| OHDSI Home | Forums | Wiki | Github |
For me a consult is not related to overnight stay nor whether this is during inpatient. I can have a consult with any healthcare provider, e.g. I can have a consult with a physiotherapist to decide if we should start treatment etc.
In our country (and i think in many others as well), i do not always have to be “referred” either. I can decide to contact certain healthcare professionals and ask for a consult.
I therefore suggest not to make this a child of either inpatient or outpatient visit.
Would it be reasonable to create new omop concepts-- inpatient-consult, outpatient-consult/referral. We may also add many parents?
This is a different thing you are asking for, Peter. That’s a consultation. We have plenty of procedures for that. A consult is if one doctor is asking another doctor for help. In the inpatient setting, that other doctor usually comes to the patient.
“Referral” is an activity where one doctor sends the outpatient to another doctor, or to the hospital. Whether that is mandatory from an insurance perspective or not is irrelevant. I still don’t think we need an outpatient consult.
@Rijnbeek nailed it:
We have many visits within our EHR where a person has one and only one visit with a Provider. This “consult” visit is for the person to seek a 2nd opinion about a condition, course of treatment, etc. This is NOT a referral. This happens regularly for Providers at large academic centers and Children’s hospitals. People seek out an “expert”/2nd opinion.
i am not asking for anything.
It shows we need to be very clear on the definitions of these types, for example how could I have known this is only between doctors? I can show you examples where the word Consult is used as Visit and this is independent on whether this is coming from request of a doctor, e.g. I pay x amount per consult of my physiotherapist. This can be a language problem sure, but we will have 100 languages in OMOP-CDM. Also, do not miss-understand me, i do not suggest to use “consult” for this but we need to avoid ambiguity by being very clear on what me mean.
Even with this explanation i find the difference between consult and referral still vague (do we need this?). My GP can consult another GP in his group, is this then a referral or consult?
Maybe it would be useful if the EHR workgroup makes a list of concepts with their descriptions that are important or frequently seen in EHR data, and then the vocabulary workgroup could review and help by proposing solution?
There seem to be different understandings what a consult is. I was under the impression a consult or “medical consultation” is a request from one provider to another, not from the patient: https://medical-dictionary.thefreedictionary.com/medical+consultation or https://www.aapc.com/blog/38856-medicare-and-consults-99241-99255/. But from rummaging through the web I realize this may not be clean.
Why don’t we add three concepts:
The problem with 3) is that it is hardly distinguishable from a normal patient to doctor visit. But if you have that information we could add.
Another problem is that we need to draw the line between a Visit and a Procedure, e.g. 2514419 or 2514420. A visit is an organizational structure for providing care, not the content. So, I am not sure whether any of these consultations or referrals can be construed as Visits at all.
Let’s think.
A few concerns:
With the new structure, how do we determine what visits are Urgent Care Visits? In the current structure, we rely heavily on the place_of_service_concept_id (8782 - Urgent Care) in the care_site table to do this. According to the hierarchy, this is being folded in as an outpatient visit (9202). For most of our researchers, because this type of visit is associated with some level of acuity we consider it to be more of an emergency (9203) visit. From a provider perspective, there is no distinct specialty for Urgent Care. The provider may have general pediatrics or emergency care as their speciality designation. We would like to avoid using the visit_source_value to determine these things.
Overall, the idea of eliminating the care site specialty is troubling form a practical standpoint of how the data is stored in our EHR and the governance tied to care sites at our insitutions (for example to attribute the “Oncology Clinic” or “Hematology Clinic” to a visit), with no way of telling what specialty the provider was “acting” in at the time besides looking at the department and the potentially the service at the visit.
As it relates to the vist designations:
For the “Home” visit type does this include Social Work Visits? If so, will there be conventions to clearly state this?
For the “Case Management” At our organization we consider case management to be opened by a discharge planner with the intent of organizing the transition from inpatient to home care for our patients. However, it was stated on a call that this is supposed to represent administrative visits as well. Do we put Emails/Telephones encounters in this category? These are true encounters in our EHR which have providers and information pertaining to the clinical care.
As it relates to the EHR structure, consults are procedures as opposed to actual visits. We would have to manufacture a visit that represents a “consult” type becase these are included in the inpatient visit and occur as outpatient visits.
@bailey Please feel free to elaborate
I deal with EHR’s,
Patients at our primary care settings have a main provider ‘a family phisician’. When the familly phisician need another specialty field to chime in (for a diagnostic or therapeutic purpose) they order a consultation (to the specialty field, not to a specific person).
