Re: Modelling not-applicable fields
On 26/06/2010 11:20, Ian McNicoll wrote: > > > We recognise a need to find a way of easily adding a comment to any > node, when a null is recorded - a similar issue arose recently with > the Glasgow Coma Scale when e.g. Eyes could not be observed. > > There are 2 potential solutions.. > > 1. Add a 'reason for null' text attribute in the Reference model > alongside the null itself. I think this would be practical and easy. I don't see the need for each of the 100 posible reasons for null to be computable, just the fact of being not applicable, not available or unknown should be enough for computing purposes (but - maybe some researcher will show why this is wrong). An additional text attribute would seem a good start. > > 2. Use the link attribute available to every node to attach an > 'annotation' - this will be recorded in a separate archetype, which > allows considerable flexibility as to whether a simple text field is > required or perhaps more complex, structured 'reasons for null' are > required. This approach is being investigated as a means of > annotating > data elements for quality or 'fitness-for-purpose'. e.g. Is a device > obtained blood pressure valid fro graphing or should it be excluded? I think this is not a good idea, as it complicates the recording of even simple data with Links, which are essentially cross-references, which are more difficult to process. - thomas beale _______________________________________________ openEHR-clinical mailing list openEHR-clinical@openehr.org http://lists.chime.ucl.ac.uk/mailman/listinfo/openehr-clinical