standardizeRisks: Estimate standardized mortality risks

In StatGent/RiskStandard: Package for benchmarking clinical care centers

Description

Estimate standardized mortality risks

Usage

standardizeRisks(patientCovariates, center, Y, method = c("indirect",
  "direct"), Firth = TRUE, alpha = 0.05, missing = c("completeCase",
  "dummyCategory"), trace = FALSE)

Arguments

patientCovariates	data frame of patient-specific covariates
center	center code (n values for n patients)
Y	binary outcome
method	method of standardization; one of 'indirect' (default) or 'direct'
Firth	logical apply Firth correction? default value TRUE
alpha	statistical significance level; default value is 0.05
missing	how to handle missing categorical data? one of 'completeCase' (default) or 'dummyCategory'; when 'dummyCategory' is chosen a separate category is added to model a missing value effect
trace	logical print summary of fitted model? default value FALSE

Value

object of class 'standardizedRisks'; data frame with columns centerName, centerSize, standardizedRisk, varStandardizedRisk, lowerCI, upperCI and observedRisk

References

Varewyck M., Goetghebeur E., Eriksson M. and Vansteelandt S. (2014), On shrinkage and model extrapolation in the evaluation of clinical center performance, Biostatistics, 15(4), p. 651–664

Examples

set.seed(130513)
simulatedData <- simulateData()
standardizedRisks <- with(simulatedData, standardizeRisks(patientCovariates = L, center = center, Y = Y))
head(standardizedRisks)

StatGent/RiskStandard documentation built on May 9, 2019, 1:57 p.m.