A methodological comparison of risk scores versus decision trees for predicting drug-resistant infections: A case study using extended-spectrum beta-lactamase (ESBL) bacteremia

Katherine E. Goodman, Justin T Lessler, Anthony D. Harris, Aaron Milstone, Pranita Tamma

Research output: Contribution to journalArticle

Abstract

Background:Timely identification of multidrug-resistant gram-negative infections remains an epidemiological challenge. Statistical models for predicting drug resistance can offer utility where rapid diagnostics are unavailable or resource-impractical. Logistic regression-derived risk scores are common in the healthcare epidemiology literature. Machine learning-derived decision trees are an alternative approach for developing decision support tools. Our group previously reported on a decision tree for predicting ESBL bloodstream infections. Our objective in the current study was to develop a risk score from the same ESBL dataset to compare these 2 methods and to offer general guiding principles for using each approach.Methods:Using a dataset of 1,288 patients with Escherichia coli or Klebsiella spp bacteremia, we generated a risk score to predict the likelihood that a bacteremic patient was infected with an ESBL-producer. We evaluated discrimination (original and cross-validated models) using receiver operating characteristic curves and C statistics. We compared risk score and decision tree performance, and we reviewed their practical and methodological attributes.Results:In total, 194 patients (15%) were infected with ESBL-producing bacteremia. The clinical risk score included 14 variables, compared to the 5 decision-tree variables. The positive and negative predictive values of the risk score and decision tree were similar (>90%), but the C statistic of the risk score (0.87) was 10% higher.Conclusions:A decision tree and risk score performed similarly for predicting ESBL infection. The decision tree was more user-friendly, with fewer variables for the end user, whereas the risk score offered higher discrimination and greater flexibility for adjusting sensitivity and specificity.

Original languageEnglish (US)
JournalInfection control and hospital epidemiology
DOIs
StatePublished - Jan 1 2019

Fingerprint

Decision Trees
beta-Lactamases
Bacteremia
Infection
Pharmaceutical Preparations
Klebsiella
Statistical Models
Drug Resistance
ROC Curve
Epidemiology
Logistic Models
Escherichia coli
Delivery of Health Care
Sensitivity and Specificity

ASJC Scopus subject areas

  • Epidemiology
  • Microbiology (medical)
  • Infectious Diseases

Cite this

@article{e68f94d9e56a4c48b825fc520dad7816,
title = "A methodological comparison of risk scores versus decision trees for predicting drug-resistant infections: A case study using extended-spectrum beta-lactamase (ESBL) bacteremia",
abstract = "Background:Timely identification of multidrug-resistant gram-negative infections remains an epidemiological challenge. Statistical models for predicting drug resistance can offer utility where rapid diagnostics are unavailable or resource-impractical. Logistic regression-derived risk scores are common in the healthcare epidemiology literature. Machine learning-derived decision trees are an alternative approach for developing decision support tools. Our group previously reported on a decision tree for predicting ESBL bloodstream infections. Our objective in the current study was to develop a risk score from the same ESBL dataset to compare these 2 methods and to offer general guiding principles for using each approach.Methods:Using a dataset of 1,288 patients with Escherichia coli or Klebsiella spp bacteremia, we generated a risk score to predict the likelihood that a bacteremic patient was infected with an ESBL-producer. We evaluated discrimination (original and cross-validated models) using receiver operating characteristic curves and C statistics. We compared risk score and decision tree performance, and we reviewed their practical and methodological attributes.Results:In total, 194 patients (15{\%}) were infected with ESBL-producing bacteremia. The clinical risk score included 14 variables, compared to the 5 decision-tree variables. The positive and negative predictive values of the risk score and decision tree were similar (>90{\%}), but the C statistic of the risk score (0.87) was 10{\%} higher.Conclusions:A decision tree and risk score performed similarly for predicting ESBL infection. The decision tree was more user-friendly, with fewer variables for the end user, whereas the risk score offered higher discrimination and greater flexibility for adjusting sensitivity and specificity.",
author = "Goodman, {Katherine E.} and Lessler, {Justin T} and Harris, {Anthony D.} and Aaron Milstone and Pranita Tamma",
year = "2019",
month = "1",
day = "1",
doi = "10.1017/ice.2019.17",
language = "English (US)",
journal = "Infection Control and Hospital Epidemiology",
issn = "0899-823X",
publisher = "University of Chicago Press",

