ROC curve and covariates: extending induced methodology to the non-parametric framework

Continuous diagnostic tests are often used to discriminate between diseased and healthy populations. The receiver operating characteristic (ROC) curve is a widely used tool that provides a graphical visualisation of the effectiveness of such tests. The potential performance of the tests in terms of distinguishing diseased from healthy people may be strongly influenced by covariates, and a variety of regression methods for adjusting ROC curves has been developed. Until now, these methodologies have assumed that covariate effects have parametric forms, but in this paper we extend the induced methodology by allowing for arbitrary non-parametric effects of a continuous covariate. To this end, local polynomial kernel smoothers are used in the estimation procedure. Our method allows for covariate effect not only on the mean, but also on the variance of the diagnostic test. We also present a bootstrap-based method for testing for a significant covariate effect on the ROC curve. To illustrate the method, endocrine data were analysed with the aim of assessing the performance of anthropometry for predicting clusters of cardiovascular risk factors in an adult population in Galicia (NW Spain), duly adjusted for age. The proposed methodology has proved useful for providing age-specific thresholds for anthropometric measures in the Galician community.

[1]  R. Tibshirani,et al.  Generalized additive models for medical research , 1986, Statistical methods in medical research.

[2]  Lex M. Bouter,et al.  Plasma insulin and cardiovascular mortality in non-diabetic European men and women: a meta-analysis of data from eleven prospective studies , 2004, Diabetologia.

[3]  Carl de Boor,et al.  A Practical Guide to Splines , 1978, Applied Mathematical Sciences.

[4]  Remi Rabasa-Lhoret,et al.  Metabolic and body composition factors in subgroups of obesity: what do we know? , 2004, The Journal of clinical endocrinology and metabolism.

[5]  Tianxi Cai,et al.  Model Checking for ROC Regression Analysis , 2007, Biometrics.

[6]  Margaret S. Pepe,et al.  Semiparametric Receiver Operating Characteristic Analysis to Evaluate Biomarkers for Disease , 2002 .

[7]  P. Mirmiran,et al.  Detection of cardiovascular risk factors by anthropometric measures in Tehranian adults: receiver operating characteristic (ROC) curve analysis , 2004, European Journal of Clinical Nutrition.

[8]  Paul H. C. Eilers,et al.  Flexible smoothing with B-splines and penalties , 1996 .

[9]  E. Nadaraya.  On Estimating Regression , 1964 .

[10]  Jaroslaw Harezlak,et al.  Comparison of bandwidth selection methods for kernel smoothing of ROC curves , 2002, Statistics in medicine.

[11]  Wenceslao González-Manteiga,et al.  ROC Curves in Non‐Parametric Location‐Scale Regression Models , 2011 .

[12]  Jianqing Fan,et al.  Local polynomial modelling and its applications , 1994 .

[13]  Carmen Cadarso-Suárez,et al.  Nonparametric estimation of conditional ROC curves: Application to discrimination tasks in computerized detection of early breast cancer , 2008, Comput. Stat. Data Anal..

[14]  Holly Janes,et al.  Adjusting for covariate effects on classification accuracy using the covariate-adjusted receiver operating characteristic curve. , 2009, Biometrika.

[15]  John A. Swets,et al.  Evaluation of diagnostic systems : methods from signal detection theory , 1982 .

[16]  C. Metz.  Basic principles of ROC analysis. , 1978, Seminars in nuclear medicine.

[17]  Chris Lloyd,et al.  Using Smoothed Receiver Operating Characteristic Curves to Summarize and Compare Diagnostic Systems , 1998 .

[18]  R. Tibshirani,et al.  An introduction to the bootstrap , 1993 .

[19]  M. Zweig,et al.  Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine. , 1993, Clinical chemistry.

[20]  J. Shaw,et al.  Global and societal implications of the diabetes epidemic , 2001, Nature.

[21]  Enrique F Schisterman,et al.  Adjusting the generalized ROC curve for covariates , 2004, Statistics in medicine.

[22]  Paul Zimmet,et al.  [A new international diabetes federation worldwide definition of the metabolic syndrome: the rationale and the results]. , 2005, Revista espanola de cardiologia.

[23]  C B Begg,et al.  A General Regression Methodology for ROC Curve Estimation , 1988, Medical decision making : an international journal of the Society for Medical Decision Making.

[24]  K. Zou,et al.  Smooth non-parametric receiver operating characteristic (ROC) curves for continuous diagnostic tests. , 1997, Statistics in medicine.

[25]  David Ruppert,et al.  Local polynomial variance-function estimation , 1997 .

[26]  Jun S. Liu,et al.  Linear Combinations of Multiple Diagnostic Markers , 1993 .

[27]  Liang Peng,et al.  Local linear smoothing of receiver operating characteristic (ROC) curves , 2004 .

[28]  M. Pepe.  The Statistical Evaluation of Medical Tests for Classification and Prediction , 2003 .

[29]  David Faraggi,et al.  Adjusting receiver operating characteristic curves and related indices for covariates , 2003 .

[30]  G A Colditz,et al.  Body fat distribution and risk of non-insulin-dependent diabetes mellitus in women. The Nurses' Health Study. , 1997, American journal of epidemiology.

[31]  B. Turnbull,et al.  NONPARAMETRIC AND SEMIPARAMETRIC ESTIMATION OF THE RECEIVER OPERATING CHARACTERISTIC CURVE , 1996 .

[32]  C. Cadarso-Suárez,et al.  Prevalence of metabolic syndrome in Galicia (NW Spain) on four alternative definitions and association with insulin resistance , 2009 .

[33]  Yingye Zheng,et al.  Semiparametric estimation of time-dependent ROC curves for longitudinal marker data. , 2004, Biostatistics.

[34]  M S Pepe,et al.  Three approaches to regression analysis of receiver operating characteristic curves for continuous test results. , 1998, Biometrics.

[35]  Jianqing Fan,et al.  Efficient Estimation of Conditional Variance Functions in Stochastic Regression , 1998 .

[36]  Jianqing Fan,et al.  Fast Implementations of Nonparametric Curve Estimators , 1994 .

[37]  Xiao-Hua Zhou,et al.  Statistical Methods in Diagnostic Medicine , 2002 .

[38]  Xiao-Hua Zhou,et al.  A Flexible Method for Estimating the ROC Curve , 2004 .

[39]  Margaret Sullivan Pepe,et al.  Combining Several Screening Tests: Optimality of the Risk Score , 2002, Biometrics.

[40]  Tianxi Cai,et al.  Semi-parametric ROC regression analysis with placement values. , 2004, Biostatistics.

[41]  P J Garry,et al.  Fat distribution and blood lipids in a sample of healthy elderly people. , 1992, International journal of obesity and related metabolic disorders : journal of the International Association for the Study of Obesity.

[42]  R L Hayes,et al.  The medical use of gallium radionuclides: a brief history with some comments. , 1978, Seminars in nuclear medicine.

[43]  Margaret Sullivan Pepe,et al.  Distribution-free ROC analysis using binary regression techniques. , 2002, Biostatistics.

[44]  G. S. Watson,et al.  Smooth regression analysis , 1964 .

[45]  Holly Janes,et al.  Adjusting for Covariate Effects on Classification Accuracy Using the Covariate-Adjusted ROC Curve , 2006 .