No one's ever asked I'll dig it out and send it to you.

However, I don't think that the provision of such information provides much of the solution to good quality data transfer between systems It just removes some of the initial guesswork which is otherwise required. 

Just looking at you file to illustrate a few of the issues that hit me on a cursory examination.

What are the constraints on the data I can expect in a given field (e.g maximum length, format) are the field contents free text, from a bounded list or coded. (if so what coding schema has been used). 

How do I relate one item to another: A drug to indication for example

What can I expect in items labelled data item 1 to 10 in test result. A numeric value, a unit, a normal value, a comment form the lab, a date the test was ordered, the name of the requesting GP.

I heard good reports of your system but there seem to be lots of things you can't record in this structure. (a referral, a recall date, the location of an encounter, date of repeat script authorisation, date of drug issue, any data about the issue of a repeat, identity of author of note entry) 

I must be missing something on your drug entries. I can't see strength or quantity (is number taken quantity?). On repeats how many are authorised and how many have been issued. Have you got you acute and repeat formats the wrong way round? They make slightly more sense if you have.

Surely this can't be the format for a full extract of data from you system.

I not trying to be destructive but the fact that you feel the provision of a format like yours is a solution to the problem of data conversion and transfer between systems worries me. To adequately represent the information in a comprehensive GP system the structure needs to be much richer than your example and there has to be a unambiguous definition of the data contained in that structure.

Have a look at http://www.clinical-info.co.uk/cenpmr.htm for an insight into the complexity of the problem
-------------------------------------------------------------------------------------------------
Ewan Davis
AAH Meditel - Voice +44 (1) 527 579414 Fax +44(1)527 837287
Email [log in to unmask] also at [log in to unmask]


-----Original Message-----
From:	P S Buttar [SMTP:[log in to unmask]]
Sent:	Monday, June 01, 1998 8:02 PM
To:	[log in to unmask]
Subject:	RE: Data conversion/transfer: a proposal

Ewan,

> It has been a requirement since RFA 1 for systems to be able to output their data in to an ASCII file for which documentation must be available. 

But I'm proposing something beyond this. Attached is the format used by my system for export - your programmers should have no difficulty with devising an import function from this (or any other) format.

Any chance that you can let me have your DOS export format in return?

Prit
 << File: PMP.TXT >>