Early prediction of acute kidney injury following ICU admission using a multivariate panel of physiological measurements

Lindsay P. Zimmerman, Paul Andrew Reyfman, Angela D.R. Smith, Zexian Zeng, Abel N Kho, Lazaro Nelson Sanchez-Pinto, Yuan Luo*

*Corresponding author for this work

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Background: The development of acute kidney injury (AKI) during an intensive care unit (ICU) admission is associated with increased morbidity and mortality. Methods: Our objective was to develop and validate a data driven multivariable clinical predictive model for early detection of AKI among a large cohort of adult critical care patients. We utilized data form the Medical Information Mart for Intensive Care III (MIMIC-III) for all patients who had a creatinine measured for 3 days following ICU admission and excluded patients with pre-existing condition of Chronic Kidney Disease and Acute Kidney Injury on admission. Data extracted included patient age, gender, ethnicity, creatinine, other vital signs and lab values during the first day of ICU admission, whether the patient was mechanically ventilated during the first day of ICU admission, and the hourly rate of urine output during the first day of ICU admission. Results: Utilizing the demographics, the clinical data and the laboratory test measurements from Day 1 of ICU admission, we accurately predicted max serum creatinine level during Day 2 and Day 3 with a root mean square error of 0.224 mg/dL. We demonstrated that using machine learning models (multivariate logistic regression, random forest and artificial neural networks) with demographics and physiologic features can predict AKI onset as defined by the current clinical guideline with a competitive AUC (mean AUC 0.783 by our all-feature, logistic-regression model), while previous models aimed at more specific patient cohorts. Conclusions: Experimental results suggest that our model has the potential to assist clinicians in identifying patients at greater risk of new onset of AKI in critical care setting. Prospective trials with independent model training and external validation cohorts are needed to further evaluate the clinical utility of this approach and potentially instituting interventions to decrease the likelihood of developing AKI.

Original languageEnglish (US)
Article number16
JournalBMC Medical Informatics and Decision Making
Volume19
DOIs
StatePublished - Jan 31 2019

Fingerprint

Acute Kidney Injury
Intensive Care Units
Critical Care
Creatinine
Logistic Models
Patient Admission
Area Under Curve
Demography
Preexisting Condition Coverage
Vital Signs
Chronic Renal Insufficiency
Urine
Guidelines
Morbidity
Mortality
Serum

Keywords

  • Acute kidney injury
  • Artificial neural networks
  • Intensive care unit
  • Multivariate logistic regression
  • Physiological measurements
  • Predictive modeling
  • Random forest

ASJC Scopus subject areas

  • Health Policy
  • Health Informatics

Cite this

@article{b7c1eefbff3a4cabb3745bd234c8dece,
title = "Early prediction of acute kidney injury following ICU admission using a multivariate panel of physiological measurements",
abstract = "Background: The development of acute kidney injury (AKI) during an intensive care unit (ICU) admission is associated with increased morbidity and mortality. Methods: Our objective was to develop and validate a data driven multivariable clinical predictive model for early detection of AKI among a large cohort of adult critical care patients. We utilized data form the Medical Information Mart for Intensive Care III (MIMIC-III) for all patients who had a creatinine measured for 3 days following ICU admission and excluded patients with pre-existing condition of Chronic Kidney Disease and Acute Kidney Injury on admission. Data extracted included patient age, gender, ethnicity, creatinine, other vital signs and lab values during the first day of ICU admission, whether the patient was mechanically ventilated during the first day of ICU admission, and the hourly rate of urine output during the first day of ICU admission. Results: Utilizing the demographics, the clinical data and the laboratory test measurements from Day 1 of ICU admission, we accurately predicted max serum creatinine level during Day 2 and Day 3 with a root mean square error of 0.224 mg/dL. We demonstrated that using machine learning models (multivariate logistic regression, random forest and artificial neural networks) with demographics and physiologic features can predict AKI onset as defined by the current clinical guideline with a competitive AUC (mean AUC 0.783 by our all-feature, logistic-regression model), while previous models aimed at more specific patient cohorts. Conclusions: Experimental results suggest that our model has the potential to assist clinicians in identifying patients at greater risk of new onset of AKI in critical care setting. Prospective trials with independent model training and external validation cohorts are needed to further evaluate the clinical utility of this approach and potentially instituting interventions to decrease the likelihood of developing AKI.",
keywords = "Acute kidney injury, Artificial neural networks, Intensive care unit, Multivariate logistic regression, Physiological measurements, Predictive modeling, Random forest",
author = "Zimmerman, {Lindsay P.} and Reyfman, {Paul Andrew} and Smith, {Angela D.R.} and Zexian Zeng and Kho, {Abel N} and Sanchez-Pinto, {Lazaro Nelson} and Yuan Luo",
year = "2019",
month = "1",
day = "31",
doi = "10.1186/s12911-019-0733-z",
language = "English (US)",
volume = "19",
journal = "BMC Medical Informatics and Decision Making",
issn = "1472-6947",
publisher = "BioMed Central",

}

TY - JOUR

T1 - Early prediction of acute kidney injury following ICU admission using a multivariate panel of physiological measurements

AU - Zimmerman, Lindsay P.

