E-mail
TSC, May 12, 2010
Attending: Joel Farrell, chair; James Fiore, Peter Greene, Steve Kinney, Dan Rehak, Carl Singer, Valerie Smothers,
Recap of annual meeting
Joel summarized that it was a more academic conference than in years past. They were more universities present and the content was more oriented towards medical school education. There was a strong turnout of people working on the virtual patient standard as well as those working with virtual patients in general. There also topics on Web 20, including sessions on the mEducator project. They have a social networking site to share educational artifacts including virtual patients. There are relationships between objects that describe how objects get reused. In addition, the virtual patient working group is determining their next steps. Joel asked Peter if there were further developments.
Peter commented that there is an interest in exchange of data between VP systems and physical simulators of various kinds. Some work has been funded; Rachel is the lead on that work. She presented some information, a loosely coupled framework for data interchange. That's of great interest. The dental simulator company at the meeting was very interested. That is increasingly being looked at – coupling of vps and simulation. We'd love to get this group's help on the topics that come up there.
Joel commented that he talked to Rachel about that as well. Having an interface between virtual and physical simulation could open opportunities for high fidelity simulation environment. Joel added that eViP was cosponsor of the meeting, but they are winding down. Peter added that many reports will come out of the eViP group that inform future development. mEducator is exploring opportunities for Web 2.0 and semantic web, wondering how that crosswalks with healthcare lom. Joel and Peter talked to the leadership of that group. We want their leadership to present their latest ideas. Joel agreed we'd set that up.
Peter added that the virtual patient group had an interest in describing pathways and sequencing through a virtual patient. It has been looked at in high fidelity simulation, but the national Board of medical examiners center for innovation has expressed concern that we don't know what was looking for from competency assessment perspective. There is a lot of data to manage, and it's not clear how to get out data that is valuable.
Joel agreed; there is an opportunity to use virtual patients for clinical practice research. Peter added that one of the Dutch investigators of overlaid a belief network. Comparative effectiveness research is hot area right now.
Joel asked what the plan for next year's meeting was. Valerie commented that tentatively the conference is scheduled for May 9-11 at Johns Hopkins University.
Dan commented that he was meeting with White House staff tomorrow to discuss the use of semantic Web technologies for educational technologies. He asked that the Valerie sent him more information on mEducator. Valerie agreed to send that today.
Outlook for tsc for coming year
Joe commented that there may be new work for the technical steering committee to monitor. He asked if Rachel would come to the committee with ideas. Peter agreed that was a great idea. It would be good to have her provide some background on HLA and other background on her work. Someone from that group will draft a spec; he believes the exec committee would have the virtual patient working group work on that.
Joel agreed that could be important. He added that the virtual patient working group had mentioned work on profiles. They may need some consultation on what constitutes a profile, and the technical steering committee could help.
Another item for consideration is leap 2A and the educational trajectory. Joel suggested we reviewed that on the future call.
Finally the common attachments approach will require talking to the working groups. Joel spoke with the professional profile working group. We will try to nail down a common approach in the coming months.
Long description in comp wg
Valerie explained that the questions arose as a result of Tufts implementation of the competency framework specification. There dental school has competencies they consist of multiple paragraph descriptions. For example, a competency entitled professionalism and ethics may have a two paragraph description: one paragraph describing professionalism and another describing ethics. That single competency would then have several sub competencies. But the sub competencies do not neatly fall under either professionalism or ethics. This reinforces the idea of professionalism and ethics as a single competency.
They would like to have a way of encoding multiple paragraphs in the description.
Dan commented that he is not aware of this coming up elsewhere. He has seen people put long descriptions in metadata, but not with markup. Since it's a text string, it can be encoded, but there is no requirement to parse it. As a best practice, we should leave markup out of the description. He added that there is nothing wrong with having multiple descriptions, but the intent for that functionality was to support multiple languages. It is not widely used.
Valerie asked if they should have to description elements in their own system and then concatenate on export.
Dan commented that it would be simpler if they had two new lines. He recommended encoding two new Unicode characters for line breaks, one for each paragraph.
James warned that some parsers would omit the white space. Joel commented that you can tell the parser to accept the line breaks.
Common attachments approach
Joel reviewed the professional profile working group requirements. They need a way to attach binary objects to XML. Some elements in the XML would specify binary data, usually existing in a database. They need to keep track of data in their own database and when they send it. He could go in as base 64 encoding or could be an attachment, as in MTOM Web services, which uses mime multipart encoding. Or you could have references in the XML that resolve to those binary objects. The group was originally planning base 64 encoding, which is extremely inefficient. They could use Web services, but they may not want to be tied to Web services.
James recalled that the TSC recommended embedding resource locators.
Joel agreed that is another option. I URL or other identifier would allow the resource to be retrieved from an external site. When the document is imported, the system could change it to a reference in your own system.
Valerie reviewed the possibilities:
- In line data
- Mime multipart encoding
- Reference to shared location
Joel added we would have to look a little more at some issues. For example, is mime appropriate outside web services? If you are using soap Web services, you would use them comp for attachments. But if you are exchanging data using another mechanism, reference a shared location.
James added it could be a relative file location if you sent out the XML in a zip file. He added that in the line data increases the risk of corruption.
Joel agreed we may not include that first option at all. Just describe mime and reference to shared location.
Joel asked what had happened with professional profile Web services. Valerie commented that the AMC may be using it. Peter added that certifying boards are uploading data to the ABMS using these Web services. Peter added that evidentiary data is important and that coming up with a lucid way to exchange attachments is important as well. The simpler the better.
James agreed. He mentioned that the board of surgery uses rest URLs and blocks things by IP address.
Joe commented that SSL may be sufficient. We don't want data stolen or corrupted. The scenarios are simple, and rest is probably sufficient for most environments, and it's not too different from uploading data. The http endpoint for the rest service could be the same thing called by the a form doing upload. We may not need to elaborate a lot on the Web services approach at this point. We may want to detail the rest approach, security ramifications, etc.
The committee agreed to continue this discussion on the next call and schedule a mEducator presentation.