**** 2011/06/27 - version 0.2 **** This is the first version available for public download. We wait your comments for futur improvements. **** 2011/12/05 - version 0.3 **** This version takes into account the new estimator of the net survival proposed by Pohar-Perme et al. Moreover, the multi-core computing is abandoned: the stability of this algorithm is very different depending on the computer and the operating system. **** 2012/05/31 - version 0.4 **** This version offers a correction of a bug during the installation. We expect that this bug was due to the same names between the ROCt() function and the ROCt package. In this new version, the name of the function in order to compute the time-dependent ROC curve is allcause.ROCt(). **** 2012/07/02 - version 0.5 **** The changes concern the function relative.ROCt() only. When the expected mortality (from lifetime tables) is closed or even higher than the observed mortality, the net survival can be higher than 1 (package relsurv). In this situation, we have constrained the higher bound of the net survival at 1 in the computation of the ROC curve. If this situation occurs, a warning message is given in the output. We also have changed the name of the function with net.ROCt(). **** 2012/07/02 - version 0.6 **** The change concern the addition of a lifetime table. This table is an object of the class ratetable which contains the expected mortality of French patients with End Stage Kidney Disease (ESKD) in dialysis and registered previously on waiting list for renal transplantation. **** 2013/08/01 - version 0.7 **** This version is compatible with R > 3.0. **** 2013/08/02 - version 0.7.1 **** An error was corrected in the usage associated with the object "rein.ratetable". **** 2013/08/02 - version 0.8 **** We have completed the bibliography and we have reduced the dimensions of the databases and the cut-off values in order to reduce the time necessary for the example computations (CRAN policies). **** 2014/28/06 - version 0.9 **** In this version, we have added a new function in order to estimate optimal cut-off for medical decision making. This estimation is based on expected utility maximisation. We also have added the naive estimator of the time-dependent ROC curve for the all-cause analysis. **** 2015/15/05 - version 0.9.1 **** In this version, we have changed the function EUt as follows: i) the argument "gain" may be a vector, ii) the additional argument "reference" in order to identify the group of patients for which the standard care will be proposed, and iii) the argument "full.efficacy" was removed because the user may specify an important gain in order to reach the full efficacy. **** 2015/06/23 - version 0.9.2 **** Two additional functions, entitled adjusted.ROC and adjusted.ROCt, have been included. These functions respectively propose confounder-adjusted estimators of ROC and time-dependent ROC curves. For that purpose, the Inverse Probability Weighting (IPW) approach was used. For the confounder-adjusted ROC curve (without censoring), we also proposed the implementation of the estimator based on placement values (Pepe and Cai, Biometrics, 2004). A simple function, entitled AUC, has also been added to compute the area under the ROC curve from both sensitivities and specificities vectors. The function allcause.ROCt() was renamed in crude.ROCt(). **** 2015/11/04 - version 0.9.3 **** In this version, we have removed the function EUt, which was divided into two different functions: i) EUt1 when only one treatment is observed and assumptions are required for the alternative treatment, and ii) EUt2 when the two treatments are observed. **** 2016/03/09 - version 0.9.4 **** In this version, we just have proposed minor revisions in order to respect the CRAN policy.