Confidence intervals for the Youden index and corresponding optimal cut-point

Enrique F. Schisterman, Neil Perkins

Research output: Contribution to journalArticle

Abstract

A global measure of biomarker effectiveness is the Youden index, the maximum difference between sensitivity, the probability of correctly classifying diseased individuals, and 1-specificity, the probability of incorrectly classifying healthy individuals. The cut-point leading to the index is the optimal cut-point when equal weight is given to sensitivity and specificity. Using the delta method, we present approaches for estimating confidence intervals for the Youden index and corresponding optimal cut-point for normally distributed biomarkers and also those following gamma distributions. We also provide confidence intervals using various bootstrapping methods. A comparison of interval width and coverage probability is conducted through simulation over a variety of parametric situations. Confidence intervals via delta method are shown to have both closer to nominal coverage and shorter interval widths than confidence intervals from the bootstrapping methods.

Original languageEnglish (US)
Pages (from-to)549-563
Number of pages15
JournalCommunications in Statistics Part B: Simulation and Computation
Volume36
Issue number3
DOIs
StatePublished - May 2007
Externally publishedYes

Fingerprint

Confidence interval
Biomarkers
Delta Method
Bootstrapping
Specificity
Interval Methods
Short Intervals
Coverage Probability
Gamma distribution
Categorical or nominal
Coverage
Interval
Simulation

Keywords

  • Confidence intervals
  • Optimal cut-point
  • ROC curve
  • Sensitivity and specificity
  • Youden index

ASJC Scopus subject areas

  • Modeling and Simulation
  • Statistics and Probability

Cite this

Confidence intervals for the Youden index and corresponding optimal cut-point. / Schisterman, Enrique F.; Perkins, Neil.

In: Communications in Statistics Part B: Simulation and Computation, Vol. 36, No. 3, 05.2007, p. 549-563.

Research output: Contribution to journalArticle

@article{777ef10e16dc4295b760c59e3cce8699,
title = "Confidence intervals for the Youden index and corresponding optimal cut-point",
abstract = "A global measure of biomarker effectiveness is the Youden index, the maximum difference between sensitivity, the probability of correctly classifying diseased individuals, and 1-specificity, the probability of incorrectly classifying healthy individuals. The cut-point leading to the index is the optimal cut-point when equal weight is given to sensitivity and specificity. Using the delta method, we present approaches for estimating confidence intervals for the Youden index and corresponding optimal cut-point for normally distributed biomarkers and also those following gamma distributions. We also provide confidence intervals using various bootstrapping methods. A comparison of interval width and coverage probability is conducted through simulation over a variety of parametric situations. Confidence intervals via delta method are shown to have both closer to nominal coverage and shorter interval widths than confidence intervals from the bootstrapping methods.",
keywords = "Confidence intervals, Optimal cut-point, ROC curve, Sensitivity and specificity, Youden index",
author = "Schisterman, {Enrique F.} and Neil Perkins",
year = "2007",
month = "5",
doi = "10.1080/03610910701212181",
language = "English (US)",
volume = "36",
pages = "549--563",
journal = "Communications in Statistics Part B: Simulation and Computation",
issn = "0361-0918",
publisher = "Taylor and Francis Ltd.",
number = "3",

}

TY - JOUR

T1 - Confidence intervals for the Youden index and corresponding optimal cut-point

AU - Schisterman, Enrique F.

AU - Perkins, Neil

PY - 2007/5

Y1 - 2007/5

N2 - A global measure of biomarker effectiveness is the Youden index, the maximum difference between sensitivity, the probability of correctly classifying diseased individuals, and 1-specificity, the probability of incorrectly classifying healthy individuals. The cut-point leading to the index is the optimal cut-point when equal weight is given to sensitivity and specificity. Using the delta method, we present approaches for estimating confidence intervals for the Youden index and corresponding optimal cut-point for normally distributed biomarkers and also those following gamma distributions. We also provide confidence intervals using various bootstrapping methods. A comparison of interval width and coverage probability is conducted through simulation over a variety of parametric situations. Confidence intervals via delta method are shown to have both closer to nominal coverage and shorter interval widths than confidence intervals from the bootstrapping methods.

AB - A global measure of biomarker effectiveness is the Youden index, the maximum difference between sensitivity, the probability of correctly classifying diseased individuals, and 1-specificity, the probability of incorrectly classifying healthy individuals. The cut-point leading to the index is the optimal cut-point when equal weight is given to sensitivity and specificity. Using the delta method, we present approaches for estimating confidence intervals for the Youden index and corresponding optimal cut-point for normally distributed biomarkers and also those following gamma distributions. We also provide confidence intervals using various bootstrapping methods. A comparison of interval width and coverage probability is conducted through simulation over a variety of parametric situations. Confidence intervals via delta method are shown to have both closer to nominal coverage and shorter interval widths than confidence intervals from the bootstrapping methods.

KW - Confidence intervals

KW - Optimal cut-point

KW - ROC curve

KW - Sensitivity and specificity

KW - Youden index

UR - http://www.scopus.com/inward/record.url?scp=34249749955&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34249749955&partnerID=8YFLogxK

U2 - 10.1080/03610910701212181

DO - 10.1080/03610910701212181

M3 - Article

VL - 36

SP - 549

EP - 563

JO - Communications in Statistics Part B: Simulation and Computation

JF - Communications in Statistics Part B: Simulation and Computation

SN - 0361-0918

IS - 3

ER -