time_varying_codes: Convert a dataframe of codes with start/stop time to a flat,...

In qPharmetra/PMDatR: Pharmacometric Data Management in R

Convert a dataframe of codes with start/stop time to a flat, time-varying covariate format

Description

Convert a dataframe of codes with start/stop time to a flat, time-varying covariate format

Usage

time_varying_codes(df, codes, stdtc, endtc, start_val = 1, end_val = 0,
  TIME = "TIME", .sticky = NULL)

Arguments

df	A dataframe (e.g. CM or AE domain)
codes	The name of a column that holds codes that will become columns
stdtc	An expression for the start time (e.g. CMSTDTC)
endtc	An expression for the end time (e.g. iso_to_posix(CMSTDTC, .time="24:00"))
start_val	An expression for the value to assign at the start time
end_val	An expression for the value to assign at the end time
TIME	A name for the new time column
.sticky	A value (set by start_val or end_val) which should be set permanently

Details

The function does the heavy lifting of separating codes into columns that are flagged at their start and stop times. The names of the columns are the codes provided in df. There is some setup that may be necessary to make proper use of this function:

• Set grouping before calling. The groups will be preserved.

• The times are assumed to be POSIXct or numeric. Conversions can be done using the mappings stdtc and endtc.

• Clean up the data to remove duplicate rows and create short codes that will make suitable column names.

• filter out codes that are not of interest. The function will create columns for ALL of the codes.

• the output of this function is inteded to be used in a Time-Varying Covariate query. Records occurring beyond the limits of the study can be removed there. The output keeps the original data on the rows to aid in further pruning/mapping of the dataset.

• The start_val and end_val mappings can be complex expressions to take into account not only the codes, but other columns as well. The values can be set arbitrarily.

Value

A reformatted dataframe

Examples

## Not run: 
cm %>% filter(CMDECOD %in% .(ATORVASTATIN,PARACETAMOL,IBUPROFEN,METFORMIN)) %>%
  group_by(USUBJID) %>%
  time_varying_codes(CMDECOD,
    iso_to_posix(CMSTDTC,.time = "00:00"),
    iso_to_posix(CMENDTC,.time = "24:00"),
    ifelse(CMENDTC=="",2,1),
    .sticky = 2) -> cm.out

## End(Not run)

qPharmetra/PMDatR documentation built on April 7, 2024, 5:42 p.m.