Build a risk calculator — risk_calculator • riskcalc

Create a (templated) risk calculator for www.riskcalc.org, optionally populating using an existing model object

Usage

risk_calculator(...)

# Default S3 method
risk_calculator(app_name = NULL, app_directory = NULL, ...)

# S3 method for class 'glm'
risk_calculator(
  model,
  app_name = NULL,
  app_directory = NULL,
  title = "",
  citation = "",
  label = "Predicted Value",
  label_header = "Result",
  value_header = "Value",
  format = NULL,
  labels = NULL,
  levels = NULL,
  placeholders = NULL,
  ...
)

# S3 method for class 'coxph'
risk_calculator(
  model,
  time,
  app_name = NULL,
  app_directory = NULL,
  title = "",
  citation = "",
  label = "Survival Probability",
  label_header = "Result",
  value_header = "Value",
  format = NULL,
  labels = NULL,
  levels = NULL,
  placeholders = NULL,
  ...
)

Arguments

...: Generic arguments
app_name: Application name (e.g., "AppExample" for www.riskcalc.org/AppExample. If no name is provided, a generic name "riskcalc_app" is used.
app_directory: Location to export the application. If none provided, defaults to the working directory.
model: A glm or coxph object
title: Title for risk calculator
citation: Citation(s) and author information
label: Label for the calculated value
label_header: Header label for the result descriptor
value_header: Header label for the result value
format: Function to format the predicted value for display
labels: Named character vector specifying labels for any inputs
levels: Named list of named character vectors specifying labels for any factor levels
placeholders: Named character vector specifying range limits for numeric input variables (of the form c(var="<min>-<max>")). Invokes calls to validate enforcing the limits.
time: Single time point to calculate survival probability (see summary.survfit)

Value

A shiny application

Examples

library(survival)

# Make a data set
dat <- gbsg
dat$meno <- factor(dat$meno)

# Build a model
mod <-
  coxph(
    formula = Surv(rfstime, status) ~ age + meno,
    data = dat
  )

# Time point of interest
time <- median(dat$rfstime) # 1084

# Create the risk calculator
if (FALSE) { # \dontrun{
risk_calculator(
  model = mod,
  time = time,
  app_name = "BreastCancer",
  app_directory = getwd(),
  title = "Risk of Death or Recurrence in Breast Cancer",
  citation =
    paste0(
     "Patrick Royston and Douglas Altman, External validation of a Cox prognostic",
     "model: principles and methods. BMC Medical Research Methodology 2013, 13:33"
    ),
  label_header = "`Probability Of Death or Recurrence`",
  label = paste0("At time = ", time),
  labels = c(age = "Age (years)", meno = "Menopausal Status"),
  levels = list(meno = c(`0` = "Premenopausal", `1` = "Postmenopausal")),
  placeholders = c(age = "21-80"),
  format = function(x) paste0(round(100 * (1 - x), 2), "%")
)
} # }