Synthetic biomarker study: outcomes

One row per participant and endpoint of the made-up study in Synthetic_Results, with a time and a censor flag. There is one endpoint, event-free survival in months.

Usage

Synthetic_Outcomes

Format

A data frame with 200 rows and 5 columns. The names are ADaM's time-to-event names.

USUBJID

character Participant identifier, BIO-001 to BIO-200.

PARAMCD

character Endpoint code, EFS.

PARAM

character Endpoint name, Event-free survival (months).

AVAL

numeric Time to the event or to censoring, in months, to two decimal places.

CNSR

integer Censor flag, ADaM's way round: 0 for an event, 1 for a censored time.

Source

Generated by data-raw/synthetic-study.R. The same table is in the package as system.file("extdata", "synthetic_outcomes.csv", package = "gsm.bio").

Details

Everything in it is synthetic: the times are random draws, not observations.

Planted effect

Participants are split at the median of their Baseline CRP in Synthetic_Results, 100 above it and 100 below, with no value on the median. Event times are exponential, and the hazard in the high group is the hazard in the low group multiplied by a constant: a hazard ratio of 2.5, high against low. CRP has no other relationship with the outcome, and no other biomarker has any.

Censoring is independent of the event: a participant drops out at an exponential time, and follow-up ends at 24 months. The true hazard ratio is a parameter of the generating model, held in Synthetic_Truth.

See also

Synthetic_Results, Synthetic_Participants, Synthetic_Truth