Stacked classifiers for individualized prediction of glycemic control following initiation of metformin therapy in type 2 diabetes

Dennis H. Murphree, Elaheh Arabmakki, Che Ngufor, Curtis Storlie, Rozalina McCoy

Research output: Contribution to journalArticle

Abstract

Objective: Metformin is the preferred first-line medication for management of type 2 diabetes and prediabetes. However, over a third of patients experience primary or secondary therapeutic failure. We developed machine learning models to predict which patients initially prescribed metformin will achieve and maintain control of their blood glucose after one year of therapy. Materials and methods: We performed a retrospective analysis of administrative claims data for 12,147 commercially-insured adults and Medicare Advantage beneficiaries with prediabetes or diabetes. Several machine learning models were trained using variables available at the time of metformin initiation to predict achievement and maintenance of hemoglobin A1c (HbA1c) < 7.0% after one year of therapy. Results: AUC performances based on five-fold cross-validation ranged from 0.58 to 0.75. The most influential variables driving the predictions were baseline HbA1c, starting metformin dosage, and presence of diabetes with complications. Conclusions: Machine learning models can effectively predict primary or secondary metformin treatment failure within one year. This information can help identify effective individualized treatment strategies. Most of the implemented models outperformed traditional logistic regression, highlighting the potential for applying machine learning to problems in medicine.

Original languageEnglish (US)
Pages (from-to)109-115
Number of pages7
JournalComputers in Biology and Medicine
Volume103
DOIs
StatePublished - Dec 1 2018

Fingerprint

Metformin
Medical problems
Type 2 Diabetes Mellitus
Learning systems
Classifiers
Hemoglobin
Prediabetic State
Hemoglobins
Medicare Part C
Insurance Claim Review
Bioelectric potentials
Therapeutics
Medicine
Glucose
Logistics
Diabetes Complications
Treatment Failure
Blood
Area Under Curve
Blood Glucose

Keywords

  • Clinical
  • Decision support systems
  • Diabetes mellitus
  • Machine learning
  • Precision medicine

ASJC Scopus subject areas

  • Computer Science Applications
  • Health Informatics

Cite this

@article{cc576d87ef9c45de81684543450b31f3,
title = "Stacked classifiers for individualized prediction of glycemic control following initiation of metformin therapy in type 2 diabetes",
abstract = "Objective: Metformin is the preferred first-line medication for management of type 2 diabetes and prediabetes. However, over a third of patients experience primary or secondary therapeutic failure. We developed machine learning models to predict which patients initially prescribed metformin will achieve and maintain control of their blood glucose after one year of therapy. Materials and methods: We performed a retrospective analysis of administrative claims data for 12,147 commercially-insured adults and Medicare Advantage beneficiaries with prediabetes or diabetes. Several machine learning models were trained using variables available at the time of metformin initiation to predict achievement and maintenance of hemoglobin A1c (HbA1c) < 7.0{\%} after one year of therapy. Results: AUC performances based on five-fold cross-validation ranged from 0.58 to 0.75. The most influential variables driving the predictions were baseline HbA1c, starting metformin dosage, and presence of diabetes with complications. Conclusions: Machine learning models can effectively predict primary or secondary metformin treatment failure within one year. This information can help identify effective individualized treatment strategies. Most of the implemented models outperformed traditional logistic regression, highlighting the potential for applying machine learning to problems in medicine.",
keywords = "Clinical, Decision support systems, Diabetes mellitus, Machine learning, Precision medicine",
author = "Murphree, {Dennis H.} and Elaheh Arabmakki and Che Ngufor and Curtis Storlie and Rozalina McCoy",
year = "2018",
month = "12",
day = "1",
doi = "10.1016/j.compbiomed.2018.10.017",
language = "English (US)",
volume = "103",
pages = "109--115",
journal = "Computers in Biology and Medicine",
issn = "0010-4825",
publisher = "Elsevier Limited",

}

TY - JOUR

T1 - Stacked classifiers for individualized prediction of glycemic control following initiation of metformin therapy in type 2 diabetes

AU - Murphree, Dennis H.

AU - Arabmakki, Elaheh

AU - Ngufor, Che

AU - Storlie, Curtis

AU - McCoy, Rozalina

PY - 2018/12/1

Y1 - 2018/12/1

N2 - Objective: Metformin is the preferred first-line medication for management of type 2 diabetes and prediabetes. However, over a third of patients experience primary or secondary therapeutic failure. We developed machine learning models to predict which patients initially prescribed metformin will achieve and maintain control of their blood glucose after one year of therapy. Materials and methods: We performed a retrospective analysis of administrative claims data for 12,147 commercially-insured adults and Medicare Advantage beneficiaries with prediabetes or diabetes. Several machine learning models were trained using variables available at the time of metformin initiation to predict achievement and maintenance of hemoglobin A1c (HbA1c) < 7.0% after one year of therapy. Results: AUC performances based on five-fold cross-validation ranged from 0.58 to 0.75. The most influential variables driving the predictions were baseline HbA1c, starting metformin dosage, and presence of diabetes with complications. Conclusions: Machine learning models can effectively predict primary or secondary metformin treatment failure within one year. This information can help identify effective individualized treatment strategies. Most of the implemented models outperformed traditional logistic regression, highlighting the potential for applying machine learning to problems in medicine.

AB - Objective: Metformin is the preferred first-line medication for management of type 2 diabetes and prediabetes. However, over a third of patients experience primary or secondary therapeutic failure. We developed machine learning models to predict which patients initially prescribed metformin will achieve and maintain control of their blood glucose after one year of therapy. Materials and methods: We performed a retrospective analysis of administrative claims data for 12,147 commercially-insured adults and Medicare Advantage beneficiaries with prediabetes or diabetes. Several machine learning models were trained using variables available at the time of metformin initiation to predict achievement and maintenance of hemoglobin A1c (HbA1c) < 7.0% after one year of therapy. Results: AUC performances based on five-fold cross-validation ranged from 0.58 to 0.75. The most influential variables driving the predictions were baseline HbA1c, starting metformin dosage, and presence of diabetes with complications. Conclusions: Machine learning models can effectively predict primary or secondary metformin treatment failure within one year. This information can help identify effective individualized treatment strategies. Most of the implemented models outperformed traditional logistic regression, highlighting the potential for applying machine learning to problems in medicine.

KW - Clinical

KW - Decision support systems

KW - Diabetes mellitus

KW - Machine learning

KW - Precision medicine

UR - http://www.scopus.com/inward/record.url?scp=85055109064&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85055109064&partnerID=8YFLogxK

U2 - 10.1016/j.compbiomed.2018.10.017

DO - 10.1016/j.compbiomed.2018.10.017

M3 - Article

VL - 103

SP - 109

EP - 115

JO - Computers in Biology and Medicine

JF - Computers in Biology and Medicine

SN - 0010-4825

ER -