A non-parametric method for the comparison of partial areas under ROC curves and its application to large health care data sets

Dong Zhang, Xia Hua Zhou, Daniel H. Freeman, Jean L. Freeman

Research output: Contribution to journalArticle

93 Citations (Scopus)

Abstract

The receiver operating characteristic (ROC) curve is a statistical tool for evaluating the accuracy of diagnostic tests. Investigators often compare the validity of two tests based on the estimated areas under the respective ROC curves. However, the traditional way of comparing entire areas under two ROC curves is not sensitive when two ROC curves cross each other. Also, there are some cutpoints on the ROC curves that are not considered in practice because their corresponding sensitivities or specificities are unacceptable. For the purpose of comparing the partial area under the curve (AUC) within a specific range of specificity for two correlated ROC curves, a non-parametric method based on Mann-Whitney U-statistics has been developed. The estimation of AUC along with its estimated variance and covariance is simplified by a method of grouping the observations according to their cutpoint values. The method is used to evaluate alternative logistic regression models that predict whether a subject has incident breast cancer based on information in Medicare claims data.

Original languageEnglish (US)
Pages (from-to)701-715
Number of pages15
JournalStatistics in Medicine
Volume21
Issue number5
DOIs
StatePublished - Mar 15 2002

Fingerprint

Nonparametric Methods
Receiver Operating Characteristic Curve
ROC Curve
Healthcare
Delivery of Health Care
Partial
Area Under Curve
Specificity
Logistic Models
Curve
Diagnostic Tests
U-statistics
Logistic Regression Model
Medicare
Breast Cancer
Routine Diagnostic Tests
Grouping
Datasets
Research Personnel
Entire

Keywords

  • Area under ROC curve
  • Breast cancer
  • Medicare claims
  • Non-parametric method
  • Portions of two ROC curves

ASJC Scopus subject areas

  • Epidemiology

Cite this

A non-parametric method for the comparison of partial areas under ROC curves and its application to large health care data sets. / Zhang, Dong; Zhou, Xia Hua; Freeman, Daniel H.; Freeman, Jean L.

In: Statistics in Medicine, Vol. 21, No. 5, 15.03.2002, p. 701-715.

Research output: Contribution to journalArticle

Zhang, Dong ; Zhou, Xia Hua ; Freeman, Daniel H. ; Freeman, Jean L. / A non-parametric method for the comparison of partial areas under ROC curves and its application to large health care data sets. In: Statistics in Medicine. 2002 ; Vol. 21, No. 5. pp. 701-715.
@article{0460231e1faa4210aa6b67216dfd49c3,
title = "A non-parametric method for the comparison of partial areas under ROC curves and its application to large health care data sets",
abstract = "The receiver operating characteristic (ROC) curve is a statistical tool for evaluating the accuracy of diagnostic tests. Investigators often compare the validity of two tests based on the estimated areas under the respective ROC curves. However, the traditional way of comparing entire areas under two ROC curves is not sensitive when two ROC curves cross each other. Also, there are some cutpoints on the ROC curves that are not considered in practice because their corresponding sensitivities or specificities are unacceptable. For the purpose of comparing the partial area under the curve (AUC) within a specific range of specificity for two correlated ROC curves, a non-parametric method based on Mann-Whitney U-statistics has been developed. The estimation of AUC along with its estimated variance and covariance is simplified by a method of grouping the observations according to their cutpoint values. The method is used to evaluate alternative logistic regression models that predict whether a subject has incident breast cancer based on information in Medicare claims data.",
keywords = "Area under ROC curve, Breast cancer, Medicare claims, Non-parametric method, Portions of two ROC curves",
author = "Dong Zhang and Zhou, {Xia Hua} and Freeman, {Daniel H.} and Freeman, {Jean L.}",
year = "2002",
month = "3",
day = "15",
doi = "10.1002/sim.1011",
language = "English (US)",
volume = "21",
pages = "701--715",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "John Wiley and Sons Ltd",
number = "5",

}

TY - JOUR

T1 - A non-parametric method for the comparison of partial areas under ROC curves and its application to large health care data sets

AU - Zhang, Dong

AU - Zhou, Xia Hua

AU - Freeman, Daniel H.

AU - Freeman, Jean L.

PY - 2002/3/15

Y1 - 2002/3/15

N2 - The receiver operating characteristic (ROC) curve is a statistical tool for evaluating the accuracy of diagnostic tests. Investigators often compare the validity of two tests based on the estimated areas under the respective ROC curves. However, the traditional way of comparing entire areas under two ROC curves is not sensitive when two ROC curves cross each other. Also, there are some cutpoints on the ROC curves that are not considered in practice because their corresponding sensitivities or specificities are unacceptable. For the purpose of comparing the partial area under the curve (AUC) within a specific range of specificity for two correlated ROC curves, a non-parametric method based on Mann-Whitney U-statistics has been developed. The estimation of AUC along with its estimated variance and covariance is simplified by a method of grouping the observations according to their cutpoint values. The method is used to evaluate alternative logistic regression models that predict whether a subject has incident breast cancer based on information in Medicare claims data.

AB - The receiver operating characteristic (ROC) curve is a statistical tool for evaluating the accuracy of diagnostic tests. Investigators often compare the validity of two tests based on the estimated areas under the respective ROC curves. However, the traditional way of comparing entire areas under two ROC curves is not sensitive when two ROC curves cross each other. Also, there are some cutpoints on the ROC curves that are not considered in practice because their corresponding sensitivities or specificities are unacceptable. For the purpose of comparing the partial area under the curve (AUC) within a specific range of specificity for two correlated ROC curves, a non-parametric method based on Mann-Whitney U-statistics has been developed. The estimation of AUC along with its estimated variance and covariance is simplified by a method of grouping the observations according to their cutpoint values. The method is used to evaluate alternative logistic regression models that predict whether a subject has incident breast cancer based on information in Medicare claims data.

KW - Area under ROC curve

KW - Breast cancer

KW - Medicare claims

KW - Non-parametric method

KW - Portions of two ROC curves

UR - http://www.scopus.com/inward/record.url?scp=0037088346&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0037088346&partnerID=8YFLogxK

U2 - 10.1002/sim.1011

DO - 10.1002/sim.1011

M3 - Article

C2 - 11870811

AN - SCOPUS:0037088346

VL - 21

SP - 701

EP - 715

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 5

ER -