In many clinical studies involving event history analysis, the event of interest is non-fatal and may occur more than once for each subject. Models based on the theory of counting processes have been developed to deal with such data, the recurrences being considered as transitions in a Markovian process. Under this setting, the experimental units can move between states over time, and it is possible to estimate the corresponding transition probabilities employing regression models that incorporate the influence of covariates. Despite of this, most of the softwares are concerned only in the estimation of regression parameters and do not provide transition probabilities estimates. The aim of this paper is to present a SAS macro developed to estimate the transition probabilities, considering three approaches for the regression modeling. The macro is flexible enough to allow the user to select the model to be fit providing, for a given set of covariates, plots of the estimates for the predicted transition probabilities as a function of time.
[1]
P. Grambsch,et al.
Modeling Survival Data: Extending the Cox Model
,
2000
.
[2]
A. V. Peterson,et al.
On the regression analysis of multivariate failure time data
,
1981
.
[3]
L. J. Wei,et al.
Regression analysis of multivariate incomplete failure time data by modeling marginal distributions
,
1989
.
[4]
Ren Johansen.
An Empirical Transition Matrix for Non-homogeneous Markov Chains Based on Censored Observations
,
1978
.
[5]
David R. Cox,et al.
Regression models and life tables (with discussion
,
1972
.
[6]
Niels Keiding,et al.
Non- and semi-parametric estimation of transition probabilities from censored observation of a non-homogeneous Markov process
,
1991
.
[7]
R. Gill,et al.
Cox's regression model for counting processes: a large sample study : (preprint)
,
1982
.