tojointdata: Function to change multi-study data into jointdata format
In joineRmeta: Joint Modelling for Meta-Analytic (Multi-Study) Data

Description Usage Arguments Details Value See Also Examples

This function is designed to take data in various formats from multiple studies and output the data in a jointdata format.

tojointdata(
  dataset = NULL,
  longitudinal = NULL,
  survival = NULL,
  baseline = NULL,
  id,
  longoutcome,
  timevarying = NULL,
  survtime,
  cens,
  time = NULL,
  longtimes = NULL
)

`dataset`	a dataset or list of datasets. The datasets can be of either long or wide format holding both the longitudinal and survival information (and any baseline information). Either the parameter `dataset` should be specified, or the parameters `longitudinal` and `survival` (and `baseline` if available) should be supplied.
`longitudinal`	a dataset or list of datasets in long format containing the longitudinal outcome and any time varying covariates. It can also contain baseline information. This should not be supplied if `dataset` is supplied to the function call.
`survival`	a dataset or list of datasets in wide format (one row per individual) containing the survival information (survival time and censoring variable). It can also contain baseline information. This should not be supplied if `dataset` is supplied to the function call.
`baseline`	a dataset or list of datasets in wide format (one row per individual) containing any baseline information. This variable does not have to be supplied if there is no baseline information, or if it is already contained in the longitudinal or survival datasets. This should not be supplied if `dataset` is supplied to the function call.
`id`	a character string holding the name of the id variable. This should be present in all datasets supplied to this function.
`longoutcome`	a character string holding the name of the longitudinal outcome.
`timevarying`	a vector of character strings indicating the names of the time varying covariates in the dataset
`survtime`	a character string denoting the name of the surivival time variable in the dataset
`cens`	a character string denoting the name of the censoring variable in the dataset
`time`	a character string to label the time variable if the data is transformed from wide to long format, or the name of the variable holding the time variable if the data is supplied in long format
`longtimes`	if wide data, the labels that denote the time varying variables to allow the longitudinal data to be returned in long data format.

The data supplied to the jointmeta1 function has to be in a jointdata format. However it is conceivable that data supplied from multiple studies could come in a range of formats.

This function can handle a range of formats in which data from multiple studies may be supplied. This are discussed below. We refer to wide data as data that contains both time varying and time stationary data, with one line per individual with variables measured over time recorded in multiple columns. We refer to long data as data with multiple lines per individual, with time varying covariates differing between rows, but time stationary covariates identical between rows. Survival data are considered data with survival outcome (survival time and censoring variable) with or without baseline data. Longitiudinal datasets are considered long format datasets containing time varying data potentially with baseline data. Baseline datasets are considered wide format datasets containing non-time varying data measured at baseline. This function can take data in the following formats and output a jointdata object.

One wide dataset: One dataset in wide format (one row per individual) supplied to the parameter dataset in the function call with longitudinal, survival and baseline set to NULL, or left unspecified in the function call. This dataset would contain all the data from all studies, and the survival, longitudinal and any baseline information would all be present in the same dataset.
One long dataset: One dataset in long format (multiple rows for each individual) supplied to the parameter dataset in the function call with longitudinal, survival and baseline set to NULL, or left unspecified in the function call. This dataset would contain all the data from all studies, and the survival, longitudinal and any baseline information would all be present in the same dataset.
A list of study specific wide datasets: One dataset for each study, each in wide format (one row per individual), supplied to the parameter dataset in the function call with longitudinal, survival and baseline set to NULL, or left unspecified in the function call. The data from each study would contain the survival, longitudinal and any baseline information.
A list of study specific long datasets: One dataset for each study, each in long format (multiple rows per individual), supplied to the parameter dataset in the function call with longitudinal, survival and baseline set to NULL or left unspecified in the function call. The data from each study would contain the survival, longitudinal and any baseline information.
One longitudinal and one survival dataset with or without an additional baseline dataset: In this case all the longitudinal and time varying data for all studies is supplied in a single dataset in long format to the parameter longitudinal. All the survival data for all studies is supplied in a single dataset in wide format to the parameter survival. Baseline data can be present in these two datasets, or can also be supplied as a dataset to the parameter baseline. If longitudinal and survival are specified, then parameter dataset should be set to NULL or left unspecified in the function call. The parameter baseline is optional, but should only be specified if parameter dataset is NULL or unspecified.
A list of longitudinal and a list of survival datasets with or without a list of baseline datasets: In this case the longitudinal and time varying data for each study is supplied as one element of a list of long format datasets to the parameter longitudinal. The survival data for each study is supplied as one element of a list of wide format datasets to the parameter survival. Baseline data can be present in these two sets of datasets, or can be supplied as an additional list of datasets one for each study to the parameter baseline. If longitudinal and survival are specified (baseline is optional), then parameter dataset should be set to NULL, or left unspecified in the function call.

The specified id variable should be present in all datasets supplied to the function. Variables containing the same information should be identically named in each supplied dataset, for example if a variable 'age' is present in one dataset, denoting age of individual at baseline, corresponding variables in other datasets also supplying age at baseline should also be named 'age'. Similarly, different variables should not share the same name across different datasets, for example there should not be a variable named 'age' in the longitudinal dataset denoting individual's age at last longitudinal measurement along with a variable 'age' in the baseline dataset that denotes age of the individual at baseline. Before supplying data to this function, names of variables in each dataset should be checked to confirm that common variables share the same name, and differing variables are appropriately distinguished from each other.

A jointdata object, see jointdata.

jointdata, jointmeta1

   #simdat is a simulated dataset available in the joineRmeta package
   #it is supplied as a list of longitudinal and a list of survival datasets,
   #each list is of length equal to the number of studies in the entire
   #dataset.
   jointdat<-tojointdata(longitudinal = simdat$longitudinal,
                         survival = simdat$survival, id = 'id',
                         longoutcome = 'Y', timevarying = c('time','ltime'),
                         survtime = 'survtime', cens = 'cens', time = 'time')