1

I have a medical data set that has all the following cases:

  • Right-censored
  • Left-censored
  • Interval-censored
  • Delayed-censored and
  • truncated-data.

I am trying to fit this data with CoxPH function in survival package in R. So my question, is there any consideration I have to take when I feed the data to Cox model, or just use the data as is.

Ferdi
  • 4,882
  • 7
  • 42
  • 62
Abdal
  • 125
  • 1
  • 8

1 Answers1

1

The analysis used by coxph() only allows for right-censored or counting process (left truncated, right censored/event) data. From the code (in survival_3.2-7):

if (type == "mright" || type == "mcounting") 
    multi <- TRUE
else if (type != "right" && type != "counting") 
    stop(paste("Cox model doesn't support \"", type, "\" survival data", 
        sep = ""))

where "mright" and "mcounting" are right censoring or counting-process data in a multi-state model.

Ways to handle other types of censoring or truncation are discussed here.

EdM
  • 57,766
  • 7
  • 66
  • 187