Using Stochastic modelling to identify unusual continuous glucose monitor measurements and behaviour, in newborn infants

Matthew Signal, Aaron Le Compte, Deborah Harris, Phil J. Weston, Jane Harding, J. G. Chase, Jane Alsweiler, Geoff Chase, Jane Harding, Deborah Harris, Ben Thompson, Trecia Wouldes, Judith Ansell, Nicola Anstice, Jo Arthur, Coila Bevan, Kate Bluett, Ellen Campbell, Arijit Chakraborty, Tineke Crawford & 16 others Karen Frost, Greg Gamble, Rob Jacobs, Kelly Jones, Aaron Le Compte, Sapphire Martin, Gill Matheson, Grace McKnight, Christina McQuoid, Janine Paynter, Jenny Rogers, Matthew Signal, Heather Stewart, Anna Timmings, Rebecca Young, Sandy Yu

Research output: Contribution to journalArticle

11 Citations (Scopus)
1 Downloads (Pure)

Abstract

Background: Abnormal blood glucose (BG) concentrations have been associated with increased morbidity and mortality in both critically ill adults and infants. Furthermore, hypoglycaemia and glycaemic variability have both been independently linked to mortality in these patients. Continuous Glucose Monitoring (CGM) devices have the potential to improve detection and diagnosis of these glycaemic abnormalities. However, sensor noise is a trade-off of the high measurement rate and must be managed effectively if CGMs are going to be used to monitor, diagnose and potentially help treat glycaemic abnormalities.Aim: To develop a tool that will aid clinicians in identifying unusual CGM behaviour and highlight CGM data that potentially need to be interpreted with care.Methods: CGM data and BG measurements from 50 infants at risk of hypoglycaemia were used. Unusual CGM measurements were classified using a stochastic model based on the kernel density method and historical CGM measurements from the cohort. CGM traces were colour coded with very unusual measurements coloured red, highlighting areas to be interpreted with care. A 5-fold validation of the model was Monte Carlo simulated 25 times to ensure an adequate model fit.Results: The stochastic model was generated using ~67,000 CGM measurements, spread across the glycaemic range ~2-10 mmol/L. A 5-fold validation showed a good model fit: the model 80% confidence interval (CI) captured 83% of clinical CGM data, the model 90% CI captured 91% of clinical CGM data, and the model 99% CI captured 99% of clinical CGM data. Three patient examples show the stochastic classification method in use with 1) A stable, low variability patient which shows no unusual CGM measurements, 2) A patient with a very sudden, short hypoglycaemic event (classified as unusual), and, 3) A patient with very high, potentially un-physiological, glycaemic variability after day 3 of monitoring (classified as very unusual).Conclusions: This study has produced a stochastic model and classification method capable of highlighting unusual CGM behaviour. This method has the potential to classify important glycaemic events (e.g. hypoglycaemia) as true clinical events or sensor noise, and to help identify possible sensor degradation. Colour coded CGM traces convey the information quickly and efficiently, while remaining computationally light enough to be used retrospectively or in real-time.

Original languageEnglish
Article number45
Pages (from-to)1-12
Number of pages12
JournalBioMedical Engineering OnLine
Volume11
DOIs
Publication statusPublished - 6 Aug 2012

Fingerprint

Glucose
Newborn Infant
Monitoring
Stochastic models
Hypoglycemia
Confidence Intervals
Blood Glucose
Noise
Sensors
Color
Blood
Mortality
Hypoglycemic Agents
Critical Illness
Morbidity
Light
Equipment and Supplies

