create_standata_outcome: Create List for Fitting Outcome Model via STAN
In jpkeller/bercs: Bayesian Exposure-Response Curves via Stan

Description Usage Arguments Details See Also

View source: R/functions_outcome_model.R

Generates a 'standata_outcome' object from a 'long' data frame of outcome data

create_standata_outcome(..., datalist = NULL, Mtlist = NULL, covarslist = NULL)

create_standata_outcome_singlestudy(
  data = NULL,
  unit_id = data$unit_id,
  conc = data$conc,
  study = data$study,
  clust_id = data$clust_id,
  case = data$case,
  at_risk = data$at_risk,
  covars = "1",
  scale_covars = TRUE,
  xdf = 1,
  xfn = "iSpline",
  xfnargs = list(),
  Mt = NULL,
  timefn = "ns",
  timedf = 0,
  timefnargs = list(),
  return_addition = FALSE
)

`...`	arguments passed to `create_standata_outcome_singlestudy`. See Details.
`datalist`	List of data frames, containing the data in 'long' format.
`Mtlist`	time spline matrix, or a list of such matrices.
`covarslist`	list of covariate variable names for each study, or a list of covariate matrices.
`data`	Optional data frame containing data in a long format.
`unit_id`	Vector identifying the distinct unit (e.g. person) for each observation.
`conc`	Vector of exposure concentrations for each observation.
`study`	Vector identifying the study of each observation.
`clust_id`	Optional vector identifying cluster membership for each observation.
`case`	Count of cases for the observations. In most cases, this is a 0/1 indicator.
`at_risk`	Time at risk for the observation
`covars`	character vector of variables to include as covariates, or a vector/matrix of covariate values. Should not include an intercept. If a character vector, this intercept is automatically removed (so default of "1" leads to no covariates).
`scale_covars`	should the covariate matrix be centered and scaled?
`xdf`	Degrees of freedom for exposure splines
`xfn`	name of function used to generate exposure splines
`xfnargs`	named list of additional arguments for `xfn`
`Mt`	Optional matrix of time splines
`timefn`	name of function used to generate time splines
`timedf`	degrees of freedom for time spline
`timefnargs`	named list of additional arugments to `timefn`
`return_addition`	should the modified version of `data` be returned in addition to the `standata_outcome` object?

An important special case of create_standata_outcome is when there is only a single study. In this case, the values are passed to create_standata_outcome_singlestudy without pre-processing.

If using a separate temporal spline for each study, first compute the time splines for each study individually using create_spline. Combine these into a list and pass that to Mtlist.

The matrices of adjustment variables are created using the names provided to covarslist. This is done using formula and model.matrix, so transformations and interactions can be used in the typical manner.

create_standata_exposure, add_spline_exposure, sample_outcome_model

jpkeller/bercs documentation built on March 24, 2021, 5:36 a.m.