SSIS fuzzy lookup and dates

Hello, 

I'm using SSIS fuzzy lookup transformation to match clients together across several databases where the following predicates are 95% similar to each other:

SURNAME (text)
FORENAME (text)
GENDER (text)
DATE OF BIRTH (date) (format 'dd/MM/yyyy')

Because SSIS can only use text strings for fuzzy lookup transformations, what is the best way of using fuzzy lookup transformation on DATE OF BIRTH? 

1) Convert 01/01/2015 to '01/02/2015' ?

2) Split date parts into three columns - e.g.  '01' , '02', '2015'?

3) Convert short date type into long date types in three columns e.g. 'ONE', FEBRUARY', '2015' ?

4) Other?

In brief testing, it looks SSIS doesn't handle option 1) particularly well - i.e. '01/02/2015' is not deemed similar to '02/02/2015'.

Thanks!

OC