OHDSI Home | Forums | Wiki | Github

Discharge Status and UB04 Pt dis status

(Nadya Zvyagina) #1

Hello, We have faced the problem with discharge statuses. The problem is that we have the list of discharge statuses in our source data and their description.
The original plan was to use ‘UB04 Pt dis status’ vocabulary to get concept_id’s. But when we compared the descriptions provided in our data to the descriptions in the vocabulary we found several discrepancies.

  1. 21
    UB04 - ‘Discharged/transferred to Court/Law Enforcement’

  2. 81-89
    UB04 - Contains meaningful descriptions of these codes, for example ‘Discharged to home or self-care with a planned acute care hospital readmission’

Thus if we simply join this vocabulary we will change the meaning of original data.

The same picture is for ‘UB04 Point of Origin’.

What can we do with this?
@Gowtham_Rao , I heard you are an expert in this topic)

(Gowtham Rao) #2

Hi - could you please clarify.

Is the concern that the concepts for this vocabulary are incorrect in Omop vocabulary, OR

You are unable to map the proprietary vocabulary in your source data to omop vocabulary.

(Nadya Zvyagina) #3

I am mostly trying to understand where is the truth, in vocabularies or in source data?

Just to clarify, I was going to join to vocabularies like this:
source join concept ON source.dstatus = concept.concept_code AND concept_vocabulary_id = ‘UB04 Pt dis status’
Should I include any other conditions?

(Gowtham Rao) #4

Here is the UB-04 Patient Discharge Status Vocabulary in OMOP, and here is some education on UB-04 patient discharge status codes.

The first step is to determine what vocabulary does your source data use to code discharge status. Is it UB04 patient discharge status codes?
Second step would be to a frequency distribution of the codes in your source data – check if they match with expected values in UB04 patient discharge status vocabularies codes. Check for number of characters, are there leading zeros? If you would like automate these two steps using a tool then use USAGI here.

If your source data vocabulary matched UB04 patient discharge status - you can use this query below to get the 'concept_id’s - it is straight forward text join.

source join concept ON source.dstatus = concept.concept_code AND concept_vocabulary_id = ‘UB04 Pt dis status’ and concept.invalid_reason is null

don’t forget the invalid_reason.

If your source data vocabulary does not match UB04 patient discharge status – then USAGI will most likely offer match recommendations, i.e. your source data is not using UB-04 patient discharge status vocabulary. In that case, you may want to use source_to_concept_map to doing mapping.

(Gowtham Rao) #5

I am tagging @Dymshyts and @Christian_Reich - please see the education link. Thelast page has education on ‘when to use each code’/frequently used codes. Should we try to incorporate such ‘code descriptions’ into concept_synonym?

(Christian Reich) #6

Yikes. We need to clean this up. It’s all over the place. We have “Discharge Status” concepts, but they are mapped to Observations. Will make a proposal soon.

How urgent is it, @nzvyagina?

(Gowtham Rao) #7

these are the concepts in Athena . Its in the domain VISIT. I think they should be used for visit_occurrence.discharge_to_concept_id or visit_detail.discharge_to_concept_id

(Christian Reich) #8

I know, but there are concepts that have been included before you dropped the UB04 stuff.

(Nadya Zvyagina) #9

My source data matches UB04 in most cases (both codes and descriptions match perfectly), the only problem is non-matching descriptions of some codes.
BTW: As for 81-95 there is information here
The following codes apply only to particular MS-DRGs*, and were new in 10/2013:
Comments: * MS-DRG codes where additional codes were available in October 2013 are:
280 (Acute Myocardial Infarction, Discharged Alive with MCC),
281 (Acute Myocardial Infarction, Discharged Alive with CC),
282 (Acute Myocardial Infarction, Discharged Alive without CC/MCC), and
789 (Neonates, Died or Transferred to Another Acute Care Facility).

The only unclear point left is what to do with 21.
Description in our data - ‘EXPIRED TO BE DEFINED AT STATE LEVEL’
UB04 - ‘Discharged/transferred to Court/Law Enforcement’

And ‘EXPIRED’ and ‘Discharged/transferred’ they are pretty different in meaning.

(Nadya Zvyagina) #10

Hi Christian,

It does not look urgent because I am going to put these discharge statuses to visit tables to discharged_to_ fields. And I do not need any corrections so far.

(Gowtham Rao) #11

What are the non matches

That would mean that your source data is not compliant with this vocabulary. I would use the concept with closest meaning