I think it is necessary to be aware of the intended scope of a
patient identifier. For a (potentially) universal identifier
the FUPI isn't good enough.
> First digit would be
> 0 = female
> 1 = male
By what criterion ? Genotype, phenotype, patient's psychological
preference ? We at least need to add a category "other".
> next 8 digits are the D.O.B. (as best can be determined!) with the year first
> (4 digits), followed by month (2 digits) followed by day (2 digits)
For sorting that would do. But not for identification. If a
patient comes up and we have to depend on the DOB _as best as
can be determined_ we need a fuzzy search algorithm. How am I
to know that the patient who today claims to be born "probably May
1923" didn't tell me "uh, maybe June 1923 CE" or "maybe May
1924" when he/she registered first time in the office ?
And how do I know who is who if the systems turns up two
similar people who match the DOB close enough ?
And again, what calendar of reference ? Hindu, Buddhist,
Muslim, Christian ?
> Next, first 4 letters of Last name (given at time of birth, adopted would be
> adopted last name) (if Name is only 1 letter, 2 letters or 3 letters, just
> complete 4 letter code with a Z)
> Next, first 3 letters of First name (given at time of birth) (if Name is
> only 1 letter, 2 letters, just complete 3 letter code with a Z)
Again, what about cultures that have several valid rules for
constructing the "true" name of a person ? How do you tell what
is the Last or First Name ? Think Arabic, Thai, ...
Use of characters also necessitates the use of Unicode or
ambiguity regarding encoding phonetic transliteration of
non-ASCII languages arises.
As I said, scope is important. For the average Western
Hemisphere industrialized country (whatever _that_ is)
we'll be in good shape. We should probably hope to not see
requests from China or India (randomly selected countries) for
providing them with a good enough unique patient identifier,
though.
Isn't this whole thing somewhat similar to the "unique diagnosis
code" discussed in Slee's book ?
I can't recall the link from memory but you might want to search
for it at either
http://www.linuxmednews.com/LMNProjects
or
http://www.txoutcome.org/scripts/zope/library/
Just my two Baht.
Regards,
Karsten Hilbert
--
GPG key ID E4071346 @ certserver.pgp.org
E167 67FD A291 2BEA 73BD 4537 78B9 A9F9 E407 1346
PGP signature
