A data analytics approach to building a clinical decision support system for diabetic retinopathy: Developing and deploying a model ensemble

Saeed Piri, Dursun Delen, Tieming Liu, Hamed M. Zolbanin

Research output: Contribution to journalArticle

18 Citations (Scopus)

Abstract

Diabetes is a common chronic disease that may lead to several complications. Diabetic retinopathy (DR), one of the most serious of these complications, is the most common cause of vision loss among diabetic patients. In this paper, we analyzed data from more than 1.4 million diabetics and developed a clinical decision support system (CDSS) for predicting DR. While the existing diagnostic approach requires access to ophthalmologists and expensive equipment, our CDSS only uses demographic and lab data to detect patients' susceptibility to retinopathy with a high accuracy. We illustrate how a combination of multiple data preparation and modeling steps helped us improve the performance of our CDSS. From the data preprocessing aspect, we aggregated the data at the patient level and incorporated comorbidity information into our models. From the modeling perspective, we built several predictive models and developed a novel “confidence margin” ensemble technique that outperformed the existing ensemble models. Our results suggest that diabetic neuropathy, creatinine serum, blood urea nitrogen, glucose serum plasma, and hematocrit are the most important variables in detecting DR. Our CDSS provides several important practical implications, including identifying the DR risk factors, facilitating the early diagnosis of DR, and solving the problem of low compliance with annual retinopathy screenings.

Original languageEnglish
Pages (from-to)12-27
Number of pages16
JournalDecision Support Systems
Volume101
DOIs
StatePublished - 1 Sep 2017
Externally publishedYes

Fingerprint

Clinical Decision Support Systems
Diabetic Retinopathy
Decision support systems
Medical problems
Urea
Diabetic Neuropathies
Blood Urea Nitrogen
Glucose
Screening
Serum
Blood
Hematocrit
Comorbidity
Early Diagnosis
Nitrogen
Creatinine
Plasmas
Chronic Disease
Demography
Ensemble

Keywords

  • Clinical decision support systems
  • Data analytics
  • Diabetic retinopathy
  • Model ensembles
  • Predictive modeling
  • Variable importance

Cite this

@article{59d5141b2b0448268846abd048f2122d,
title = "A data analytics approach to building a clinical decision support system for diabetic retinopathy: Developing and deploying a model ensemble",
abstract = "Diabetes is a common chronic disease that may lead to several complications. Diabetic retinopathy (DR), one of the most serious of these complications, is the most common cause of vision loss among diabetic patients. In this paper, we analyzed data from more than 1.4 million diabetics and developed a clinical decision support system (CDSS) for predicting DR. While the existing diagnostic approach requires access to ophthalmologists and expensive equipment, our CDSS only uses demographic and lab data to detect patients' susceptibility to retinopathy with a high accuracy. We illustrate how a combination of multiple data preparation and modeling steps helped us improve the performance of our CDSS. From the data preprocessing aspect, we aggregated the data at the patient level and incorporated comorbidity information into our models. From the modeling perspective, we built several predictive models and developed a novel “confidence margin” ensemble technique that outperformed the existing ensemble models. Our results suggest that diabetic neuropathy, creatinine serum, blood urea nitrogen, glucose serum plasma, and hematocrit are the most important variables in detecting DR. Our CDSS provides several important practical implications, including identifying the DR risk factors, facilitating the early diagnosis of DR, and solving the problem of low compliance with annual retinopathy screenings.",
keywords = "Clinical decision support systems, Data analytics, Diabetic retinopathy, Model ensembles, Predictive modeling, Variable importance",
author = "Saeed Piri and Dursun Delen and Tieming Liu and Zolbanin, {Hamed M.}",
year = "2017",
month = "9",
day = "1",
doi = "10.1016/j.dss.2017.05.012",
language = "English",
volume = "101",
pages = "12--27",
journal = "Decision Support Systems",
issn = "0167-9236",
publisher = "Elsevier",

}

A data analytics approach to building a clinical decision support system for diabetic retinopathy : Developing and deploying a model ensemble. / Piri, Saeed; Delen, Dursun; Liu, Tieming; Zolbanin, Hamed M.

In: Decision Support Systems, Vol. 101, 01.09.2017, p. 12-27.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A data analytics approach to building a clinical decision support system for diabetic retinopathy

T2 - Developing and deploying a model ensemble

AU - Piri, Saeed

AU - Delen, Dursun

AU - Liu, Tieming

AU - Zolbanin, Hamed M.

PY - 2017/9/1

Y1 - 2017/9/1

N2 - Diabetes is a common chronic disease that may lead to several complications. Diabetic retinopathy (DR), one of the most serious of these complications, is the most common cause of vision loss among diabetic patients. In this paper, we analyzed data from more than 1.4 million diabetics and developed a clinical decision support system (CDSS) for predicting DR. While the existing diagnostic approach requires access to ophthalmologists and expensive equipment, our CDSS only uses demographic and lab data to detect patients' susceptibility to retinopathy with a high accuracy. We illustrate how a combination of multiple data preparation and modeling steps helped us improve the performance of our CDSS. From the data preprocessing aspect, we aggregated the data at the patient level and incorporated comorbidity information into our models. From the modeling perspective, we built several predictive models and developed a novel “confidence margin” ensemble technique that outperformed the existing ensemble models. Our results suggest that diabetic neuropathy, creatinine serum, blood urea nitrogen, glucose serum plasma, and hematocrit are the most important variables in detecting DR. Our CDSS provides several important practical implications, including identifying the DR risk factors, facilitating the early diagnosis of DR, and solving the problem of low compliance with annual retinopathy screenings.

AB - Diabetes is a common chronic disease that may lead to several complications. Diabetic retinopathy (DR), one of the most serious of these complications, is the most common cause of vision loss among diabetic patients. In this paper, we analyzed data from more than 1.4 million diabetics and developed a clinical decision support system (CDSS) for predicting DR. While the existing diagnostic approach requires access to ophthalmologists and expensive equipment, our CDSS only uses demographic and lab data to detect patients' susceptibility to retinopathy with a high accuracy. We illustrate how a combination of multiple data preparation and modeling steps helped us improve the performance of our CDSS. From the data preprocessing aspect, we aggregated the data at the patient level and incorporated comorbidity information into our models. From the modeling perspective, we built several predictive models and developed a novel “confidence margin” ensemble technique that outperformed the existing ensemble models. Our results suggest that diabetic neuropathy, creatinine serum, blood urea nitrogen, glucose serum plasma, and hematocrit are the most important variables in detecting DR. Our CDSS provides several important practical implications, including identifying the DR risk factors, facilitating the early diagnosis of DR, and solving the problem of low compliance with annual retinopathy screenings.

KW - Clinical decision support systems

KW - Data analytics

KW - Diabetic retinopathy

KW - Model ensembles

KW - Predictive modeling

KW - Variable importance

UR - http://www.scopus.com/inward/record.url?scp=85020058737&partnerID=8YFLogxK

U2 - 10.1016/j.dss.2017.05.012

DO - 10.1016/j.dss.2017.05.012

M3 - Article

AN - SCOPUS:85020058737

VL - 101

SP - 12

EP - 27

JO - Decision Support Systems

JF - Decision Support Systems

SN - 0167-9236

ER -