{stdmest}: Post-Estimation Predictions for Standardised Hierarchical Contrasts after -mestreg- Models in Stata

This is a prototype package to implement regression standardisation for (fair) comparison between hierarchical units in multilevel survival models.

The {stdmest} package implements post-estimation predictions of standardised survival probabilities after fitting hierarchical survival models using -mestreg- in Stata. It is the R equivalent of the -stdmest- command in Stata.

This package can be used to obtain predictions of, e.g., standardised survival probabilities while fixing posterior predictions of the random effects at any level of the hierarchy.

In other words, {stdmest} can obtain marginal predictions standardising across observed covariates (i.e., the fixed effects) while fixing predicted values of the random effects. Built-in post-estimation commands for -mestreg- can do the opposite, i.e., marginalising over the random effects.

Moreover, the package includes a set of functions that can be used to import and pre-process estimation results exported from Stata, which are required to obtain predictions.

Installation

The development version of stdmest can be installed from this GitHub repository by typing the following in your Stata console:

# install.packages("devtools")
devtools::install_github("RedDoorAnalytics/stdmest-r", build_vignettes = TRUE)

References

• Gasparini, A., Crowther, M.J. & Schaffer, J.M. Standardized survival probabilities and contrasts between hierarchical units in multilevel survival models. BMC Med Res Methodol (2026). https://doi.org/10.1186/s12874-026-02782-8