple data sets. In addition, we trained a second model, which used patient SMBG values and information about patient medication administration. Results: The optimal number of SMBG values needed by the model was approximately 10 per week. The sensitivity of the model for predicting a hypoglycemia event in the next 24 hours was 92% and the specificity was 70%. In the model that incorporated medication information, the prediction window was for the hour of hypoglycemia, and the specificity improved to 90%. Conclusions: Our machine learning models can predict hypoglycemia events with a high degree of sensitivity and specificity. These models-which have been validated retrospectively and if implemented in real time-could be useful tools for reducing hypoglycemia in vulnerable patients.

Original languageEnglish (US)
Pages (from-to)86-90
Number of pages5
JournalJournal of diabetes science and technology
Volume9
Issue number1
DOIs
StatePublished - Jan 1 2015
Externally publishedYes

Fingerprint

Medical problems
Hypoglycemia
Type 2 Diabetes Mellitus
Learning systems
Blood Glucose
Glucose
Blood
Statistical Models
Machine Learning
Reading
Learning algorithms
Sensitivity and Specificity

Keywords

  • Hypoglycemia prediction
  • Machine learning
  • Type 2 diabetes

ASJC Scopus subject areas

  • Internal Medicine
  • Endocrinology, Diabetes and Metabolism
  • Bioengineering
  • Medicine(all)
  • Biomedical Engineering

Cite this

Hypoglycemia prediction using machine learning models for patients with type 2 diabetes. / Sudharsan, Bharath; Peeples, Malinda; Shomali, Mansur.

In: Journal of diabetes science and technology, Vol. 9, No. 1, 01.01.2015, p. 86-90.

Research output: Contribution to journalArticle

Sudharsan, Bharath ; Peeples, Malinda ; Shomali, Mansur. / Hypoglycemia prediction using machine learning models for patients with type 2 diabetes. In: Journal of diabetes science and technology. 2015 ; Vol. 9, No. 1. pp. 86-90.
@article{b4f2553dc6164d45866797d6f239d319,
title = "Hypoglycemia prediction using machine learning models for patients with type 2 diabetes",
abstract = "Background: Minimizing the occurrence of hypoglycemia in patients with type 2 diabetes is a challenging task since these patients typically check only 1 to 2 self-monitored blood glucose (SMBG) readings per day. Method: We trained a probabilistic model using machine learning algorithms and SMBG values from real patients. Hypoglycemia was defined as a SMBG value < 70 mg/dL. We validated our model using multiple data sets. In addition, we trained a second model, which used patient SMBG values and information about patient medication administration. Results: The optimal number of SMBG values needed by the model was approximately 10 per week. The sensitivity of the model for predicting a hypoglycemia event in the next 24 hours was 92{\%} and the specificity was 70{\%}. In the model that incorporated medication information, the prediction window was for the hour of hypoglycemia, and the specificity improved to 90{\%}. Conclusions: Our machine learning models can predict hypoglycemia events with a high degree of sensitivity and specificity. These models-which have been validated retrospectively and if implemented in real time-could be useful tools for reducing hypoglycemia in vulnerable patients.",
keywords = "Hypoglycemia prediction, Machine learning, Type 2 diabetes",
author = "Bharath Sudharsan and Malinda Peeples and Mansur Shomali",
year = "2015",
month = "1",
day = "1",
doi = "10.1177/1932296814554260",
language = "English (US)",
volume = "9",
pages = "86--90",
journal = "Journal of diabetes science and technology",
issn = "1932-2968",
publisher = "Diabetes Technology Society",
number = "1",

}

TY - JOUR

T1 - Hypoglycemia prediction using machine learning models for patients with type 2 diabetes

AU - Sudharsan, Bharath

AU - Peeples, Malinda

AU - Shomali, Mansur

PY - 2015/1/1

Y1 - 2015/1/1

N2 - Background: Minimizing the occurrence of hypoglycemia in patients with type 2 diabetes is a challenging task since these patients typically check only 1 to 2 self-monitored blood glucose (SMBG) readings per day. Method: We trained a probabilistic model using machine learning algorithms and SMBG values from real patients. Hypoglycemia was defined as a SMBG value < 70 mg/dL. We validated our model using multiple data sets. In addition, we trained a second model, which used patient SMBG values and information about patient medication administration. Results: The optimal number of SMBG values needed by the model was approximately 10 per week. The sensitivity of the model for predicting a hypoglycemia event in the next 24 hours was 92% and the specificity was 70%. In the model that incorporated medication information, the prediction window was for the hour of hypoglycemia, and the specificity improved to 90%. Conclusions: Our machine learning models can predict hypoglycemia events with a high degree of sensitivity and specificity. These models-which have been validated retrospectively and if implemented in real time-could be useful tools for reducing hypoglycemia in vulnerable patients.

AB - Background: Minimizing the occurrence of hypoglycemia in patients with type 2 diabetes is a challenging task since these patients typically check only 1 to 2 self-monitored blood glucose (SMBG) readings per day. Method: We trained a probabilistic model using machine learning algorithms and SMBG values from real patients. Hypoglycemia was defined as a SMBG value < 70 mg/dL. We validated our model using multiple data sets. In addition, we trained a second model, which used patient SMBG values and information about patient medication administration. Results: The optimal number of SMBG values needed by the model was approximately 10 per week. The sensitivity of the model for predicting a hypoglycemia event in the next 24 hours was 92% and the specificity was 70%. In the model that incorporated medication information, the prediction window was for the hour of hypoglycemia, and the specificity improved to 90%. Conclusions: Our machine learning models can predict hypoglycemia events with a high degree of sensitivity and specificity. These models-which have been validated retrospectively and if implemented in real time-could be useful tools for reducing hypoglycemia in vulnerable patients.

KW - Hypoglycemia prediction

KW - Machine learning

KW - Type 2 diabetes

UR - http://www.scopus.com/inward/record.url?scp=84924375509&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84924375509&partnerID=8YFLogxK

U2 - 10.1177/1932296814554260

DO - 10.1177/1932296814554260

M3 - Article

C2 - 25316712

AN - SCOPUS:84924375509

VL - 9

SP - 86

EP - 90

JO - Journal of diabetes science and technology

JF - Journal of diabetes science and technology

SN - 1932-2968

IS - 1

ER -