Performance of Cause-specific and Subdistribution Hazard for Large Samples - A Simulation Study
[1]
Galappaththige Hasani Sandamali Karunarathna, Department of Statistics, University of Colombo, Colombo, Sri Lanka.
[2]
Marina Roshini Sooriyarachchi, Department of Statistics, University of Colombo, Colombo, Sri Lanka.
The competing risks scenario is a complex setting for classical survival analysis when an individual is under risk of failing from various events. Since competing risk data are often found in many fields such as medicine, social science, biology etc., interest has been paid among researchers to focus towards the methodological competing risk setting. Additionally, it is not possible to have real data and thus to know about the real status, thus simulation studies lead to more advantages towards analyzing such responses. Hence, this paper focuses on investigating the performance of the most commonly used regression approaches for analyzing the competing risk responses namely, cause specific hazard model and sub-distribution hazard model by following pre-specified cause specific hazard ratio. A simulation study was carried out by varying the censoring distribution parameter and shape parameter while keeping the scale parameter constant, under nine scenarios. Summary statistics of cause specific hazard and sub-distribution hazard were different for the two methods and it showed that mean hazards of cause-specific hazard model decreases when the shape parameter of the censoring distribution is increased. As a conclusion, this simulation study reveals that cause specific and sub-distribution hazard ratios are monotonically increasing with all scenarios and all scenarios performed approximately equally with minor differences for the two types of regression models.
Competing Risks, Cause-Speccific Hazard, Sub-distribution Hazard
[1]
B. Haller, "The Analysis of Competing Risks Data with a Focus on Estimation of Cause-Specific and Subdistribution Hazard Ratios from a Mixture Model," 2014.
[2]
H. Putter, M. Fiocco and R. B. Geskus, "Tutorial in biostatistics: competing risks and multi state models," Stat Med, vol. 26, no. 11, pp. 2389-2430, 2007.
[3]
P. K. Anderson and N. Keiding, "Interpretability and importance of functional in competing risk and multistate models," Stat Med, vol. 31, pp. 1074-1088, 2012.
[4]
J. D. Kalbfleisch and R. L. Prentice, The Statistical Analysis of Failure Time Data (2nd ed.), New York: Wiley, 2002.
[5]
P. Andersen, S. Abildstrom and S. Rosthoj, "Competing risks as a multi state model," Statisical methods in Medical Research, vol. 11, no. 2, pp. 203-215, 2002.
[6]
R. Prentice, J. Kalbfleish, A. Peterson, N. Flournoy, V. Farewell and N. Breslow, "The Analysis of Failure Times in the presense of Competing Risks," Biometrics, vol. 34, pp. 541-554, 1978.
[7]
J. P. Fine and R. J. Gray, "A Proportional hazard model for the subdistribution of a competing risk," Journal of the American Statistical Association, pp. 496-509, 1999.
[8]
J. Beyersmann, A. Latouche, A. Buchholz and M. Schumacher, "Simulating Competing risks data in survival analysis," Statistics in Medicine, pp. 956-971, 2009.
[9]
G. H. S. Karunarathna and M. R. Sooriyarachchi, "Investigating Hospital Discharge and Mortality: Contribution of Competing Risk Regression Model," 2017.
[10]
J. J. Dignam, Q. Zhang and M. N. Kocherginsky, "The Use and Interpretion of Competing Risks Regression Models," 2012.
[11]
H. T. Kim, "Cumulative incidence in competing risk data and competing risk regression analysis," Clin Cancer Res, vol. 13, pp. 559-565, 2007.
[12]
B. Tai, R. Grundy and D. Machin, "On the importance of accounting for competing risks in pediatric brain cancer: II. Regression modeling and sample size," Radiat Oncol Biol Phys, vol. 79, pp. 1139-1146, 2011.
[13]
P. C. Austin and J. P. Fine, "Practical Recommendation for reporting Fine-Gray model analyses for competing risk data," Statistics in Medicine, vol. 36, pp. 4391-4400, 2017.
[14]
B. Haller, G. Schmidt and K. Ulm, "Applying competing risk regression models: An overview," Springer Science Business media, 2012.
[15]
A. Burton, D. Altman, P. Royston and R. Holder, "The design of simulation studies in medical statistics," Statistics in Medicine, vol. 25, pp. 4279-4292, 2006.
[16]
R. Bender, T. Augustin and M. Blettner, "Generating survival times to simulate cox proportional hazard models," Statistics in Medicine, vol. 24, pp. 1713-1723, 2005.
[17]
R. J. Madachy and D. Houston, What every engnieer should know about Modeling and Simulation, CRC Press, 2017.
[18]
J. A. Bayersmann and M. Schumacher, "Competing Risks and Multistate Models with R," Springer, 2011.
[19]
I. Poguntke, M. Schumacher, J. Bayersmann and M. Wolkewitz, "Simulation shows undesirable results for competing risks analysis with time-dependent covariates for clinical outcomes," BMC Medical Research Methodologyvolume, vol. 18, 2018.
