Use of multi-state models for time-to-event data
Background & Aim: In medical sciences, the outcome is the time until the occurrence of an event of interest. A multi-state model (MSM) is used to model a process where subjects’ transition takes place from one state to the next. For instance, a standard survival curve can be thought of as a simple MSM with two states (alive and dead) and the transition between these two-state models is a method used to analyze time to event data. The most important aspect of this model is that it considers intermediate events and models the effect of covariates on each transition intensity. Some diseases like cancer, human immunodeficiency virus (HIV), etc. have several stages. In the present study, these models were reviewed using cardiac allograft vasculopathy (CAV) data focusing on different approaches.
Methods & Materials: The data of 576 CAVs were collected. A time dependent simple Cox regression model (CRM) was fitted and a three-state illness-death model was considered for the MSMs.
Results: In the simple CRM, only the individuals with the age of > 50 were significant, however for Cox Markov model (CMM) and Cox semi-Markov model (CSMM), the donor’s age > 40, sex, and the individuals with the age of > 50 were other significant covariates.
Conclusion: The CMM and CSMM showed more accurate results about risk factors compared to the simple CRM.
Andersen PK, Esbjerg S, Sorensen TI. Multi-state models for bleeding episodes and mortality in liver cirrhosis. Stat Med 2000; 19(4): 587-99.
Abner EL, Charnigo RJ, Kryscio RJ. Markov chains and semi-Markov models in time-to-event analysis. J Biom Biostat 2013; Suppl 1(e001): 19522.
Commenges D. Inference for multi-state models from interval-censored data. Stat Methods Med Res 2002; 11(2): 167-82.
Meira-Machado L, de Una-Alvarez J, Cadarso-Suarez C, Andersen PK. Multi-state models for the analysis of time-to-event data. Stat Methods Med Res 2009; 18(2): 195-222.
Kay R. A Markov model for analyzing cancer markers and disease states in survival studies. Biometrics 1986; 42(4): 855-65.
Cox DR. Regression models and life-tables. J R Stat Soc Series B 1972; 34(2): 187-220.
Meira-Machado L, Roca-Pardinas R. p3state.msm: Analyzing survival data from an. J Stat Software; 2011; 38(3): 1-18.
Sharples LD, Jackson CH, Parameshwar J,94 http://jbe.tums.ac.ir
Wallwork J, Large SR. Diagnostic accuracy of coronary angiography and risk factors for post-heart-transplant cardiac allograft vasculopathy. Transplantation 2003; 76(4): 679-82.
Jackson C. Multi-state modelling with R: The msm package. J Stat Softw 2011; 38(8): 1-29.
Fisher LD, Lin DY. Time-dependent covariates in the Cox proportional-hazards regression model. Annu Rev Public Health 1999; 20: 145-57.
Heutte N, Huber-Carol C. Semi-markov models for quality of life data with censoring. In: Mesbah M, Cole BF, Lee MT, Editors.Statistical methods for quality of life studies. Berlin, Germany: Springer Science & Business Media, 2002. p. 207-18.
Shu Y, Klein JP, Zhang MJ. Asymptotic theory for the Cox semi-Markov illness-death model. Lifetime Data Anal 2007; 13(1): 91-117.
Titman AC. Transition probability estimates for non-Markov multi-state models. Biometrics 2015; 71(4): 1034-41.
Chang IS, Chuang YC, Hsiung C. Goodness-of-fit tests for semi-markov and markov survival models with one intermediate state. Scand J Stat 2001; 28(3): 505-25.
|Issue||Vol 3 No 3/4 (2017)|
|Markov chains Survival analysis Risk factors Proportional hazards models Disease progression|
|Rights and permissions|
|This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.|