capturetb

Installation

remotes::install_github("CAPTURE-TB-HIV/capturetb")

Basic usage

To predict the unitcost of one outpatient visit given facility characteristics:

# Load model
model <- capturetb::unitcost()

# View covariates
covariates <- model$covariates()
print(covariates)
[1] "public"       "urban"        "healthcentre" "primary"      "secondary"   
[6] "tertiary"     "logVisits"   

# Generate predictions from posterior
inputs <- list(
  logVisits = 6.9, 
  healthcentre = FALSE,
  primary = TRUE,
  secondary = FALSE, 
  tertiary = FALSE,
  urban = FALSE, 
  public = TRUE,
  output = "op_treatmentvisit",
  fc_country = "Ethiopia"
)
prepared_inputs <- capturetb::prepare_covariates(inputs, model)
pred <- model$predict(
    prepared_inputs,
    scale = "natural",
    summarised = TRUE)

# Expected unit cost is mean prediction
expected_unit_cost <- pred$Mean
print(expected_unit_cost)
[1] 17.72955

Advanced usage

If fitting models, you will also need to install JAGS and the runjags package.

See vignettes for full documentation.

Testing

To run tests:

devtools::load_all()
devtools::test()

To build pkgdown documentation locally:

devtools::load_all()
pkgdown::build_site()

Files will appear in the docs/ folder.