Thomas Beale wrote: > pablo pazos wrote: >> >> Consider this ER scenario: a BP value could be recorded each 30" or so, and >> the system could be used 1. for many patients, 2. by many users, 3. on the >> same machine. > > this is most likely a 1-event-per-Observation scenario. I realise it is not > always obvious when to use which recording approach! But the other design > aspect of a COMPOSITION is that it is a 'single health system event for the > patient'. Here it sounds like 1 nursing observation every 30 mins. Therefore > we would expect 1 COMPOSITION for each one, each containing one OBSERVATION, > containing one EVENT.
An important consideration here is the composer of the composition. Different nurses will be recording the readings during the course of the day (or days, or weeks ...), but each composition can have only one composer. (You could get around that by adding an updated version of the composition with each reading, so the latest version would contain all of the data, but that would be a truly baroque approach! It would make it difficult to figure out which nurse had recorded which reading.) Another consideration is that the nurse is likely to be recording other observations at the same time as the BP. It seems logical to me that all of these observations should go into the same composition, because they were all done at the same time, by the same committer, for the same subject of care. On the other hand, if the BP readings are coming from a patient monitor, say, every 30 seconds, then it would make sense to store all of these BP readings in one composition. When would you decide, okay, that's enough, let's start another composition? Maybe every hour? Each day? Or maybe at the point in time when a clinician reviews them and says, "Yep, I've reviewed those BPs, commit 'em"? Peter