Skip to contents

Add time-to-event (TTE) follow-up outcomes to a transplant cohort

Source: R/add_time_to_event_variables.R
add_txf_outcomes.Rd

Uses the cohort's file_key attribute (as set by your SRTR load helpers) to locate and load the corresponding follow-up table, then derives:

Usage

add_txf_outcomes(df)

Arguments

df

A data frame representing a transplant cohort (one row per recipient or transplant). Must include a TRR_ID column and have an attribute file_key (a string) indicating the originating SRTR table key (e.g., "KI_txf"). The companion follow-up key is inferred as the same prefix with "F" inserted after the first two characters (e.g., "KI_Ftxf").

Value

The input df with additional columns:

FIRST_FU_MALIG_DATE

Date of first follow-up malignancy (per TRR).

FIRST_FU_REJ_DATE_ORG

Date of first acute rejection per organ (wide columns by organ code).

Existing columns/attributes of df are preserved.

Details

  • FIRST_FU_MALIG_DATE — first follow-up malignancy date per TRR_ID

  • FIRST_FU_REJ_DATE_{ORG} — first acute rejection date per organ code (e.g., FIRST_FU_REJ_DATE_KI, FIRST_FU_REJ_DATE_PA)

The function limits the follow-up load to the cohort's TRR_IDs for efficiency, and merges the derived dates back into df.

This helper:

  1. Infers the follow-up table key from attr(df, "file_key").

  2. Loads the follow-up subset for the cohort's TRR_IDs via load_srtr_file().

  3. Normalizes organ to a two-letter code (ORG_TYPE := first two characters of ORG_TY), selects core follow-up fields, and orders by TRR_ID and TFL_PX_STAT_DT.

  4. Computes the earliest malignancy date per TRR_ID where TFL_MALIG == "Y".

  5. Computes the earliest acute rejection date per (TRR_ID, ORG_TYPE) where TFL_ACUTE_REJ_EPISODE indicates at least one treated episode, and pivots to wide organ-specific columns.

  6. Left-joins both results back to df.

Assumptions/requirements

  • The follow-up table contains: TRR_ID, ORG_TY, TFL_PX_STAT_DT, TFL_MALIG, and TFL_ACUTE_REJ_EPISODE.

  • TFL_PX_STAT_DT should be a Date; if stored as YYYYMMDD, convert upstream (e.g., as.Date(as.character(x), "%Y%m%d")).

  • Rejection is identified by the labeled value "1: Yes, at least one episode treated with anti-rejection agent". Adjust the filter if your coding differs.

Notes

  • Combined organ codes (e.g., "HL") are not split; they will produce a column FIRST_FU_REJ_DATE_HL. If you prefer to split into heart/lung, duplicate those rows before pivoting.

  • This function only derives event dates. To build full TTE variables (censor date, indicator, elapsed time), pass these dates to your srtr_time_to_event() helper along with transplant and last-follow-up dates.

See also

srtr_time_to_event, load_srtr_file

Examples

if (FALSE) {
# df loaded earlier, e.g., df <- load_srtr_file("KI_txf")
attr(df, "file_key")
#> "KI_txf"

df2 <- add_txf_outcomes(df)
names(df2)
# ... includes FIRST_FU_MALIG_DATE, FIRST_FU_REJ_DATE_KI, etc.

# Then compute time-to-event from transplant date:
df2 <- srtr_time_to_event(
  df2,
  event_date  = FIRST_FU_MALIG_DATE,
  start_date  = REC_TX_DT,
  censor_date = TFL_LAFUDATEKI,
  prefix      = "REC_MALIG"
)
}