Cite this

Signal, Matthew ; Le Compte, Aaron ; Harris, Deborah ; Weston, Phil J. ; Harding, Jane ; Chase, J. G. ; Alsweiler, Jane ; Chase, Geoff ; Harding, Jane ; Harris, Deborah ; Thompson, Ben ; Wouldes, Trecia ; Ansell, Judith ; Anstice, Nicola ; Arthur, Jo ; Bevan, Coila ; Bluett, Kate ; Campbell, Ellen ; Chakraborty, Arijit ; Crawford, Tineke ; Frost, Karen ; Gamble, Greg ; Jacobs, Rob ; Jones, Kelly ; Le Compte, Aaron ; Martin, Sapphire ; Matheson, Gill ; McKnight, Grace ; McQuoid, Christina ; Paynter, Janine ; Rogers, Jenny ; Signal, Matthew ; Stewart, Heather ; Timmings, Anna ; Young, Rebecca ; Yu, Sandy. / Using Stochastic modelling to identify unusual continuous glucose monitor measurements and behaviour, in newborn infants. In: BioMedical Engineering OnLine. 2012 ; Vol. 11. pp. 1-12.
@article{6e6f275b1d454cd5bc5b7cf317017814,
title = "Using Stochastic modelling to identify unusual continuous glucose monitor measurements and behaviour, in newborn infants",
abstract = "Background: Abnormal blood glucose (BG) concentrations have been associated with increased morbidity and mortality in both critically ill adults and infants. Furthermore, hypoglycaemia and glycaemic variability have both been independently linked to mortality in these patients. Continuous Glucose Monitoring (CGM) devices have the potential to improve detection and diagnosis of these glycaemic abnormalities. However, sensor noise is a trade-off of the high measurement rate and must be managed effectively if CGMs are going to be used to monitor, diagnose and potentially help treat glycaemic abnormalities.Aim: To develop a tool that will aid clinicians in identifying unusual CGM behaviour and highlight CGM data that potentially need to be interpreted with care.Methods: CGM data and BG measurements from 50 infants at risk of hypoglycaemia were used. Unusual CGM measurements were classified using a stochastic model based on the kernel density method and historical CGM measurements from the cohort. CGM traces were colour coded with very unusual measurements coloured red, highlighting areas to be interpreted with care. A 5-fold validation of the model was Monte Carlo simulated 25 times to ensure an adequate model fit.Results: The stochastic model was generated using ~67,000 CGM measurements, spread across the glycaemic range ~2-10 mmol/L. A 5-fold validation showed a good model fit: the model 80{\%} confidence interval (CI) captured 83{\%} of clinical CGM data, the model 90{\%} CI captured 91{\%} of clinical CGM data, and the model 99{\%} CI captured 99{\%} of clinical CGM data. Three patient examples show the stochastic classification method in use with 1) A stable, low variability patient which shows no unusual CGM measurements, 2) A patient with a very sudden, short hypoglycaemic event (classified as unusual), and, 3) A patient with very high, potentially un-physiological, glycaemic variability after day 3 of monitoring (classified as very unusual).Conclusions: This study has produced a stochastic model and classification method capable of highlighting unusual CGM behaviour. This method has the potential to classify important glycaemic events (e.g. hypoglycaemia) as true clinical events or sensor noise, and to help identify possible sensor degradation. Colour coded CGM traces convey the information quickly and efficiently, while remaining computationally light enough to be used retrospectively or in real-time.",
keywords = "Classify, Continuous Glucose Monitor, Glycaemia, Intensive care unit, Neonatal",
author = "Matthew Signal and {Le Compte}, Aaron and Deborah Harris and Weston, {Phil J.} and Jane Harding and Chase, {J. G.} and Jane Alsweiler and Geoff Chase and Jane Harding and Deborah Harris and Ben Thompson and Trecia Wouldes and Judith Ansell and Nicola Anstice and Jo Arthur and Coila Bevan and Kate Bluett and Ellen Campbell and Arijit Chakraborty and Tineke Crawford and Karen Frost and Greg Gamble and Rob Jacobs and Kelly Jones and {Le Compte}, Aaron and Sapphire Martin and Gill Matheson and Grace McKnight and Christina McQuoid and Janine Paynter and Jenny Rogers and Matthew Signal and Heather Stewart and Anna Timmings and Rebecca Young and Sandy Yu",
year = "2012",
month = "8",
day = "6",
doi = "10.1186/1475-925X-11-45",
language = "English",
volume = "11",
pages = "1--12",
journal = "BioMedical Engineering OnLine",
issn = "1475-925X",
publisher = "BioMed Central",

