From: owner-mnm@lists.hl7.org on behalf of Julie James [Julie_James@BlueWaveInformatics.co.uk]
Sent: Friday, November 12, 2004 5:05 AM
To: pharmacy@lists.hl7.org; 'MNM listserv'
Cc: 'Bishop Charlie'; 'Hugh_Glover'; 'Charlie McCay'; 'David Markwell'
Subject: Variable Doses - Data Type proposal

Attachments: footer.txt; Dose Calculation.doc

HI

We discussed the proposition of data type extension for managing “mg/kg” type dosage expression at the HL7 UK TC meeting.

HL7 UK seemed happy to stick with the derivation.Expr way of doing things, on the grounds that everyone was going to have to get to grips with an OCL at some stage (maybe GELLO…) and that putting structure into xml was possibly less valuable than working with OCL, in the long run.  I think this was the crux of the debate, anyway – it was very much at the limits of my range of skills….. Although I presented it pretty much as in Gunther’s e-mail, and even though folk kept saying “no, that’s not right” actually, when we dug, we decided “yes, it is because….”, all to do with the relationship between the meaning of derivation.Expr and the mood of the act and the related observation, and how it all hangs together……

We did feel the RIM definition for derivation.Expr was a little ambiguous, as there were folk who thought this attribute specifically should be used to describe how a stated “value” has been calculated when there are more than one way, whereas the actual definition “allows” for the interpretation we are thinking of for SBADM dosing but it would be good to have it stated that the derivation.Expr can be used “to give calculation information to be used to reach a value” as well as “to show how a value was reached”.

Definition: A character string containing a formal language expression that specifies how the Act's attributes are derived from input parameters associated with derivation relationships.

Possibly…..

Definition: A character string containing a formal language expression that specifies how the Act's attributes are (should be/have been) derived from input parameters associated with derivation relationships.

 

I don’t feel that there is consensus on how to do variable doses, but I do need to get somewhere more concrete quickly, please….. and it’s not something I have enough expertise in to know what is best……..so, please, those of you who do…..

Maybe this should be copied to folk on “data type” lists – wherever they may be……

I’ve copied this to MnM as a start……, with the attachment of my summary of Gunther’s original proposal, which is what I presented to HL7 UK

I look forward to the discussion, even if there are bits I cannot follow

 

Best wishes

Julie

 

Julie M James MRPharmS

Blue Wave Informatics LLP

91a Queen Street

EXETER

EX4 3RP

 

julie_james@bluewaveinformatics.co.uk

+44 (0)7747 633216

+44 (0)1392 271274