This ordering gets treated as any ordering of a procedure (not any different than ordering a ‘CT scan’)
the patient may then get a visit from a person/provider of said specialty.
which is to say that there are 2 sides of these consultations, 1) the fact of ordering it, 2) and the fact of doing the thing ordered (which happens to be a visit).
“the PROCEDURE_OCCURRENCE table contains records of activities or processes ORDERED BY, or carried out by, a healthcare provider on the patient to have a diagnostic or therapeutic purpose.”
and them both may be linked together by the visit_occurrence_id foreing key in the procedure_occurrence table.
as for our distinction of referals from consultations/consults, when doing a referal it is assumed the new provider will be the main phisician from now on. while a consult is the family phisician asking for diagnosing or therapeutic tasks to be defined or suggested from the other provider but the family phisician remains the main provider.
hope this helps.
Was waiting for you guys to weigh in.
Hm. Really? I have been in Urgent care clinics for a bronchitis and a otitis. Just didn’t want to wait 3 weeks for a Primary Care Provider appointment. Was acute, but not an emergency. The 9203 really is Emergency Department, Emergency Room or Accident and Emergency (in the UK) visit. We all know this service gets abused for all sorts of reasons, but the idea is that this is the place where the ambulance takes you. Does that make sense?
But I can be persuaded and we roll it up to 9203 if folks think this is better.
I realize that. Here is how you would do that: You would assign visit_concept_id = “Inpatient hospital”, and 38004507 “Medical Oncology” and 38004502 “Hematology/Oncology” to the provider_concept_id. Would that work?
The provider is acting based on the department (s)he is in? And you have data to support that?
I happen to be a Hematologist/Oncologist, or I used to be one when I was actually practicing. These demarcations are somewhat arbitrary, and it depends on the patient’s case. An oncology patient can have a paraneoplastic syndrome with anemia, and you are in full hematology mode, even though the primary is in the colon.
I think the specialty is very clearly defined by a degree from some institution.
Well, the rule for deciding home or not was “Does the patient come to the provider, or does the provider come to the patient?” Isn’t always clear.
Social Work appears more like a service from somebody, rather than a visit (organizational constellation of how healthcare is provided).
Well, here is the thing how I think it would work: The actual consult is just an assessment. Doesn’t matter if it happens as part of a consult (another provider called you in) or because the patient showed up in the waiting room. The fact that the provider was called by another one and comes makes it a consult visit. Yes, you would have to create a visit. Not sure how much there is “manufacturing” involved: You write a VISIT_DETAIL record, and you put an assessment procedure in.
@imendezboo:
The problem is that we probably won’t be able to agree on what consult or consultation really means. The choices are as above:
We should allow all three. Depending on the organizations use of the jargon term “consult” folks can use whatever they have in mind. We could use SNOMED’s
4014829 “Consultation”
4144684 “Patient referral”
But that doesn’t help us, because it is important what the consulted or referred to provider does. Not what the consultation seeking or referring provider does. We have all the Visits we need.
That seems to be the only place where consultation or referral becomes relevant. Currently, we don’t have an ability to characterize how a visit was initiated in the VISIT or VISIT_DETAIL tables. We may want to do that together with overhauling the admission and discharge concepts.
Don’t think so. Procedures are activities on the patient for diagnostic or therapeutic purposes. Not for shipping the patient around between different Visits.
@mmatheny
We are working with the VA Information Resource Center on transforming the CMS Medicare/Medicaid data to OMOP. To your point above about a “physical” place of service, some of the claims files were are converting do not have an explicit place of service. For example, the CMS home health agency revenue data file does not have an explicit care site. Although care may have been at the patient’s home address of record, it could alternatively have been at the patient’s winter location (snowbirds), grown child’s home…the claim data does not specify. The “agency” (submitting the claim) is the best proxy we would have for care site. After much discussion on how to fit the data into the model, we are are planning to create a concept for “agency”. This is also the case for hospice claim files.
The hospice and home health files also bill claims for a period that may have multiple “visits.” Peeling out the plan for how to load this detail into OMOP is equally challenging. We plan to link the visit occurrence level to the claim ID and populate the visit detail with the data defined by the revenue codes and dates of care populated for the “claim period”.
We recognize there are multiple ways to model this type of data but want to provide the right level of detail available from the source to meet the needs of researchers. We welcome feedback.
