Ontology Mapper Minutes

7/22/08

Present on call: Hillari, Marco, Paul, Ketty, Rob, Prakash, Davera, Mark, Maggie, Larry Errecart (please welcome Larry from UCD)

In the Minutes section you will find that redline assignments on Maggie’s paper are marked in Red .

IMPORTANT ANNOUNCEMENT :  In response to RFP-08 two grant applications were filed both of which involved the usage of the Ontology Mapper.  The grant submitted by this team was unfortunately rejected.  However, the grant application submitted by the University of Washington for i2b2 is moving forward.  That i2b2 grant also involved the deployment of Ontology Mapper but only within the context of i2b2.  Since our team has already focused our efforts solely on i2b2 this is not a bad thing.

In light of these developments I have the following recommendation.  Those tasks which may be covered by the University of Washington i2b2 grant should be delayed in the hope that the U. Wash grant receives funding as planned.  Our team should attempt to complete all other high priority tasks on OntoMapper by the end of August.

What does that mean for the Ontology Mapper team?  The University of Washington still wants to collaborate with our team.  So at the beginning of September we would see an expansion of the Ontology Mapper team to address the issues below.  I will remain as the project director of the Ontology Mapper effort specifically with Nick Anderson at U. Washington directing the overall grant and many other i2b2 tasks (such as the new i2b2 Federated Query).

Specifically the scope of work for the U Wash i2b2 grant lists these items as the responsibility of UCSF:

1)       Development of the I2B2 Ontology Mapper Wrapper Cell.  This I2B2 module will allow I2B2 implementations to make use of the CTSA Ontology Mapper server that will be developed at UCSF.  This cell will provide advanced terminology services and data translation for the I2B2 platform.

2)       Development of software quality assurance regression tests for the I2B2 platform.  The I2B2 project has several CTSA related initiatives under development.  A coordinated suite of tests is required to ensure that any changes made to the I2B2 framework are all compatible and therefore can be used within any of the CTSA sites.  UCSF will be responsible for the creation of a regression test suite that can automatically and quickly test the I2B2 platform relative to a baseline of functionality.

3)       Data governance for the purpose of sharing data for cohort discovery.  UCSF will seek internal approval for the sharing of de-identified data between our 3 institutions for the purpose of cohort discovery.  Once approval for the inter-institutional sharing of data has been achieved researchers will be able to perform cohort discovery across patient populations within all three organizations.

So tasks which may be funded by the University of Washington i2b2 grant:

1)     The i2b2 integration of Ontology Mapper into an i2b2 Cell.

2)     A regression test suite for i2b2 and the Ontology Mapper Wrapper Cell.

3)     John Gennari’s tasks are still tentatively possible:

1. The creation of a Protégé Prompt plug-in to make mapping rule authoring easier (basically a user-friendly UI that can generate Prakash’s mapping rule xml files.
2. The development of alternate transformation languages (other than our current language which is SQL), (perhaps SWRL or Horn?)

So the Ontology Mapper project should focus on (by Sept 1 st ):

1)     The research lifecycle paper and its Use Cases

2)     The java design and coding of the Data Discovery and Data Request interfaces

3)     Getting OntoMapper running on actual rules derived from those Use Cases

4)     So basically we should get Ontology Mapper running with just SQL as a transformation language as a stand-alone system that uses the i2b2 schema.

Minutes:

1)     We walked through Maggies document over a webex session.  We decided to use the term BioMedical Research.  A very active discussion followed…

2)     Mark mentioned that the bridging of information from different systems is made easier by our system.  We should include that in the paper – Mark please suggest redlines for that.

3)     Davera mentioned IRB waivers – all of that workflow current falls under a single diamond on the diagram.  Should we include more detail about IRB oversight?  Is a workflow for IRB oversight within the scope for this paper?  Does this constitute a relevant gap in i2b2 that should be addressed? – Rob talk to Maggie

4)     Data Collection Phase may go away in future state – Mark – Also Paul has mentioned that he is compiling a list of research types that can be done completely within the IDR without full IRB approval (building on Mark’s earlier works).  Paul and Mark please provide redlines for this topic.

5)     Paul – talked about Patient Intervention – under clinical trial – imported from CTMS?  At UCSF we are – Maggie does this have relevance?

6)     Mark talked about the Retrospective vrs. Prospective perspective on data.

7)     Patient IDs, Sample IDs, etc make connections in data very hard – we are fixing that problem completely – Maggie is this relevant to the paper ?

8)     Maggie , please fix document date and add authors list sorted by amount and relevance of input – please just use your best judgment on that.   Please add people of the authors list of they have added a paragraph or more of text or if they have had a significant impact on the papers content.

9)     Ketty - Add current challenges to the blank section below Figure 3.  This is both a need for further explanation on the diagram and it also really will help with our page breaks.

10)                        Davera – Use of term quality – Davera - Mark please supply a paragraph on this topic

11)                        Importance of quality of care data and Quality Process Improvement (cost lowering) – Rob and the Marshfield clinic have been working on an i2b2 UI for this.  As it turns out Marco is also already doing this.   Both Davera and Ketty are interested in this topic and we have experts at UCD and UCSF on this topic.  We decided to mention this in the paper but to address this in a follow up paper.   So we just need to identify this as a gap in i2b2 but not address it as part of this paper specifically. (note: we are not talking about an operational system here) – Maggie, Paul, Davera and Rob to work on this?

12)                        Hillari mentioned bad data – what do we do if we discover bad data within the IDR?  Do we report that back to the operational system?  How do we do that?  How is the data subsequently marked as bad within the system to avoid the re-import of the data before it is fixed? – This is a gap – We need CAPA – Maggie and Rob to talk about this.

13)                        Do we need more work on the Future research diagram?   - Maggie ?

14)                        Data Request UI – screenshot needs update – Maggie

15)                        Section 5 needs to be beefed up.

Okay everyone, great meeting and please send Maggie redlines!

Rob Wynden