Hello. We're working on some things here at Houston and one of the other Devs had a question he wanted me to pass along and see if anyone had an insight. We were wondering if there's a "standard practice" for extracting encounter Diagnosis codes from Clarity to DIAGNOSIS, especially making sure we pull the codes that the provider intended, and nothing else.
Our original extraction code was based on an old example from University of Iowa. (attached) PAT_ENC_DX.ICD9 was used as the final DX CODE. Needless to say that was very outdated so we switched to a system that generated one row per encounter diagnosis in PAT_ENC_DX, joined to CLARITY_EDG, where CLARITY_EDG.REF_BILL_CODE_SET_C is 1 or 2. (ICD-10 or ICD-9) Then CLARITY_EDG.REF_BILL_CODE was used as the final DX code. Eventually we have changed to pulling the first code listed in CLARITY_EDG.CURRENT_ICD10_LIST, then the first code in CURRENT_ICD9_LIST, and then only use REF_BILL_CODE if none of those are available. I'm not sure this logic is correct, and it ignores any other mapped codes in CLARITY_EDG. So before we proceed with more changes I'm curious what other sites are doing. 1. Does Clarity have a canonical "ICD code" column populated for all hospitals? Or do we have to find out which table/column our hospital populates, or use case/coalesce statements to try one column and then another? 2. In the case of using CLARITY_EDG, should we be generating a DIAGNOSIS row for every code in the CURRENT_ICD10_LIST? Every other output column would be the same, and the data may change if Epic updates the code mappings in the future. Or is it one row per DX_ID? Thanks for any clarification. Appreciate any help. Brian Shukwit ######################################################################## To unsubscribe from the GPCDEV-L list, click the following link: https://po.missouri.edu/scripts/wa.exe?SUBED1=GPCDEV-L&A=1
pcornet_uiowa.sql
Description: pcornet_uiowa.sql
