A threshold‐free summary index of prediction accuracy for censored time to event data

Statistics in Medicine - Tập 37 Số 10 - Trang 1671-1681 - 2018
Yan Yuan1, Qian M. Zhou2,3, Bingying Li3, Hengrui Cai1, Eric J. Chow4, Gregory T. Armstrong5
1School of Public Health, University of Alberta, Edmonton, AB T6G1C9 Canada
2Department of Mathematics and Statistics, Mississippi State University, Starkville, Mississippi 39762 USA
3Department of Statistics and Actuarial Science, Simon Fraser University, Burnaby, B.C. V5A1S6 Canada
4Fred Hutchinson Cancer Research Center, Seattle Children's Hospital, University of Washington, Seattle, Washington, USA
5Department of Epidemiology and Cancer Control, St. Jude Children's Research Hospital, 262 Danny Thomas Place, MS 735,, Memphis, TN 38105 USA

Tóm tắt

Prediction performance of a risk scoring system needs to be carefully assessed before its adoption in clinical practice. Clinical preventive care often uses risk scores toscreenasymptomatic population. The primary clinical interest is to predict the risk of having an event by a prespecifiedfuturetimet0. Accuracy measures such as positive predictive values have been recommended for evaluating the predictive performance. However, for commonly used continuous or ordinal risk score systems, these measures require a subjective cutoff threshold value that dichotomizes the risk scores. The need for a cutoff value created barriers for practitioners and researchers. In this paper, we propose a threshold‐free summary index of positive predictive values that accommodates time‐dependent event status and competing risks. We develop a nonparametric estimator and provide an inference procedure for comparing this summary measure between 2 risk scores for censored time to event data. We conduct a simulation study to examine the finite‐sample performance of the proposed estimation and inference procedures. Lastly, we illustrate the use of this measure on a real data example, comparing 2 risk score systems for predicting heart failure in childhood cancer survivors.

Từ khóa


Tài liệu tham khảo

Wright CF, 2014, Conceptual issues for screening in the genomic era‐time for an update?, Epidemiol Biostat Public Health, 11, e9944

10.2337/dc07-0048

10.1016/S0140-6736(09)62004-3

10.1016/j.jacc.2013.11.005

10.1093/epirev/mxq019

10.1200/JCO.2014.56.1373

10.2337/diacare.26.3.725

10.1002/sim.2995

10.1016/S0140-6736(02)07948-5

10.1161/CIRCULATIONAHA.106.672402

10.1093/biostatistics/5.1.113

10.1198/016214507000001481

10.1177/0969141313517497

10.1111/j.1541-0420.2009.01246.x

10.1145/65943.65945

Manning C. D, 1999, Foundations of Statistical Natural Language Processing

DavisJ GoadrichM.The relationship between precision‐recall and roc curves. In: Proceedings of the 23rd International Conference on Machine Learning ICML'06 ACM;2006;New York NY USA:233‐240.

SuW YuanY ZhuM.A relationship between the average precision and the area under the ROC curve. In: Proceedings of the 2015 International Conference on the Theory of Information Retrieval ACM.New York NY USA;2015:349‐352.

10.3389/fpubh.2015.00057

10.1016/j.jclinepi.2015.02.010

10.1200/JCO.2009.22.3339

10.1111/j.0006-341X.2000.00337.x

10.1198/016214507000000149

10.1002/sim.3758

Pepe MS, 2003, The Statistical Evaluation of Medical Tests for Classification and Prediction, 10.1093/oso/9780198509844.001.0001

10.1214/aos/1176344552

10.1093/jnci/djn310

10.1111/j.1541-0420.2010.01456.x

10.1002/sim.5958

10.1056/NEJMsa060185

10.1093/biostatistics/kxi005

10.1111/j.0006-341X.2002.00657.x

10.1080/01621459.1971.10482346

10.1177/0272989X06295361

10.1186/1472-6947-11-45

Thông tin
Thông tin xuất bản

Statistics in Medicine

Tập 37 Số 10

1671-1681

Thông tin tác giả