- Examples updated
- dutility function no longer casts warnings when handling formulas
- Efficient estimation of recurrent events mean: recurrentMarginalAIPCW
- Average treatment effect for competing risks and binary data: logitATE, binregATE
- Recurrent events regression with IPCW adjustment : recreg

- Maintenance release

- Augmentation of binomial regression model: BinAugmentCifstrata
- Augmentation of Fine-Gray model: FG_AugmentCifstrata
- Double Fine-Gray model for two causes.
- Likelihood evaluation of mvn uses Moore-Penrose pseudo-inverese (threshold set via lava.options(itol=…)
- Vignette updates

- Maintenance release

- Cumulative incidence regression cifreg function
- Fine-Gray model with cloglog link of (1-F_1(t,x))
- Logit link

- Prototype of wildbootstrap for Cox regression with
- confidence bands for baseline
- with confidenence bands cumulative incidence for two cox’s

- Piecewise constant hazard: rpch, ppch
- Test-version of multinomial regression model (via phreg): mlogit
- Simulation for illness-death model: simMultistate
- Haplotype modelling for discrete time-to-pregnancy models: haplo.surv.discrete
- Interval censoring for discrete time logit-survival model: interval.logitsurv.discrete
- Binomial Regression for competing risks data with censoring and one time point only: binreg

- Updated predict function for phreg
- with plotting functionality
- with robust standard errors

- New vignettes started
- phreg robust se’s for marginal Cox model
- twostage survival model
- multivariate competing risks
- recurrent events

- logitSurv for fitting semiparametric proportional odds model
- gof
- robust standard errors for clustered case

- twostageMLE for fast twostage fitting for clustered survival data with robust standard errors.
- standard errors for twostage models now also with uncertainty from Cox baseline
- cumulative score process test gof now also for marginal Cox models

- functions km (Kaplan-Meier) and cif (cumulative incidence probability) with robust standard errors.
- computation of probability of exceeding “k” events for recurrents processs
- computation of probability of exceeding “k1” and “k2” events for bivariate recurrents processseses
- dspline simple spline decomposition on a data frame
- rmvn, dmvn: RNG and density for multivariate normal distribution with varying correlation coefficients.

- starting values updated for twinlm method

- twinlm now supports ordinal outcomes
- optimized strata calculations in phreg
- optimized robust standard errors in phreg
- weights and offsets in phreg
- weights argument added to lifetable
- gof of phreg with fast cumulative residuals (Lin, Wei, Ying)
- graphical gof of phreg
- recurrent events function for marginal mean with standard errors
- simulating recurrent events with possibly two recurrent events and death
- covariance calculation for recurrent events data and related bootstrap

- Vignettes updated
- Compatibility with lava version 1.5

- New documentation/vignettes
- Additional examples and unit tests
- lifecourse plot function: lifecourse
- block sampling function: dsample

- Namespace cleaning (twostage)…
- Dependency on R>=3.3 radix algorithm
- Case-Control sampling for twostage model.
- Two-stage additive gamma survival model. Additive random effects for two-stage survival model via pairwise composite likelihood. Simulation of family ace survival model. Function for computing Kendall’s tau for pairs with additive gamma random effects model via simulations.
- Two-stage additive gamma binomial model. Additive random effects for binomial model via pairwise composite likelihood. Simulation of family ace model. Function for computing pairwise concordance for for pairs with additive gamma random effects model.
- Updated divide.conquer
- Extra unit tests
- force.same.cens argument with IPWC methods
- New utility functions for data.frames Data processing
- dsort
- dreshape
- dcut
- drm, drename, ddrop, dkeep, dsubset
- drelevel
- dlag
- dfactor, dnumeric Data aggregation
- dby, dby2
- dscalar, deval, daggregate
- dmean, dsd, dsum, dquantile, dcor
- dtable, dcount Data summaries
- dhead, dtail,
- dsummary,
- dprint, dlist, dlevels, dunique

- Support for left-truncation

- fast.approx with ‘type’ argument
- scoreMV
- lifetable updated and new survpois function (piecewise constant hazard)

- New functions biprobit.time, binomial.twostage.time. Automatically samples time points (approximately equidistant) up to last double jump time. Intial support for left truncation. contrast argument added to biprobit.time.
- ipw removed (from namespace)
- biprobit optimized for tabular data (non-continuous covariates). Regression design for dependence parameter (tetrachoric correlation) now possible.
- predict method implemented for biprobit
- arc-sinus transformation used for probability estimates
- updated output of bptwin with relative recurrence risk + log-OR estimates
- iid method for bptwin (influence function)
- survival probabilities and start and end of intervals added to lifetable
- new function ‘jumptimes’ for extracting jump times and possibly sample (equidistant)
- fast.pattern updated to handle more than two categories
- demos added to the mets package
- divide.conquer function, folds function

- Normal orthant probabilities via ‘pmvn’ (vectorized)
- Parametric proportional hazards models via ‘phreg.par’
- twinlm.time function for censored twin data. Wraps the ‘ipw’ function that now also supports parametric survival models via phreg.par. ‘grouptable’ for tabulating twin-data.
- Relative recurrence risk ratios now reported with bptwin/twinlm.
- Grandom.cif more stable

- Adapted to changes in ‘timereg::comp.risk’
- cluster.index with ‘mat’ argument for stacking rows of a matrix according to cluster-variable
- New lava-estimator: ‘normal’, for ordinal data (cumulative probit)
- fast.reshape more robust. Now also supports ‘varying arguments of the type ’varying=-c(…)’ choosing everything except ‘…’.

- C++ source code cleanup
- Optimization of fast.reshape

- New datasets: dermalridges, dermalridgesMZ
- Grouped analysis updated in twinlm (e.g. sex limitation model)
- Confidence limits for genetic and environmental effects are now based on standard (symmetric) Wald confidence limits. (use the ‘transform’ argument of the summary method to apply logit-transform)
- Improved output in twinlm

- fast.reshape :labelnum option for both wide and long format (see example)
- Compilation flags removed from Makevars files

- fast.reshape bug-fix (column names)

- Updated twinlm. bptwin: OS analysis
- Better starting values for twinlm

- Fixed claytonaokes.cpp
- New fast cox ph regression: phreg
- Updated two-stage estimator
- Improved fast.reshape

- fast.reshape
- easy.binomial.twostage

- Fixed cor.cpp
- New datasets: twinstut, twinbmi, prtsim

- twinlm moved to mets package, and wraps the bptwin function

- code clean-up and minor bug-fixes

- Random effects CIF models moved from MultiComp to mets
- new data sets: np, multcif
- Documentation via roxygen2
- bug fixes

- Initialization of the new package ‘mets’ with implementation of the Clayton-Oakes model with piecewise constant marginal hazards, and the bivariate probit random effects model (Liability model) for twin-data.