I've merged the branches noted in the previous message into the respective master branches of the KUMC-BMI repositories -
https://github.com/kumc-bmi/heron https://github.com/kumc-bmi/i2p-transform I'm adding new work to the following branches - https://github.com/kumc-bmi/heron/tree/dispensing_5266 https://github.com/kumc-bmi/i2p-transform/tree/dispensing_5266 Though initially named for the dispensing task in trac, I plan to add all outstanding, priority CDM work to these branches. It makes for easier testing and the work has similar themes throughout. Work in the heron repository includes * Transfer of the "supplemental facts" needed for dispensing (i.e. dose, dose unit and route). Work in the i2p-transform repository includes * Raw preferred language in the demographics table * Dose, dose unit and route in the dispensing table * Present on admission in the diagnosis table As it stands, the dispensing code in i2p-transform relies on raw source system values for route. I'm not yet mapping these to CDM values. I plan to return to this after roughing in the other priority CDM tasks. I've not yet determined if dose unit needs mapping, as well. I've hard-coded "Y" and "Yes" in the new poafact table of the diagnosis.sql script, as our I2B2 ETL does not provide the source system codes for POA. It only provides a fact when POA is true. Changes to Heron and H2P mapping code can correct this, but as our ETL has already started, the work is on hold. If time permits, I can transfer the missing values as supplemental facts. From: James Gaupp Sent: Friday, July 06, 2018 4:58 PM To: 'phillip.ree...@utsouthwestern.edu' <phillip.ree...@utsouthwestern.edu>; 'todd.mcnee...@cerner.com' <todd.mcnee...@cerner.com>; 'gpc-dev@listserv.kumc.edu' <gpc-dev@listserv.kumc.edu> Subject: KUMC's CDM 4.1 Code By request, I'm providing an update on the status of KUMC's CDM 4.1 code. I've been working on CDM 4.1 tasks for a few weeks now but commits have been slow in coming, as much of the data pipeline and codebase are new to me. I've spent most of my time working on preliminary HERON (I2B2) updates. The code, with a pending pull request of new work, can be found at - https://github.com/kumc-bmi/heron/tree/insurance_and_payer_4687 (plus specimen source and ordering physician) I'm not sure who, among the recipients, has access to KUMC's Heron project. If the code isn't accessible, or is unclear, feel free to send me questions. In brief, the changes are - * Insurance data is being communicated through the standard I2B2 pipeline, from the Epic and IDX source systems, as observation facts. The facts contain the payer id as the part of the concept and payer name as a text value. * Prescribing provider is being communicated through the standard I2B2 pipeline as part of the root medication facts. * All other data needed by CDM, that's not already part of the I2B2 pipeline, is being loaded by Heron code from source systems into a new supplemental_fact table. At this point, the table comprises a de-identified instance number and source column text string (acting together as a primary key), plus a data item as a text string. o For example, the financial class of a payer is - 65678778284269129 "FINANCIAL_CLASS" "BCBS" The supplemental_fact table was created to (temporarily) take some of the pressure off the Heron ETL process which is realtively long running in its current form. Ideally, we'll be able to transfer supplemental facts in a fraction of the time, providing more dev time versus the CDM 4.1 deadline. I'm planning to add all other CDM data to this pipeline including the values needed for encounter, diagnosis, prescribing and dispensing. I assume the major issue for most other institutions working off KUMC's code is that the data is not delivered through the traditional CDM pipeline (no terms, no scilhs, no pcornet tables and no synonyms for facts and such). If there's an improved approach supported by the timeline, let me know. . . . My changes to the CDM common code in i2p-transform are modest at this point but include DRNOC approved mappings for specimen source and payer. They can be found at - https://github.com/kumc-bmi/i2p-transform/tree/lab_specimen_source_4815 https://github.com/kumc-bmi/i2p-transform/tree/insurance_and_payer_4687 https://github.com/kumc-bmi/i2p-transform (includes Oracle script updates for demographic.pat_pref_language_spoken and med_admin) I'm now working on dispensing, diagnosis, prescribing and encounter (facility) in that order.
_______________________________________________ Gpc-dev mailing list Gpc-dev@listserv.kumc.edu http://listserv.kumc.edu/mailman/listinfo/gpc-dev
