Author:
Salazar Adriana Marcela,Huertas Jaime
Abstract
The survival competing risks model in discrete time based on multinomial logistic regression, proposed by Luo et al. (2016), models the non-linear and irregular shape of hazard functions by incorporating a time-dependent spline into the multinomial logistic regression. This model also directly includes longitudinal variables in the regression. Due to the issues arising from including both baseline and longitudinal covariates in the extended form as proposed, and considering that the latter may be subject to error, this article suggests an extension of the existing model. The proposed extension utilizes the concept of joint models for longitudinal and survival data, which is an effective approach for integrating simultaneousness both baseline and time-dependent covariates into the survival model.,
Publisher
Universidad Nacional de Colombia
Subject
Statistics and Probability
Reference30 articles.
1. Annesi, I., Moreau, T. & Lellouch, J. (1989), ‘Efficiency of the logistic regression and Cox proportional hazards models in longitudinal studies’, Statistics in Medicine 8, 1515–1521.
2. Begg, C. B. & Gray, R. (1984), ‘Calculation of polychotomous logistic regression parameters using individualized regressions’, Biometrika 71(1), 11–18.
3. Bowman, A. W. & Azzalini, A. (1997), Applied Smoothing Techniques for Data Analysis. The Kernel Approach with S-Plus Illustrations, 1 edn, New York, Oxford University Press Inc.
4. Cox, D. R. (1972), ‘Regression models and life tables (with discussion)’, Journal of the Royal Statistical Society, Series B 34, 187–220.
5. Elandt-Johnson, R. C. (1980), Time dependent logistic models in follow-up studies and clinical trials, I. Binary data, Technical report, Institute of Statistics Mimeo Series No 1309, University of North Carolina.