If I read this correct you got this right. And we already have that “agency” concept: 38004519 “Home Health Agency”. It will be rolled up to 581476 “Home Visit”.
What do you need there? We got 8546 “Hospice”, which is going to be rolled up in the hierarchy to 42898160 “Non-hospital institution Visit”.
And we should create standard conventions. Thanks for helping.
Greetings all - First time caller, I’m Robert Winter a Technical PM on the AoU project doing Curation. I also have a separate pipeline that I manage here at Vanderbilt.
A couple of thoughts RE: Urgent Care - Urgent Care should be a Standard Concept as more and more facilities are moving to the less expensive models in an attempt to avoid costly Facility fees. Rolling UC into Outpatient would make it more difficult to pull apart the UC to ED, UC from outpatient, UC to IP, etc. In terms of cost and service there is a clear separation between normal outpatient services and UC, and UC with ED or IP.
Also, perhaps this is discussed in a different topic, but I think it’s important because the practice is changing fast… Telehealth should be its own Standard Class of Visit. I don’t know that I would separate Video from Telephone, but broadly speaking a physically separated care experience will become more and more common, and outcomes from those experiences should be able to be looked at in research exclusively without lots of chart digging.
Welcome to the family.
Don’t know if you are fully familiar with the new hierarchical structure. You will code the Visit as detailed as you can. However, all concepts are in a hierarchy, and if folks want a summary report concepts with higher granularity would roll up into the 10 or so main concepts:
Makes sense?
I am confused around the hierarchy…based on what you are saying, VISIT Visit is the highest level with VISIT Place of Service being under that (Outpatient Visit - Urgent Care Place of Service) (XYZ Visit - Telehealth Place of Service) is that correct? We are seeing both Telehealth and UC Facility as Place of Service, and UC is mapping to Office Visit (not Outpatient) which would be the top level based on what we are seeing. I’m looking at Athena though, and it’s a bit confusing. Is there a better resource?
I also need a little clarification And the VISIT OCCURRENCE conventions need clarification
We went from 5 standard visit concepts (IP, OP, ER, OP>IP, Long Term Care) where domain_id = ‘visit’ to 12 standards where domain_id = ‘visit’ and concept_class_id = ‘visit’ (new convention which adds concept_class_id to mapping requirement & needs to be added to the conventions, along with the fact that there are now more than 5 standard visit_concept_ids). In the 12 standards we now have visits that folks needed (Laboratory, Pharmacy) and visits that don’t make much sense: ‘Office Visit’ isn’t that an Outpatient Visit? And ‘Intensive Care’ is definitely an Inpatient Visit. Shouldn’t Office Visit & Intensive Care be children of their respective parents?
That’s why we started the post.
Yes, Ma’am. It’s parent is 9202 “Outpatient Visit”.
Correct. And so it is. The parent is 8717 “Inpatient Hospital”, and the parent of that is 9201 “Inpatient Visit”.
Neither one is part of the 12 top concepts. See my initial posting.
Not quite. “Place of Service” is no longer a domain. It is a vocabulary in the Visit domain. So is “NUCC” and “CMS Specialty”.
Not quite. There is no top concept called “Visit”. The top concepts are listed in the initial posting (scroll all the way up).
As Visit domain concepts…
Correct…
No. Office Visit itself has the parent 9202 “Outpatient Visit”. It’s iterative, hierarchical. You have to go up until you find no parent. Then you reached the top.
It’s not there, yet. We are rolling it out as we speak, and then you can look at it in Athena. (That may be the reason for your confusion, and @MPhilofsky’s confusion). Use the Google Doc listed above.
Correct, this is the reason for my confusion.
To sum up the recent discussion:
Athena has new standard visit concepts, some are NOT correct. (i.e. Office Visit & Intensive Care are standard concept_ids in Athena today, but will not be once the Visit domain has been refreshed by the Vocabulary team)
@Christian_Reich’s Google spreadsheet & original post list out the concept_ids to use for visit_concept_id. These include: inpatient, outpatient, ER, ER to inpatient, Non-hospital institutional visit (previously knows as ‘Long Term Care’), Laboratory Visit, Telehealth Visit, Pharmacy Visit, Home Visit, Transportation Visit, Rehabilitation Visit and Case Management Visit. All other visits will be source codes & will map to non-standard visit_source_concept_ids
The Vocabulary team will create parent-child/is-a relationships in the CONCEPT RELATIONSHIP & CONCEPT ANCESTOR table.