extract_time_until: Extract a 'time until' type variable

In rcprd: Extraction and Management of Clinical Practice Research Datalink Data

Extract a 'time until' type variable

Description

Query an RSQLite database and a data frame with the time until first code of interest or censoring, and an event/censoring indicator.

Usage

extract_time_until(
  cohort,
  varname_time = NULL,
  varname_indicator = NULL,
  codelist = NULL,
  codelist_vector = NULL,
  indexdt,
  censdt,
  censdt_lag = 0,
  t = NULL,
  t_varname = TRUE,
  db_open = NULL,
  db = NULL,
  db_filepath = NULL,
  tab = c("observation", "drugissue", "hes_primary", "death"),
  out_save_disk = FALSE,
  out_subdir = NULL,
  out_filepath = NULL,
  return_output = FALSE
)

Arguments

cohort	Cohort of individuals to extract the variable for.
varname_time	Name of time variable in the outputted data frame.
varname_indicator	Name of event/censoring indicator in the outputted data frame.
codelist	Name of codelist (stored on hard disk) to query the database with.
codelist_vector	Vector of codes to query the database with. This takes precedent over codelist if both are specified.
indexdt	Name of variable in cohort which specifies the index date. The extracted variable will be calculated relative to this.
censdt	Name of variable in cohort which specifies the censoring date.
censdt_lag	Number of days after censoring where events will still be considered, to account for delays in recording.
t	Number of days after indexdt at which to extract the variable.
t_varname	Whether to alter the variable name in the outputted data frame to reflect t.
db_open	An open SQLite database connection created using RSQLite::dbConnect, to be queried.
db	Name of SQLITE database on hard disk (stored in "data/sql/"), to be queried.
db_filepath	Full filepath to SQLITE database on hard disk, to be queried.
tab	Table name to query in SQLite database.
out_save_disk	If TRUE will attempt to save outputted data frame to directory "data/extraction/".
out_subdir	Sub-directory of "data/extraction/" to save outputted data frame into.
out_filepath	Full filepath and filename to save outputted data frame into.
return_output	If TRUE will return outputted data frame into R workspace.

Details

Specifying db requires a specific underlying directory structure. The SQLite database must be stored in "data/sql/" relative to the working directory. If the SQLite database is accessed through db, the connection will be opened and then closed after the query is complete. The same is true if the database is accessed through db_filepath. A connection to the SQLite database can also be opened manually using RSQLite::dbConnect, and then using the object as input to parameter db_open. After wards, the connection must be closed manually using RSQLite::dbDisconnect. If db_open is specified, this will take precedence over db or db_filepath.

If out_save_disk = TRUE, the data frame will automatically be written to an .rds file in a subdirectory "data/extraction/" of the working directory. This directory structure must be created in advance. out_subdir can be used to specify subdirectories within "data/extraction/". These options will use a default naming convetion. This can be overwritten using out_filepath to manually specify the location on the hard disk to save. Alternatively, return the data frame into the R workspace using return_output = TRUE and then save onto the hard disk manually.

Codelists can be specified in two ways. The first is to read the codelist into R as a character vector and then specify through the argument codelist_vector. Codelists stored on the hard disk can also be referred to from the codelist argument, but require a specific underlying directory structure. The codelist on the hard disk must be stored in a directory called "codelists/analysis/" relative to the working directory. The codelist must be a .csv file, and contain a column "medcodeid", "prodcodeid" or "ICD10" depending on the input for argument tab. The input to argument codelist should just be a character string of the name of the files (excluding the suffix '.csv'). The codelist_vector option will take precedence over the codelist argument if both are specified.

If the time until event is the same as time until censored, this will be considered an event (var_indicator = 1)

If dtcens.lag > 0, then the time until the event of interest will be the time until the minimum of the event of interest, and date of censoring.

Value

A data frame with variable patid, a variable containing the time until event/censoring, and a variable containing event/censoring indicator.

Examples

## Connect
aurum_extract <- connect_database(file.path(tempdir(), "temp.sqlite"))

## Create SQLite database using cprd_extract
cprd_extract(aurum_extract,
filepath = system.file("aurum_data", package = "rcprd"),
filetype = "observation", use_set = FALSE)

## Define cohort and add index date and censoring date
pat<-extract_cohort(system.file("aurum_data", package = "rcprd"))
pat$indexdt <- as.Date("01/01/1955", format = "%d/%m/%Y")
pat$fup_end <- as.Date("01/01/2000", format = "%d/%m/%Y")

## Extract time until event/censoring
extract_time_until(pat,
codelist_vector = "187341000000114",
indexdt = "fup_start",
censdt = "fup_end",
db_open = aurum_extract,
tab = "observation",
return_output = TRUE)

## clean up
RSQLite::dbDisconnect(aurum_extract)
unlink(file.path(tempdir(), "temp.sqlite"))

rcprd documentation built on April 12, 2025, 1:57 a.m.