OHDSI Home | Forums | Wiki | Github

Observation "visits"

(Melanie Philofsky) #1

In the US, we have this very specific type of Visit called an “Observation” visit. Currently, we map it to visit_concept_id = 0, since there isn’t a concept_id for this. Now we are re-visiting this decision. WWCD? (What would @Christian_Reich do?)

The Person may be in the hospital being observed for a few hours or the Person may be observed overnight or more. Since a Person hasn’t been given a formal “admit as inpatient” order from the Provider, do others map this to visit_concept_id = outpatient? This is Medicare’s point of view. However, a Person may have an ER Visit, then an Observation Visit and then an Inpatient Visit. And the Observation may take place in a department of the inpatient hospital. The care site varies, but when the care site is an inpatient department, the Visit seems more like an inpatient visit from a world-wide/OMOP point of view. It’s just not labeled by the hospital as an inpatient visit. How should we map this? And can we get a visit_source_concept_id = ‘observation visit’, so we can distinguish it from other visits?

(Christian Reich) #2


Want to use 581385 “Observation Room”?

(Melanie Philofsky) #3

I would, but the parent is Outpatient Hospital. And these data come from Inpatient Hospitals.

(Christian Reich) #4

Use it, and we will reposition in the hierarchy.

(Melanie Philofsky) #5

Done, thanks!

(Dmytry Dymshyts) #6

So, the Observation Room can belong to Intpatient Visit and Outpatient visit as well.
We should not reposition but add, right?

(Melanie Philofsky) #7

I’m not sure. I am not familiar with an “Observation Room”. I only know about the Observation Visit. It’s a very billing centric view of a Person’s Visit. And it isn’t an Outpatient Visit or an Inpatient Visit. It’s an “Observation” Visit.

(Christian Reich) #8


Well, wait. Is the patient in bed and the professional providers come to the bedside? And does the patient stay overnight?

These are the criteria. What the CMS thinks about it and how it classifies the claims is not that relevant, this is for the healthcare setting we are trying to capture.

(Roger Carlson) #9

I’m treating “Observation” visits as Inpatient visits. They just have a LOS less than 24 hours. I can’t see how it makes any difference from a research perspective. As far as I can see, observation visits are just an artifact created to address Medicare.

(Melanie Philofsky) #10

Providers go to the bedside. The person may leave after 2 hours or after 2 days/1 overnight.

“healthcare setting” sounds like a care site to me.

I am trying to make sense of my EHR whose purpose is to facilitate patient care and facilitate the retrieval of money for the care. The retrieval of money/billing is how this whole Observation Visit has come about in the US. Should there be a heuristic for these Observation Visits? > 24 hours = inpatient? < 24 hours = outpatient? Overnight isn’t a good indicator because 10pm to 1am is only 3 hours. Care site also isn’t good because the post-op department has IP, OP and Observation Visits all happening in the same place at the same time.

I’m in agreement with @roger.carlson What do others think?

(Christian Reich) #11

Sounds like Inpatient to me.

Sounds like Inpatient to me.

No. We made that split, remember? A care site can accommodate many different settings.

The heuristic is a combination of criteria, none of which are absolute:

  • Outpatient: Less than a day vs Inpatient: typically more than a day (but the patient could be kicked out the same day from Inpatient, so, that is an optional. It is not an optional for Outpatient visits. Outpatient visits don’t generally happen at 11 pm, so that would be an extreme fringe case).
  • Outpatient: the patient is not in a bed (stretcher is ok) vs Inpatient: Patient is in bed.
  • Outpatient: The patient comes to the provider vs Inpatient: the provider comes to the patient (bedside).

So, let’s make the Observation Room Visits a descendant of Inpatient.

(Michael Kahn) #12

How does @Christian_Reich’s proposed heuristic work with same-day ambulatory surgery? Seems to have conflicts with heuristic #1 versus heuristic #2.

(Christian Reich) #13

Which is why @Christian_Reich said that this is not exact science. But still: The patient comes to the surgery, leaves the same day, there is no bed. So, 8883 “Ambulatory Surgery Center” Isa 38004207 “Ambulatory Clinic/Center” Isa 9202 “Outpatient Visit”.