Skip to contents

Create counting-process dataset for time-dependent Cox model

Source: R/cox_analysis.R
cox_create_data.Rd

Constructs a long-format dataset suitable for survival::coxph() using counting-process notation Surv(tstart, tstop, event). The function prepares a follow-up cohort starting at baseline and models a time-dependent exposure diagnosis affecting the hazard of a response diagnosis.

Usage

cox_create_data(
  dpop,
  data_dates,
  data_socioeconomic,
  reference_values = list(bmi_cat1 = "Healthy Weight", bmi_cat2 = "Healthy Weight", edu =
    "1 - Low"),
  censoring_date = as.Date("2024-12-31")
)

Arguments

dpop

A data frame containing population-level variables. Must include ID, DATE_BIRTH, DATE_MIGRATION, DATE_DEATH, and diagnosis dates exp.DATE and resp.DATE.

data_dates

A data frame containing baseline dates. Must include ID and baseline date variable vpvmbl.

data_socioeconomic

A data frame containing socioeconomic and baseline questionnaire variables. Must include ID and covariates such as edu, bmi, bmi_cat1, and bmi_cat2.

reference_values

A named list defining reference levels for factor variables (e.g., education or BMI categories). Passed internally to healthpopR:::.relevel_by_reference().

censoring_date

Administrative censoring date. Default is as.Date("2024-12-31").

Value

A long-format data frame with the following variables:

ID

Individual identifier

tstart

Start of interval (days since baseline)

tstop

End of interval (days since baseline)

event

Event indicator (1 = response diagnosis, 0 = censored)

exposure_td

Time-dependent exposure indicator (0/1)

age_bs

Age at baseline (years)

edu

Education level (factor)

bmi

Baseline BMI

The output is ready for use in:


coxph(Surv(tstart, tstop, event) ~ exposure_td + ..., data = output)

Details

Baseline covariates (e.g., age, education, BMI) are treated as fixed. Exposure is handled as a time-dependent variable that switches from 0 to 1 at the exposure diagnosis date, if it occurs before the end of follow-up.

The function performs the following steps:

  1. Filters individuals with a valid baseline date.

  2. Computes age at baseline.

  3. Recodes exposure and response diagnoses occurring before baseline to the baseline date.

  4. Defines follow-up end as the minimum of migration, death, administrative censoring, or response diagnosis.

  5. Computes follow-up times (in days) from baseline.

  6. Splits follow-up into one or two intervals depending on whether exposure occurs before the end of follow-up.

Each row in the output represents a time interval during which exposure status is constant.

See also

coxph, Surv