Hi all,
Some quick thoughts from somebody who's been through this before;
albeit with centralizing patient identification -- MPI's.
1) Master patient / person indices (MPI's) have the same problem
characteristics as payer indices; though with volumes several orders
of magnitude larger. Nonetheless the problems are the same.
2) The United States is a decentralized environment when it comes
to identifying most individuals and organizations. While we may think
of the Federal Government as big brother, when it comes to issues such
as these, it's simply not a player. Neither is any other governmental
agency except within its jurisdiction (e.g. state licensure of
professionals, state and city business licenses, library cards,
drivers licenses). Establishment of a commerical enterprise to do
this is possible, but so far nobody has done it all that well (e.g.,
Dow Jones, and Dunn & Bradstreet try, but are forever telephoning
folks to try to keep their listings current).
3) Even where jurisdictional-based lists of individuals or
organizations exist, there is presently virtually no way to
cross-reference the data. [Witness latest request by state motor
vehicle departments for federal money to build integratable data
bases.]
4) Given the above, I think we should consider for our starting
point the landing point for national MPI mediation -- whose groups
ceased any significant efforts a few years ago. [I know 'cause I was
there at HL7!]
So, bottom line ...
A) The probability of any centralized system being established, let
alone maintained, is nil. [See Rachel Foerster's post of earlier
today.]
B) We need to find a way to obtain electronic addressing
information from each patient / payor / provider / VAN / clearinghouse
/ whatever we do business with in the course of establishing any
relationship. [This is one of those cases where I think the current
paper and telephone / facsimile model works quite well.]
C) We need to let the communications links / networks handle the
routing [end-to-end and point-to-point within end-to-end] as a
severable issue from addressing. As one example, SMTP does routing
quite nicely for e-mail when only an end-point address is known. X12N
transactions are just ASCII text and could be e-mailed [with
appropriate encryption as necessary for HIPAA security / privacy].
D) For those familiar with the International Standards
Organization's Open Systems Interconnect model, think about the HIPAA
transactions being something like ...
level 7 data content (that which is only required by HIPAA
DDE)
level 6 data format (e.g., X12 format)
level 5 addresses for end-points of communication
level 4 routing the messages through the intermediate
way-points
[this is what TCP/IP does, amongst other things]
level 3 and below -- not really significant to our disucssion just
now.
OK, my contribution on this holiday when I'm not supposed to be
working :-)
Dave Feinberg
Member, sunsetted HL7 MPI Special Interest Group
Rensis Corporation
206-617-1717
[EMAIL PROTECTED]
