Meeting Minutes-DSTU-Testing Team 5-20-2010
Return to main RPS R2 DSTU Technical Team Page
Harv Martens, Extedo; Lenore Palma, Datafarm; Sriram Thiruvengadam, Datafarm; Jennifer Denomme, i4i; Matthew Lukela, Otsuka; Tracy Naughton, AstraZeneca; Bill Friggle, Sanofi-Aventis; Joel Finkle, ISI; Ikram Baig, Virtify; Bernadette Billet, Liquent
Technical Topic Prioritization and discussion
The team reviewed the list of outstanding technical topics that had been circulated for prioritization. The prioritization with approximately half of the team responding is: 1) Controlled Vocabularies; 2) Use of SetID; 3) Lifecycle; 4) 2-way communication; 5) Submission units across applications/submissions; 6) High-level metadata. The team agreed that the controlled vocabulary topic will be a longer term discussion. There was agreement that a consistent format for structuring and delivering the controlled vocabularies would be the preferred solution and will require interaction with the controlled vocabulary team. In the meantime, we would like to focus the next data model walkthrough around use of the SetID element and managing the lifecycle in the RPS XML. During the initial walkthrough there was significant discussion on this topic, but no final resolution on the proper use of the various ID values used to manage lifecycle. Clarification on the rules of this is needed.
Test Case Availability
The team asked if more structure/organization could be given to the test cases on the wiki to better group and identify the information. Additionally, including the date when the documents are posted directly on the page will assist in identifying new information.
Joel Finkle indicated the ISI has developed a DSTU-specific non-disclosure agreement that will be used for anyone wishing to partner for testing. They are targeting mid- to late June for tool availability and have one customer committed for testing. Harv Martens from Extedo indicated they are waiting on final internal decisions to announce further information, but that should be coming in the very near-term. Lenore Palma from Datafarm indicated they will provide a client-based tool; a specific timeline for availability is not yet available. Bernadette Billet from Liquent indicated they are waiting for finalization of dates, but targeting late June. Ikram Baig from Virtify and Jennifer Denomme from i4i both indicated they are in the process of determining dates for their tool.
The team briefly reviewed the sample RPS XML messages provided by Jason Rock. It appears that the samples provided were created against a version of the schema earlier than the one that was balloted and distributed. This resulted in some old element names being included, such as <referencedBy1> and <refernecedBy2> for keyword and file references which was consolidated to just <referencedBy> in the DSTU ballot. The team is asked to review the samples and send comments/questions to Bernadette by Wednesday, May 26. The consolidated list will be forwarded to Jason Rock for clarification/discussion. In some cases the draft implementation guide may provide the answers to these questions, in which case we will include the implied understanding for confirmation. The team did note that a number of the HL7-derived mandatory attributes will have fixed values for RPS messages, such as moodCode and classCode. The allowable value lists for these attribute vary depending on their parent element, therefore the implementation guide should explicitly list their use and expected values for RPS messages. The final discussion focused on the concept of Reviewable Unit. The team’s understanding is that this provides a way to group related context of use elements together, independent of keywords, for example all components of a clinical study report. This is believed to provide a referencing mechanism to allow another submission unit to just reference the relevant reviewable unit rather than the full submission unit. Neither the current controlled vocabulary list, nor the test case scenarios include reviewable unit information, although it may be explored as we get further in the testing process.