AU - Reyfman, Paul Andrew

AU - Smith, Angela D.R.

AU - Zeng, Zexian

AU - Kho, Abel N

AU - Sanchez-Pinto, Lazaro Nelson

AU - Luo, Yuan

PY - 2019/1/31

Y1 - 2019/1/31

N2 - Background: The development of acute kidney injury (AKI) during an intensive care unit (ICU) admission is associated with increased morbidity and mortality. Methods: Our objective was to develop and validate a data driven multivariable clinical predictive model for early detection of AKI among a large cohort of adult critical care patients. We utilized data form the Medical Information Mart for Intensive Care III (MIMIC-III) for all patients who had a creatinine measured for 3 days following ICU admission and excluded patients with pre-existing condition of Chronic Kidney Disease and Acute Kidney Injury on admission. Data extracted included patient age, gender, ethnicity, creatinine, other vital signs and lab values during the first day of ICU admission, whether the patient was mechanically ventilated during the first day of ICU admission, and the hourly rate of urine output during the first day of ICU admission. Results: Utilizing the demographics, the clinical data and the laboratory test measurements from Day 1 of ICU admission, we accurately predicted max serum creatinine level during Day 2 and Day 3 with a root mean square error of 0.224 mg/dL. We demonstrated that using machine learning models (multivariate logistic regression, random forest and artificial neural networks) with demographics and physiologic features can predict AKI onset as defined by the current clinical guideline with a competitive AUC (mean AUC 0.783 by our all-feature, logistic-regression model), while previous models aimed at more specific patient cohorts. Conclusions: Experimental results suggest that our model has the potential to assist clinicians in identifying patients at greater risk of new onset of AKI in critical care setting. Prospective trials with independent model training and external validation cohorts are needed to further evaluate the clinical utility of this approach and potentially instituting interventions to decrease the likelihood of developing AKI.

AB - Background: The development of acute kidney injury (AKI) during an intensive care unit (ICU) admission is associated with increased morbidity and mortality. Methods: Our objective was to develop and validate a data driven multivariable clinical predictive model for early detection of AKI among a large cohort of adult critical care patients. We utilized data form the Medical Information Mart for Intensive Care III (MIMIC-III) for all patients who had a creatinine measured for 3 days following ICU admission and excluded patients with pre-existing condition of Chronic Kidney Disease and Acute Kidney Injury on admission. Data extracted included patient age, gender, ethnicity, creatinine, other vital signs and lab values during the first day of ICU admission, whether the patient was mechanically ventilated during the first day of ICU admission, and the hourly rate of urine output during the first day of ICU admission. Results: Utilizing the demographics, the clinical data and the laboratory test measurements from Day 1 of ICU admission, we accurately predicted max serum creatinine level during Day 2 and Day 3 with a root mean square error of 0.224 mg/dL. We demonstrated that using machine learning models (multivariate logistic regression, random forest and artificial neural networks) with demographics and physiologic features can predict AKI onset as defined by the current clinical guideline with a competitive AUC (mean AUC 0.783 by our all-feature, logistic-regression model), while previous models aimed at more specific patient cohorts. Conclusions: Experimental results suggest that our model has the potential to assist clinicians in identifying patients at greater risk of new onset of AKI in critical care setting. Prospective trials with independent model training and external validation cohorts are needed to further evaluate the clinical utility of this approach and potentially instituting interventions to decrease the likelihood of developing AKI.

KW - Acute kidney injury

KW - Artificial neural networks

KW - Intensive care unit

KW - Multivariate logistic regression

KW - Physiological measurements

KW - Predictive modeling

KW - Random forest

UR - http://www.scopus.com/inward/record.url?scp=85060861835&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85060861835&partnerID=8YFLogxK

U2 - 10.1186/s12911-019-0733-z

DO - 10.1186/s12911-019-0733-z

M3 - Article

VL - 19

JO - BMC Medical Informatics and Decision Making

JF - BMC Medical Informatics and Decision Making

SN - 1472-6947

M1 - 16

ER -