Extension of the cox proportional hazards model for time dependent variables is an alternative method used if the PH assumption is not satisfied for one or more of the predictors in the model. The step of this method are testing the PH assumption, interacting the time independent variable not satisfying the PH assumption with function of time (linear, logarithm and the combination), testing the parameters using the Likelihood Ratio test and the Wald test, determine the better model of both models by the AIC values, calculating the hazard ratio of each variable that significantly affected ASD. Based on the smallest AIC values, a better model is the extension of the cox proportional hazards model for time dependent variables interacted with logarithm time function. From the better model obtained the variables that affect the recurrence time ASD is patient with heart beat, noisy heart, blood pressure, body mass index, and treatment with catheter.