Yes, email is still the state-of-the art. I've added exploring GitHub back to my todo list. We're focused on gearing up for a public release currently, which will go on the i2b2 wiki. Email me for a link to our development version.
Update since the 9/22 notes - we have a CDMv3 transform for MSSQL now (Oracle forthcoming). Death, PRO, and Clinical_Trial tables are not yet supported. Thanks, Jeffrey Klann, PhD Instructor of Medicine, Harvard Medical School From: Dan Connolly [mailto:dconno...@kumc.edu] Sent: Tuesday, October 27, 2015 5:19 PM To: Stroup, Jeff; Klann, Jeffrey G. Cc: French, Tony; Allen, Katie Sue; <gpc-dev@listserv.kumc.edu> Subject: RE: CDM refresh +cc gpc-dev, with permission What I meant by the SCILHS network version builds on gpc-pcornet-cdm<https://bitbucket.org/gpcnetwork/gpc-pcornet-cdm> (#145<https://informatics.gpcnetwork.org/trac/Project/ticket/145>), but it's not the same thing. We're tracking evaluation of the SCILHS code as #347<https://informatics.gpcnetwork.org/trac/Project/ticket/347> (@MU)/ #381<https://informatics.gpcnetwork.org/trac/Project/ticket/381> (@CMH). Jeff K., as of our Sep 22 gpc-dev call<http://listserv.kumc.edu/pipermail/gpc-dev/2015q3/002055.html> when you joined to discuss it, you offered to share the code to anyone in our community who asked you by email, and you were thinking about other ways to share; is email to you still the state of the art? Jeff S., I'll excerpt relevant parts of that call and a later one: ________________________________ 3. Compare approaches to ETL around i2b2 and CDM i. Jeff - the process to get the CDM is to take data mapped into the ontology; consistent set of paths with flexible local codes. 1. SQL transform to pull from the database with that as the source. Looks up relevant basecodes; new transform column. 2. There's a lookup in the ontology table for pcori basecode; that gets plugged into the diagnosis table (etc.); not generalized, but automatic once the data is mapped into the SCILHS ontology. 3. Covers CDM v1, 2, 3; transform is only in place for CDM v1. Skipping to working on the CDM v3 transform. a. Internally labs work; meds in progress.data formats and auto-incrementing keys are updates between the versions. b. Will have this work done by mid-October * Can share with the PCORnet community. ________________________________ And from our 6 Oct gpc-dev call<http://listserv.kumc.edu/pipermail/gpc-dev/2015q4/002104.html>: ________________________________ 1. milestone:data-domains3<https://informatics.gpcnetwork.org/trac/Project/milestone/data-domains3>, drn-query-1<https://informatics.gpcnetwork.org/trac/Project/milestone/drn-query-1> a. #160<https://informatics.gpcnetwork.org/trac/Project/ticket/160> reproducible GPC, CDM refresh at a critical mass of GPC sites (PMO milestone 2.4)<https://informatics.gpcnetwork.org/trac/Project/ticket/160> i. DC: So the goal is to get a critical mass of sites using the same approach. ... ii. PR: My thought it depends on what GPC wants to do. i2b2 to CDM and get the process working well? I think that's fine. If not, everyone is feeling pressure to get data in CDM and figuring out a way to do it. So I'm looking for a GPC higher level decision. DC: As customer/reporter of #160, Russ asked us to go i2b2->CDM; in #145, we provided code to do it; the code has some holes, and the Harvard folks have filled some (all?) of them. Earlier Laurel put together a table of DRN status at GPC sites<https://docs.google.com/spreadsheets/d/1Lg5fAHV5_rBT_FyF4_8NDuZC7KlNapLai90LEkdnxpI/edit#gid=0>; I'm not sure whether she intends to continue to use/maintain it, but it shows several sites using the shared approach from #145. So i2b2->CDM is plan A as far as I know. iii. Susan Morrison: Are going back to do version 1 with the code from Harvard or not? Will we risk incompatible approaches? DC: I expect to evolve towards the direction of using Harvard's refinement of the approach in #145. Right now MU has the ball on evaluating it (the Harvard CDM code; cf #347<https://informatics.gpcnetwork.org/trac/Project/ticket/347>). We have the code from Harvard, it's just a question of trying it out. To get the code, the shortest path is the e-mail Jeff Klan. ________________________________ -- Dan ________________________________ From: Stroup, Jeff [stro...@regenstrief.org] Sent: Tuesday, October 27, 2015 3:06 PM To: jeff.kl...@mgh.harvard.edu<mailto:jeff.kl...@mgh.harvard.edu> Cc: French, Tony; Dan Connolly; Allen, Katie Sue Subject: RE: CDM refresh Jeff, Based on some links in a "GPC CDM ETL" pdf on an associated ticket #145<https://informatics.gpcnetwork.org/trac/Project/ticket/145>, I stumbled on what look like instructions here https://bitbucket.org/gpcnetwork/gpc-pcornet-cdm. I plan to start implementing these instructions so please alert me if these are not the "Skills Network" version that Dan alluded to. Thanks, Jeff From: Stroup, Jeff Sent: Tuesday, October 27, 2015 3:42 PM To: 'jeff.kl...@mgh.harvard.edu' Subject: CDM refresh Hi Jeff, This is in regards to GPC Ticket #160<https://informatics.gpcnetwork.org/trac/Project/ticket/160>. During our onboarding meeting, Dan Connolly suggested that we try out the "Skills Network" code for creating/refreshing our CDM data. Jim Meeks-Johnson mentioned you had emailed him a copy but he didn't share it with me and is no longer with RI. Can you please forward that email to me or direct me to a link where I might download it. Any instructions would be helpful as well. So far we know that it involves adding another column to the metadata table (PCORI based code) and then running some SQL. Thanks, Jeff The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
_______________________________________________ Gpc-dev mailing list Gpc-dev@listserv.kumc.edu http://listserv.kumc.edu/mailman/listinfo/gpc-dev
