Datatypes R2 Issue 27

Introduction

EIVL<TS>: Deprecate many of the current event codes that are of the form Ax, Px, Ix (ante-x, post-x, inter-x), such as ACV, PCV, ICV. Instead put in codes for simply C, CM, CD, CV, etc. to indicate the event. Then the offset interval would be interpreted as follows: negative offsets are measured from the start of the event (e.g., 30 min before breakfast) and positive offsets are measured from the end of the event (e.g., 2 hours after meals.) The problem with the present definition is that the offset direction (whether it is before or after) is indicated by the sign AS WELL AS the event code.

And the EIVL component summary is wrong

? backward compatible: parts of this are not backward compatible

Discussion

Removal of Unrelated Ammedment

I (Gschadow 17:51, 7 June 2007 (CDT)) removed the following ammendment:

Also add the following event codes (with the offset semantics as specified in previous change)

* COIT Sexual intercourse

* DIAP Diaper change

* DEFEC Defecation

this is an issue separate from EIVL. The original proposal was vetted by Pharmacy SIG. This ammendment, however, was discussed and rejected by the same group back in 2004 and after heavy confusion it was unanimously determined that it should not be done. The consensus can be seen in some of the following statements from the email archive:

Lloyd McKenzie [Sep 9 2005 - 6:12pm]:

The EIVL datatype should be restricted to use for [...] setting where the specified timings can be resolved to specific times or time-windows based on [...] schedules. EIVL event types should be restricted to those which are resolvable [...]. All other events should be communicated as preconditions with offsets [pauseQty].

Gunther Schadow [Sep 9 2005 - 6:30pm]:

I agree to the intent. [...] I think it comes darn close to the original intent of EIVL, which was to spefify *regular* events that are usually relatively regularly timed, such as meal and sleeping, but not sex and other urges for relief.

Julie James [Sep 12 2005 - 5:44am ]:

I stick with my view that anything that can be "mathematically measured time" for a dosage is part of GTS; therefore if you say "one at breakfast" when you mean "one on the 8am drug round" then those are synonymous phrases - if you want to capture that synonymy in a part of [...] EIVL, then that's OK.

If a dosage timing is in any sense "circumstantial" and cannot be exactly predicted without using another variable other than the passage of time [...] is not catered for in GTS, which is "mathematical".

I am trying to follow the same logic for GTS that UCUM uses for measurement of quantity; if you can reproducibly measure the pattern of time for a dosage, use GTS, if you can't, don't.

My rationale is to try to move to sensible maths for dosage checking (we kill more people by wrong dosage than by anything else) and also to know when there is *not* maths for dose checking, so that applications can handle that too, rather than trying (as we have been for so many years) to make inferences that are all too often wrong.

Tom De Jong [10 September 2005 21:27]:

I think the difference is between 'regularly scheduled' and 'regularly occurring'. Meals and bedtime can be regularly scheduled, but things like menstruation and defecation are only regularly occurring. They aren't scheduled. I apologize for trying to squeeze these concepts into EIVL last year, because that would have stretched the concept too far (which resulted in the complete data type being suspended' by the Pharmacy SIG until now).

I hope that settles it. Gschadow 17:51, 7 June 2007 (CDT)

Disposition

Motion: Approve these changes

Approve: Grahame
Oppose:
Abstain:

Links

Back to Data Types R2 issues