Stata 18 offers several new features in survival analysis. The new commands lasso cox and elasticnet cox can be used to select covariates using lasso or elastic net and fit Cox models to survival-time data. In addition to standard lasso and elastic-net postestimation tools, you can also use stcurve to plot the penalized survivor, failure, hazard, or cumulative hazard function.

Stata 17 introduced stintcox for fitting interval-censored Cox models using novel genuinely semiparametric estimation methods. Stata 18 adds support for time-varying covariates, commonly encountered in practice, to stintcox. In addition to estimation, postestimation features such as prediction and plots of survivor and other functions incorporate time-varying covariates too! StataNow™ extends the support of Cox proportional hazards model for interval-censored data from the univariate case to the multiple-event case. Now you can fit a marginal Cox proportional hazards model to interval-censored multiple-event data using the new stmgintcox command.

Last but not least, you can now check model fit of your survival models more easily by using the new estat gofplot command. The supported features include right-censored and interval-censored data, parametric and semiparametric models, by-group and stratified estimation, and more.

Read more about these features and see them in action:

•  Interval-censored multiple-event Cox model StataNow

•  Lasso for Cox model

•  TVCs with interval-censored Cox model

•  GOF plots for survival models

•  And more

View all the new features in Stata 18.