Changes in version: JM_0.8-0

  * the new argument 'interFact' added in jointModel() allows the specification of interaction terms between the longitudinal outcome and baseline covariates.

  * for all joint models fitted in JM there is now the option to use a pseudo adaptive Gauss-Hermite rule. This is much faster than the default option and produces results of equal or better quality.

  * a predict() method has been added. Currently this only calculates fitted average longitudinal evolutions based on the information provided in the 'newdata' argument.

  * a new algorithm for calculating the starting values has been implemented. In most of the cases these will be closer to the MLEs than in the previous version.

  * some small changes have been made in the default Gauss-Hermite quadrature rule. This will result in minor changes in parameter estimates, standard errors and log-likelihood value compared to the previous version.

  * a bug has been corrected in the code used to specify the design matrix for the random effects in the longitudinal outcome, that did not allow this matrix not to be a subset of the design matrix of the fixed effects.


==============================

Changes in version: JM_0.7-0

  * the new function rocJM() has been added that calculates time-dependent ROC curves and the corresponding AUCs for joint models.

  * methods "weibull-AFT-GH", "weibull-PH-GH", "piecewise-PH-GH", and "spline-PH-GH" support now the true slope parameterization. This is invoked be specifying the 'parameterization' and 'derivForm' arguments accordingly.


==============================

Changes in version: JM_0.6-2

  * a small bug was corrected in summary.jointModel().


==============================

Changes in version: JM_0.6-1

  * jointModel() has now the extra argument 'scaleWB' that allows to fix the scale parameter for the Weibull baseline hazard to a specific value.


==============================

Changes in version: JM_0.6-0

  * method = "spline-PH-GH" allows now to include stratification factors for which different spline coefficients are estimated. By default the knots positions are the same across strata -- this can be changed by either directly specifying the knots or by setting the control argument 'equal.strata.knots' to FALSE.

  * the new function wald.strata() can be used to test for equality of the spline coefficients among strata.

  * a confint() method has been introduced for 'jointModel' objects.

  * jointModel() has now the extra argument 'lag' that allows for lagged effects in the time-dependent covariate represented by the linear mixed model.

  * a bug was corrected in joint models with piecewise constant baseline risk function. In particular, the 'xi' parameters were reported as double their actual value.


==============================

Changes in version: JM_0.5-0

  * function dynC() has been added that calculates a dynamic concordance index for joint models.

  * method = "ch-GH" has been replaced by method = "spline-PH-GH" that fits a relative risk model with a B-spline-approximated baseline risk function.

  * method = "ph-GH" that fits a relative risk with an unspecified baseline risk function has been renamed to method = "Cox-PH-GH".


==============================

Changes in version: JM_0.4-0

  * function survfitJM() has been added that calculates predictions of subject-specific probabilities of survival given a history of longitudinal responses.

  * the multiple-imputation residuals now work also for joint models with piecewise constant baseline risk functions.

  * faster optimization algorithms have implemented for 'method = "weibull-PH-GH"' and 'method = "piecewise-PH-GH".


==============================

Changes in version: JM_0.3-0

  * the Weibull model is now available under both the relative risk and accelerated failure time parameterizations.

  * a number of enhancements have been implemented in the functions that compute the MI-based residuals.

  * new more robust algorithms have been written for the numerical approximation of integrals; this will lead to some discrepancies in the results, especially in the survival part, compared to the previous versions of the package.


==============================

Changes in version: JM_0.2-1

  * changes in e-mail addresses.


==============================

Changes in version: JM_0.2-0

  * the jointModel method for the residuals generic has further options: (i) MI residuals for fixed and random visit times for the longitudinal process, and (ii) martingale, Cox-Snell, and AFT residuals for the survival process.

  * Function weibull.frailty() is introduced (along with supporting methods) for fitting multivariate survival data using the Weibull model with Gamma multiplicative frailties under maximum likelihood.

  * several typos have been corrected in .Rd files.


==============================

Changes in version: JM_0.1-1

  * corrected some typos in .Rd files.

