Package index • hvtiRpropensity

Package Overview

Package-level documentation and the base S3 class shared by all hvtiRpropensity data objects.

hvtiRpropensity-package hvtiRpropensity: hvtiRpropensity: Propensity Score Methods for Cardiac Surgery Studies
is_ps_data(): Test whether an object is a ps_data instance

Data

Synthetic datasets for examples and unit tests.

sample_ps_data(): Generate sample data for propensity score examples
sample_ps_data_ordinal(): Generate sample data with an ordinal treatment for ps_ordinal() examples
sample_ps_data_nominal(): Generate sample data with a nominal treatment for ps_nominal() examples
sample_ps_data_count(): Generate sample data with a count outcome for bs_count() examples

Propensity Score Estimation

Functions for estimating propensity scores and balancing scores from raw patient data. Supports binary, ordinal and nominal treatments and continuous and count exposures, with optional multiple-imputation averaging.

ps_logistic(): Estimate propensity scores via binary logistic regression
ps_forest(): Random-forest propensity score (out-of-bag)
ps_ordinal(): Estimate propensity scores via ordinal (cumulative logit) logistic regression
ps_nominal(): Estimate propensity scores via nominal (generalised logit) logistic regression
bs_continuous(): Estimate a balancing score via linear regression
bs_count(): Estimate a balancing score via negative-binomial or Poisson regression

Outcome Models

Logistic and linear outcome-model bundles, and validation of a saved logistic bundle without refitting.

fit_logistic(): Fit a general logistic model bundle
fit_linear(): Fit an ordinary least-squares model bundle
validate_logistic(): Validate a saved binary logistic model

Propensity Score Methods

Functions for balancing treated and control groups via matching or inverse-probability-of-treatment weighting.

ps_match(): Propensity Score Nearest-Neighbour Matching
ps_weight(): Inverse-Probability-of-Treatment Weighting (IPTW)

Balance Diagnostics

Standardized differences between treated and control groups, by variable type and optionally weighted.

ps_stddiff(): Standardized differences between two groups
ps_stddiff_perm(): Permutation reference for standardized differences

Treatment Effects

Matching-weight differences in means and restricted mean survival contrasts, with bootstrap intervals.

ps_mw_var(): Bootstrap variance of a matching-weight treatment effect
ps_rmst(): Observed-outcome restricted mean survival contrast, weighted

Sensitivity Analysis

Rosenbaum bounds and E-values for unmeasured confounding; overlap, support and trimming diagnostics for positivity.

sa_rosenbaum(): Rosenbaum Sensitivity Bounds for Matched Analyses
sa_evalue(): E-Value Sensitivity Analysis
sa_overlap(): Propensity Score Overlap and Positivity Diagnostics
sa_trim_sweep(): IPTW Weight-Trimming Sensitivity Sweep
ps_support(): Compare support (overlap) rules

Internal

S3 methods for printing and summarising ps_data objects.

print(<ps_data>): Print a ps_data object
summary(<ps_data>): Summarise a ps_data object