On Wed, 09 Oct 1996 09:49:15 +0100, you wrote:
>> What I would need to be able to do is submit a request to the
>> directory service with some reasonably uniquely identifying attribute
>> of my patient (New NHS Number ?), and receive back a list of the
>> locations where record fragments for that individual are stored
>
>The Administrative Registers are supposed to do something like this.
Ineed so, but its a significant rethink of an Administrative Register
- it would be a URL server, if you like, and need contain no actual
patient info at all, unlike the AR as we know it today. Access rights
to the information pointed to by the "AR" could be embodied in the
pointers themselves, or (and I think perhaps the more sensible
approach) controlled locally by the systems/ applications that
actually house the information fragment.
>The problem of course is that you might not want an attendance at a
>GUM clinic to become widely known. Come to think of it, if you're
>the French president you wouldn't want an attendance at an oncology
>clinic to become known
>
>Ross
Quite so. As John pointed out, an appropriate security/
confidentiality framework would be required to deal with such issues.
That doesn't detract, however, from the practicality of the approach
he outlined.
Someone would have to be responsible for determining what patient
record fragments are pointed to by the central pointer system - there
could be several models - maybe the "owner" of the data, the patient's
Urologist, say, decides how much GUM related info will be pointed to
by the "AR" and who can access it - an access control list. He
probably bases his decisions on discussion with the patient and on
nationally agreed guidelines. The Central component ("AR") would then
reflect locally controlled access decisions/rules, not Central
decisions. The only reason for it being a Central resource would be
to save time and duplication having to trawl everyone's systems to
construct the whole patient record each time. However with
Intelligent Agent technologies, as John points out, maybe a Central
pointer service would be uneccesary - it would be down to performance
and practicality, perhaps, to determine which is the better approach.
Strikes me, therefore, the technology is here to support this kind of
approach, and is in line with your overall thinking, not at odds
unless there's something I'm missing or you have a better underlying
technical approach/solution/model to suggest..... ?
---
Rob Tweed
IM&T Consulting Ltd; Health Web Services Ltd;
M/Gateway Developments Ltd
http://www.hwsl.co.uk/mgw
Tel: (+44) 181 540 1325
Fax: (+44) 181 715 4337
---
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|