}

Signal, M, Le Compte, A, Harris, D, Weston, PJ, Harding, J, Chase, JG, Alsweiler, J, Chase, G, Harding, J, Harris, D, Thompson, B, Wouldes, T, Ansell, J, Anstice, N, Arthur, J, Bevan, C, Bluett, K, Campbell, E, Chakraborty, A, Crawford, T, Frost, K, Gamble, G, Jacobs, R, Jones, K, Le Compte, A, Martin, S, Matheson, G, McKnight, G, McQuoid, C, Paynter, J, Rogers, J, Signal, M, Stewart, H, Timmings, A, Young, R & Yu, S 2012, 'Using Stochastic modelling to identify unusual continuous glucose monitor measurements and behaviour, in newborn infants', BioMedical Engineering OnLine, vol. 11, 45, pp. 1-12. https://doi.org/10.1186/1475-925X-11-45

Using Stochastic modelling to identify unusual continuous glucose monitor measurements and behaviour, in newborn infants. / Signal, Matthew; Le Compte, Aaron; Harris, Deborah; Weston, Phil J.; Harding, Jane; Chase, J. G.; Alsweiler, Jane; Chase, Geoff; Harding, Jane; Harris, Deborah; Thompson, Ben; Wouldes, Trecia; Ansell, Judith; Anstice, Nicola; Arthur, Jo; Bevan, Coila; Bluett, Kate; Campbell, Ellen; Chakraborty, Arijit; Crawford, Tineke; Frost, Karen; Gamble, Greg; Jacobs, Rob; Jones, Kelly; Le Compte, Aaron; Martin, Sapphire; Matheson, Gill; McKnight, Grace; McQuoid, Christina; Paynter, Janine; Rogers, Jenny; Signal, Matthew; Stewart, Heather; Timmings, Anna; Young, Rebecca; Yu, Sandy.

In: BioMedical Engineering OnLine, Vol. 11, 45, 06.08.2012, p. 1-12.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Using Stochastic modelling to identify unusual continuous glucose monitor measurements and behaviour, in newborn infants

AU - Signal, Matthew

AU - Le Compte, Aaron

AU - Harris, Deborah

AU - Weston, Phil J.

AU - Harding, Jane

AU - Chase, J. G.

