Friday, June 1, 2012

Mapping MOH 257 to CIEL/MVP

I mentioned early on that we want to use the MVP/CIEL concept dictionary, because curating concept dictionary is a lot of work, and we should really be leveraging existing resources and expertise.

IntelliSOFT is just coming on board the project, and they're going to start by mapping the MOH 257 data elements to the MVP/CIEL dictionary. First off, you guys should say hi and introduce yourselves. :-)

It would be great to know if someone out there has already done the work of mapping MOH 257 to MVP/CIEL (or to another OpenMRS dictionary like AMPATH, or to SNOMED/ICD, ...) who could share that work with us. Anybody?

Anyway, I spoke with Andy Kanter today, and he suggested a process for how you should go about the mapping process. After typing it up in a private email, I figured it would be useful to post here, after the jump...

High-level points:
  • We shouldn't just send MVP/CIEL a list of the missing concepts, but we should sent them a spreadsheet that also includes the concepts we've already found, so they can comment on whether our mappings seem right or not.
  • We should send them regular updates of our incomplete work, and not send them what we consider to be the complete request at the very end.
  • It is best to use a structured spreadsheet, and he's provided an example, which (I don't know how to attach here)
  1. To start, we should collect all of the concepts that we want to map, filling out the "Actual text on form" and "Other text" columns in the spreadsheet. (Possibly also Class and Datatype.)
  2. Look through instructions/documentation for the forms, and put any supporting information in the "Other text" column, and if available, SNOMED CT or ICD 10 maps.
    ---It ends up more accurate if we separate the above steps from the next one---
  3. Take that entire spreadsheet, and use mcl:search to look for matching concepts that are already in the dictionary. For any that are found, fill in the "CIEL/MVP ID" and "CIEL/MVP Concept Name" columns.
  4. At this point (and actually several times before this point) share progress with everyone so we can make any necessary adjustments early.
  5. Ensure that Class, Datatype, and Submitter question/comments are all filled out for the remaining unmapped concepts.
  6. At this point (and several times beforehand) share progress with everyone, and we can decide whether to start creating the missing concepts immediately, or wait for feedback from CIEL/MVP.


  1. Thanks Darius for the intro and support. Hi :-) I am Peter, a BA with IntelliSoft Consulting, as mentioned we just yesterday dug our teeth into the task of mapping the MOH 257 data elements to the MVP/CIEL dictionary, the template and process guideline have been useful - we will share progress later today. Exciting stuff!

  2. My name is Steven, a senior technical advisor in biomedical and health informatics for IntelliSOFT.We are currently knee deep mapping MOH 257 to the available listed dictionaries, and as Darius pointed out, it would be useful to hear from someone (especially AMPATH team) anyone who has already done this - but in the meantime, we are working on it.

  3. I have mapped 2 pages of the MOH 257 to the CIEL/MVP and AMPATH dictionaries on the MCL and it is available on the Kenya OpenMRS DropBox.

  4. Steven, I'd really like to see this work!

    Can't find where you've put this.

  5. Hey Paul, Jan put it in the "Dropbox for OpenMRS Kenya work" and she can give you access to it.

    I got very useful feedback from Andy for that first iteration so now working hard at refining, updating it and posting.

  6. Darius, could you put these strategies in this OpenMRS page that I created for a similar need

  7. A big welcome to Steven and Peter! Looking forward to working with you more as this interesting piece of work progresses.