When is the correct time to create an Observation

We have a question on the lifecycle for an Observation. Consider the following steps in a flow:

  1. Blood is drawn from the patient.
  2. Vial is shipped to the lab.
  3. Vial is received at the lab.
  4. Analysis is started.
  5. Analysis is done but not approved.
  6. Result is technically approved.
  7. Result is medically approved.

We internally agree to set the status to “preliminary” in step 5 and “final” in step 7. What we are discussing is when to create it with status “registered”. A few options:

A. In step 1, thus viewing the vial as a “pending observation that will eventually come into being”.
B. In step 3 since then that is the starting point for the analysis process that will lead to the observation.
C. In step 4 since that is when the actual analysis starts.
D. At another time altogether.

We would really appreciate if someone could share some insight on best practices here. Thank you in advance.

Best regard
Erik

most systems don’t create an observation at all until step #5 but some workflows are predicated on having it earlier than that - usually step 3. Or even earlier, in step #0, when the clinical system registers the forth-coming test with the lab system (but that only happens when there’s online ordering with some kind of pre-specimen workflow, which isn’t always the case)

In short - what do you need? There are a number of tools available - ‘filler’ DiagnosticRequests, Tasks and non-final Observation instances. Different workflows will drive different behaviors. If the real-world system has created a business object that will eventually correspond to the ‘final’ Observation, then that should be exposed as an Observation instance - with whatever status is most appropriate.

Thx for the response. We just wanted to be sure that we didn’t miss any best practice but then we know that there are several “correct ways” of doing this.

Thx for the pointers, we will check those resources to see if they match what we need.