Derive Age where DOB = Feb 29th

Question

I'm looking for a method to derive age that will accurately handle the date of birth = Feb 29th. I've tried using the new_interval function on the lubridatepackage, however for a child born on 29-Feb-2004, this returns an age of 2 years old on 28-Feb-2006. The child would be 2 years old on 01-Mar-2006.
I've also tried using the decimal_date function, but that gives me an error when date = 01-Jan-2006.

I'm using R version 2.15.3 (2013-03-01) -- "Security Blanket"
Here is my code:

library (lubridate)

MyTable <- data.frame(Eval.Date = c(as.Date("2006-01-01"), 
                                    as.Date(0:5, origin = "2006-02-26")),
                      Birth.Date = as.Date("2004-02-29"))

MyTable$Age <- floor(new_interval(MyTable$Birth.Date, MyTable$Eval.Date) / 
                     duration(num = 1, units = "years"))

MyTable$DecDate <- decimal_date(MyTable$Eval.Date)

MyTable[,c("Birth.Date","Eval.Date","Age","DecDate")]

Output:

  Birth.Date  Eval.Date Age  DecDate
1 2004-02-29 2006-01-01   1      NaN
2 2004-02-29 2006-02-26   1 2006.153
3 2004-02-29 2006-02-27   1 2006.156
4 2004-02-29 2006-02-28   2 2006.159
5 2004-02-29 2006-03-01   2 2006.162
6 2004-02-29 2006-03-02   2 2006.164
7 2004-02-29 2006-03-03   2 2006.167

Help either with decimal_date or the age calculation will be appreciated.
Thanks!

Answer 1

Using brute force:

library(lubridate)

age <- function(ED, BD) {
    year(ED) - year(BD) - 1 +
    (month(ED) > month(BD) | 
       (month(ED) == month(BD) & day(ED) >= day(BD)))
}

transform(MyTable, age = age(Eval.Date, Birth.Date))
#    Eval.Date Birth.Date age
# 1 2006-01-01 2004-02-29   1
# 2 2006-02-26 2004-02-29   1
# 3 2006-02-27 2004-02-29   1
# 4 2006-02-28 2004-02-29   1
# 5 2006-03-01 2004-02-29   2
# 6 2006-03-02 2004-02-29   2
# 7 2006-03-03 2004-02-29   2