Learn R Programming

survival (version 3.8-3)

tcut: Factors for person-year calculations

Description

Attaches categories for person-year calculations to a variable without losing the underlying continuous representation

Usage

tcut(x, breaks, labels, scale=1)
# S3 method for tcut
levels(x)

Value

An object of class tcut

Arguments

x

numeric/date variable

breaks

breaks between categories, which are right-continuous

labels

labels for categories

scale

Multiply x and breaks by this.

See Also

cut, pyears

Examples

Run this code
# For pyears, all time variable need to be on the same scale; but
# futime is in months and age is in years
test <- mgus2
test$years <- test$futime/30.5   # follow-up in years

# first grouping based on years from starting age (= current age)
# second based on years since enrollment (all start at 0)
test$agegrp <- tcut(test$age, c(0,60, 70, 80, 100), 
                     c("<=60", "60-70", "70-80", ">80"))
test$fgrp  <- tcut(rep(0, nrow(test)), c(0, 1, 5, 10, 100),
                   c("0-1yr", "1-5yr", "5-10yr", ">10yr"))

# death rates per 1000, by age group
pfit1 <- pyears(Surv(years, death) ~ agegrp, scale =1000, data=test)
round(pfit1$event/ pfit1$pyears) 

#death rates per 100, by follow-up year and age
# there are excess deaths in the first year, within each age stratum
pfit2 <- pyears(Surv(years, death) ~ fgrp + agegrp, scale =1000, data=test)
round(pfit2$event/ pfit2$pyears)  

Run the code above in your browser using DataLab