> >To allow exchange of EMRs (portability), these UPICs have to be not
> only unique
> Helpful, but not at all essential, actually.
> >but also allocated always by the same algorithm.
> Definitely not required. If the algorithm is good and free of course
> everyone might see the sense of using it, butthe NHS will clearly
> stick to its 10 character (all digit) string of which the last
> haracter is a check digit.
They certainly will. Doesn't matter. That does not help me at all to exchange EMRs
when a tourist from the UK gets medical attention here and comes back to the UK. Of
course, you will need some standard format such as GEHR to exchange the data, but you
need to identify the patient unambiguosly first.
Manual update of EMRs is error prone. Automatic update is only possible if you have
this persistent identifier. The UK solution is not persistent.
> Uses modulo 11 arithmetic to produce it, helping to protect against
> typos.
> I posted code for checking these a couple of years ago, and for
> producing numbers that validate in case anyone wanted some. (to test
> their system of course)
> Search GP-UK archives or ask me. In VB3
Could you please email it to me?
> >It is obvious, that the chances for duplicates will still be high
> >if you don't use a clearinghouse for those UPICs when the population
> is
> >large enough.
> Strangely, no. however the numbers are large to avoid that and may
> not be user-friendly.
> I refer to the guid that is used in a lot of MS code.
> A combination of the issuing place, and a 4 byte(?) number is very
> unlikely to collide, and much less so if each loation doesn't reissue
> any numbrs.
I think it is important that the identifier can be re-created at any time. Therefore I
insist on an algorithm that is based solely on information the patient can provide.
I can't understand why virtually all systems fail to recognize that. I believe it is
idle to discuss interoperability and exchange standards if we can't come forward with
unambiguous, unique, persistent and reproducible person identifiers, the _absolute_
prerequisite for any automated data exchange. I believe that user unfriendly codes
will ultimately produce errors. The UK NHS is not exactly a positive example for
working health IT.
I don't care that different nations have their own home baken solutions. I don't care
that in some nations there is an almost paranoid aversion against identifier codes. We
all need a solution to this problem. The "reproducible, patient info based" algorithm
could help us out in that aspect. Future EMR software can use that algorithm besides
any legally required identifiers, so data exchange can happen even internationally
despite the reluctance of nations to agree on standards. We could just do it!
Everybody just use the same algorithm and ignore the authorities! Even if the
algorithm would not be perfect, it would be far better thannothing and at least a
pointto start with and force the authorities to think and come up with something
better.
Horst