AU - Alsweiler, Jane

AU - Chase, Geoff

AU - Harding, Jane

AU - Harris, Deborah

AU - Thompson, Ben

AU - Wouldes, Trecia

AU - Ansell, Judith

AU - Anstice, Nicola

AU - Arthur, Jo

AU - Bevan, Coila

AU - Bluett, Kate

AU - Campbell, Ellen

AU - Chakraborty, Arijit

AU - Crawford, Tineke

AU - Frost, Karen

AU - Gamble, Greg

AU - Jacobs, Rob

AU - Jones, Kelly

AU - Le Compte, Aaron

AU - Martin, Sapphire

AU - Matheson, Gill

AU - McKnight, Grace

AU - McQuoid, Christina

AU - Paynter, Janine

AU - Rogers, Jenny

AU - Signal, Matthew

AU - Stewart, Heather

AU - Timmings, Anna

AU - Young, Rebecca

AU - Yu, Sandy

PY - 2012/8/6

Y1 - 2012/8/6

N2 - Background: Abnormal blood glucose (BG) concentrations have been associated with increased morbidity and mortality in both critically ill adults and infants. Furthermore, hypoglycaemia and glycaemic variability have both been independently linked to mortality in these patients. Continuous Glucose Monitoring (CGM) devices have the potential to improve detection and diagnosis of these glycaemic abnormalities. However, sensor noise is a trade-off of the high measurement rate and must be managed effectively if CGMs are going to be used to monitor, diagnose and potentially help treat glycaemic abnormalities.Aim: To develop a tool that will aid clinicians in identifying unusual CGM behaviour and highlight CGM data that potentially need to be interpreted with care.Methods: CGM data and BG measurements from 50 infants at risk of hypoglycaemia were used. Unusual CGM measurements were classified using a stochastic model based on the kernel density method and historical CGM measurements from the cohort. CGM traces were colour coded with very unusual measurements coloured red, highlighting areas to be interpreted with care. A 5-fold validation of the model was Monte Carlo simulated 25 times to ensure an adequate model fit.Results: The stochastic model was generated using ~67,000 CGM measurements, spread across the glycaemic range ~2-10 mmol/L. A 5-fold validation showed a good model fit: the model 80% confidence interval (CI) captured 83% of clinical CGM data, the model 90% CI captured 91% of clinical CGM data, and the model 99% CI captured 99% of clinical CGM data. Three patient examples show the stochastic classification method in use with 1) A stable, low variability patient which shows no unusual CGM measurements, 2) A patient with a very sudden, short hypoglycaemic event (classified as unusual), and, 3) A patient with very high, potentially un-physiological, glycaemic variability after day 3 of monitoring (classified as very unusual).Conclusions: This study has produced a stochastic model and classification method capable of highlighting unusual CGM behaviour. This method has the potential to classify important glycaemic events (e.g. hypoglycaemia) as true clinical events or sensor noise, and to help identify possible sensor degradation. Colour coded CGM traces convey the information quickly and efficiently, while remaining computationally light enough to be used retrospectively or in real-time.

AB - Background: Abnormal blood glucose (BG) concentrations have been associated with increased morbidity and mortality in both critically ill adults and infants. Furthermore, hypoglycaemia and glycaemic variability have both been independently linked to mortality in these patients. Continuous Glucose Monitoring (CGM) devices have the potential to improve detection and diagnosis of these glycaemic abnormalities. However, sensor noise is a trade-off of the high measurement rate and must be managed effectively if CGMs are going to be used to monitor, diagnose and potentially help treat glycaemic abnormalities.Aim: To develop a tool that will aid clinicians in identifying unusual CGM behaviour and highlight CGM data that potentially need to be interpreted with care.Methods: CGM data and BG measurements from 50 infants at risk of hypoglycaemia were used. Unusual CGM measurements were classified using a stochastic model based on the kernel density method and historical CGM measurements from the cohort. CGM traces were colour coded with very unusual measurements coloured red, highlighting areas to be interpreted with care. A 5-fold validation of the model was Monte Carlo simulated 25 times to ensure an adequate model fit.Results: The stochastic model was generated using ~67,000 CGM measurements, spread across the glycaemic range ~2-10 mmol/L. A 5-fold validation showed a good model fit: the model 80% confidence interval (CI) captured 83% of clinical CGM data, the model 90% CI captured 91% of clinical CGM data, and the model 99% CI captured 99% of clinical CGM data. Three patient examples show the stochastic classification method in use with 1) A stable, low variability patient which shows no unusual CGM measurements, 2) A patient with a very sudden, short hypoglycaemic event (classified as unusual), and, 3) A patient with very high, potentially un-physiological, glycaemic variability after day 3 of monitoring (classified as very unusual).Conclusions: This study has produced a stochastic model and classification method capable of highlighting unusual CGM behaviour. This method has the potential to classify important glycaemic events (e.g. hypoglycaemia) as true clinical events or sensor noise, and to help identify possible sensor degradation. Colour coded CGM traces convey the information quickly and efficiently, while remaining computationally light enough to be used retrospectively or in real-time.

KW - Classify

KW - Continuous Glucose Monitor

KW - Glycaemia

KW - Intensive care unit

KW - Neonatal

UR - http://www.scopus.com/inward/record.url?scp=84864519217&partnerID=8YFLogxK

U2 - 10.1186/1475-925X-11-45

DO - 10.1186/1475-925X-11-45

M3 - Article

VL - 11

SP - 1

EP - 12

JO - BioMedical Engineering OnLine

JF - BioMedical Engineering OnLine

SN - 1475-925X

M1 - 45

ER -