OHDSI Home | Forums | Wiki | Github

Visit vs place of service

  1. Vocabulary Place of Service belongs to the domain Place of Service and is used to populate place_of_service_concept_id in the care_site table.
  2. Vocabulary Visit belongs to the domain Visit and is used to populate Visit_concept_id.
  3. Vocabulary Place of Service is mapped to Visit in both concept_relationship and concept_ancestor table (I dont know how complete this is @Dymshyts ?) Place of Service is a child of Visit
  4. There is a n:1 relationship between visit_occurrence table and care_site table - so there may not be more than one care_site_id per visit_occurrence_id record. As an extension, there may not be more than one place_of_service_concept_id per visit_occurrence_id.

Would it then be reasonable to add all the conceptā€™s in Place of Service vocabulary from Place of Service domain into Visit domain, as a standard_concept and maintain hierarchical relationship?

I wanted to ask - of the 59 standard and about 487 total conceptā€™s in the Place of Service domain, can we change their domain to visit - so that they may be used in the visit_occurrence table as either visit_concept_id and visit_source_concept_id?

The classification of visitā€™s into inpatient and outpatient is then natural using the concept ancestor table

1 Like

I am currently working with encounter data that contains scenarios like the following ā€“ A patient had a series of micro visits/encounters during a major encounter/visit_occurrence on a particular day. During this visit, the patient after seeing his/her PCP at the out-patient clinic had subsequent encounters at Cardiology (care site?/place of service?), Endocrinology(care site?/place of service?) and ER where he was admitted for observation. All these encounters occurred at the same facility/location for example, Emory Hospital on Emory University campus Atlanta, Ga. How will this scenario be represented in the CDM within the ontology described below? Any help will be greatly appreciated.

@kDarko:

You got a two-layer logic for this: VISIT and VISIT_DETAIL.

  1. You got 5 Visits: three (i) outpatient Visits at the PCP, the Cardiology Outpatient Department and the Endocrinology Outpatient Department, the (ii) ER Visit, where the patient was sent to the (iii) hospital Visit for the observation (not sure that is a hospital visit, some institution donā€™t count the observation visits as such). The Case Site for these are all Emory Hospital Atlanta.
  2. You got non or one Visit Detail: The observation hospitalization, with the specific Department at Emory Hospital, where the observation took place as Care Site.

Note: Outpatient and ER Visits are usually simple: The patient is in and out of these. Not much happens inside. In your case, there is a lot of outpatient activity and the funny ill-defined observation visit. This is not typical. Usually, the patient feels sick, goes to the doctor, who quickly decides that the situation requires rapid action and refers the patient into the ER. From there, the patient takes some career through the various hospital departments, for which is why we have the Visit Details.

Make sense?

We probably need to clean them up, Gowtham. Letā€™s find out from Themis Group 4 if they made some progress on this. @krfeeney and @mvanzandt?

Makes sense to me. @Dymshyts?

1 Like

@kDarko, if it helps: The fields in care_site that are analytically useful are the location_id (because we could use lat and long) and place_of_service_concept_id (because it has concept_idā€™s); rest of the fields in that table are not standardized and pretty much not useful for standardized analytics.

@kDarko,

You can use the care_site_id and care_site_name field in the care_site table to represent the PCP, Cardiology, and Endocrinology. All will have the same place_of_service_id to represent Emory hospital. And most likely the same location_id, but it could be different depending on the granularity of your address/location data. Care_site_id is:

but it may be useful for your internal purposes. If it isnā€™t, not sure you want to create a bunch of care_site_ids (and extra work) for your data.

Actually, I have doubts:
CARE_SITE
place_of_service_concept_id describes type of the clinic, i.e. Rural Clinic, Mental asylum, etc.

VISIT_OCCURRENCE
Visit_concept_id describes type of visit: in or out-patient, ER

So you suggest to populate place_of_service_concept_id and Visit_concept_id with the same concept, say place_of_service_concept_id = ā€œInpatient Psychiatric Facilityā€
Visit_concept_id = ā€œInpatient Psychiatric Facilityā€

instead of
place_of_service_concept_id = ā€œInpatient Psychiatric Facilityā€
Visit_concept_id = "Inpatient "
?

Can this change be confusing for the people used to ā€œin- or out-patient, ERā€ visit IDs?

Friends:

I think we are bumping here into a problem that came up the other day in the Atlanta hackathon discussion: place of service really denote two things:

  1. A type of service, like surgery, psych assessment, emergency, etc. (which is adequate for Visit).
  2. A type of provider organization, like Assisted Living Facility, Outpatient NEC, Inpatient Psychiatric Facility etc. (which is adequate for Care Site).

The Visit seems to need 1), not 2). So, from that perspective we need not combine them. Instead, we need to think of more granular service concepts.