}

TY - JOUR

T1 - A methodological comparison of risk scores versus decision trees for predicting drug-resistant infections

T2 - A case study using extended-spectrum beta-lactamase (ESBL) bacteremia

AU - Goodman, Katherine E.

AU - Lessler, Justin T

AU - Harris, Anthony D.

AU - Milstone, Aaron

AU - Tamma, Pranita

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Background:Timely identification of multidrug-resistant gram-negative infections remains an epidemiological challenge. Statistical models for predicting drug resistance can offer utility where rapid diagnostics are unavailable or resource-impractical. Logistic regression-derived risk scores are common in the healthcare epidemiology literature. Machine learning-derived decision trees are an alternative approach for developing decision support tools. Our group previously reported on a decision tree for predicting ESBL bloodstream infections. Our objective in the current study was to develop a risk score from the same ESBL dataset to compare these 2 methods and to offer general guiding principles for using each approach.Methods:Using a dataset of 1,288 patients with Escherichia coli or Klebsiella spp bacteremia, we generated a risk score to predict the likelihood that a bacteremic patient was infected with an ESBL-producer. We evaluated discrimination (original and cross-validated models) using receiver operating characteristic curves and C statistics. We compared risk score and decision tree performance, and we reviewed their practical and methodological attributes.Results:In total, 194 patients (15%) were infected with ESBL-producing bacteremia. The clinical risk score included 14 variables, compared to the 5 decision-tree variables. The positive and negative predictive values of the risk score and decision tree were similar (>90%), but the C statistic of the risk score (0.87) was 10% higher.Conclusions:A decision tree and risk score performed similarly for predicting ESBL infection. The decision tree was more user-friendly, with fewer variables for the end user, whereas the risk score offered higher discrimination and greater flexibility for adjusting sensitivity and specificity.

AB - Background:Timely identification of multidrug-resistant gram-negative infections remains an epidemiological challenge. Statistical models for predicting drug resistance can offer utility where rapid diagnostics are unavailable or resource-impractical. Logistic regression-derived risk scores are common in the healthcare epidemiology literature. Machine learning-derived decision trees are an alternative approach for developing decision support tools. Our group previously reported on a decision tree for predicting ESBL bloodstream infections. Our objective in the current study was to develop a risk score from the same ESBL dataset to compare these 2 methods and to offer general guiding principles for using each approach.Methods:Using a dataset of 1,288 patients with Escherichia coli or Klebsiella spp bacteremia, we generated a risk score to predict the likelihood that a bacteremic patient was infected with an ESBL-producer. We evaluated discrimination (original and cross-validated models) using receiver operating characteristic curves and C statistics. We compared risk score and decision tree performance, and we reviewed their practical and methodological attributes.Results:In total, 194 patients (15%) were infected with ESBL-producing bacteremia. The clinical risk score included 14 variables, compared to the 5 decision-tree variables. The positive and negative predictive values of the risk score and decision tree were similar (>90%), but the C statistic of the risk score (0.87) was 10% higher.Conclusions:A decision tree and risk score performed similarly for predicting ESBL infection. The decision tree was more user-friendly, with fewer variables for the end user, whereas the risk score offered higher discrimination and greater flexibility for adjusting sensitivity and specificity.

UR - http://www.scopus.com/inward/record.url?scp=85062369986&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85062369986&partnerID=8YFLogxK

U2 - 10.1017/ice.2019.17

DO - 10.1017/ice.2019.17

M3 - Article

C2 - 30827286

AN - SCOPUS:85062369986

JO - Infection Control and Hospital Epidemiology

JF - Infection Control and Hospital Epidemiology

SN - 0899-823X

ER -