I think we need to answer the question: why do we need the domain_id ā€œPlace of Serviceā€. Why canā€™t it be ā€œVisitā€.

I donā€™t think we need ā€œPlace of Serviceā€ domain_id, because we are mapping the standard Conceptā€™s in that domain to standard concepts in ā€œVisitā€ domain.

@Gowtham_Rao:

Because Place of Service doesnā€™t denote something that further qualifies a Visit, but a Care Site where the visit happened. And for that we have the table CARE_SITE.

I think it is needed for care site. Should this then be changed to a ā€œcare siteā€ domain?

For an ambulatory surgery, how would one code the visit and care site?

Yes, if we change the field name from place_of_service_concept_id to caresite_source_concept_id and caresite_concept_id. This is obviously not backwards compatible, but will be consistent with Omop CDM conventions. Maybe done for CDM 6.x

@christian_reich , why did place_of_service_concept_id break out of the visit table into care_site table when we moved from CDM 4 to 5?

Is ambulatory surgery, an

  1. Attribute of the physical place where care is provided.
  2. Attribute of the business/administrative entity for billing or documentation purpose i.e. the same physical entity can take on many place_of_service.
  3. Attribute of the (service) visit like a condition, procedure is an attribute of a visit.

All change can cause confusion, but the risk in this case should be very low because we are not deprecating existing concepts. Further, our proposal will have robust ancestor mapping - and the changed concepts will be child if the existing concepts.

A data partner may map to any any level of the hierarchy, and we can go up or down the hierarchy

I would say itā€™s more #3 and maybe #1. Right now, we classify ambulatory surgeries as outpatient visits since patients do not have a room; however, there are cases when the procedure might not go as expected and a patient has to be admitted. If we want to look at all procedures that might have this transition, itā€™s a little unclear to me. I can see going the #1 route and classifying a care site using OMOP code 8883 (Ambulatory Surgical Center), but I would think it would be nice to also have a new visit type for ambulatory surgery for easier analysis since from a practical point listing locations and their services can be difficult. I am curious to see what others think.

You hit the nail on the head @cukarthik ; the problem is that in most data sources, although it is called ā€˜place of serviceā€™ - it is not an attribute of the place, but it is an attribute of the service itself. In US health care, care-site is not a ā€˜credentialingā€™ information, but a ā€˜claimā€™. e.g. place of service information comes in every claim on CMS 1500 form.

1 Like

Thatā€™s not what happened. In 4, we had ORGANIZATION and CARE_SITE. That was the same two-layer hierarchy we have now for VISIT_OCCURRENCE and VISIT_DETAIL. We abandoned it, because there is no way this can be harmonized or standardized in any reasonable way. We got rid of ORGANIZATION, and left CARE_SITE, which is a physical place for any type of Visit to happen.

@Gowtham_Rao: Show them our proposed table.

We (@Christian_Reich @Dymshyts @Dave_Kern @Gowtham_Rao) worked on cleaning up these concepts, and would like the communities input. See document. We propose to change the domain for the vocabulary ā€˜Place of Serviceā€™ (standard concept) to ā€˜Visitā€™, because we found that ā€˜Place of Serviceā€™ is an attribute of the Visit and not attribute of ā€˜Care Siteā€™ as discussed here.

We then mapped (using concept_relationship and concept_ancestor table) the concepts in ā€˜Place of Serviceā€™ vocabulary belonging to the ā€˜Visitā€™ domain to other standard conceptā€™s in Visit domain.

1 Like

What is the definition for ā€œintensive careā€ visit?

Why is ā€œUrgent Careā€ an Office Visit with the definition: "Office Visit = ā€˜Patient goes to provider, but not inpatient, emergency room, ambulatory surgical center or other specialized unitsā€™?

Why is ā€œSkilled Nursing Facilityā€ labeled as ā€œInpatientā€ and not ā€œLong Term Careā€? Maybe I should ask the definition of Long Term Care instead?

These are services that are also called ā€˜critical careā€™, ā€˜ICUā€™ etc. Intensive care are for patients with severe and life-threatening illnesses and injuries, which require constant, close monitoring and support from specialist equipment and medications in order to ensure normal bodily functions (Ref)

From patients point of view, the ā€˜Urgent careā€™ does not lead to overnight stay in the hospital. The experience is like going to a Doctorā€™s office visit. (Ref) The main difference between traditional office visit and urgent care is a) no need for appointment, b) they will handle minor stuff like splinters, small wounds, acute infections, etc. that are not life-threatening emergencies.

In both Inpatient and Long Term Care, patient sleep in a healthcare facilities bed that is not ā€˜homeā€™. But the distinction between long-term vs short-term is not unambigious. How do we know that ā€˜Skilled Nursingā€™ was a long-term or short-term. When does a short-visit become long, 7 days, 30 days, 180 days, 365 days? Inpatient was thought to be the safest mapping.

1 Like
t