 |
 |
Howdy Jacob, One thing you'll want to consider in choosing a vocabulary is to find one that's optimized for purposes/topics similar to yours. For example, SNOMED is designed to provide standardized terminology for storing/retrieving information from clinical care EHRs, ICD-10 is for reporting out for statistical purposes, and MeSH was designed for indexing articles, books, documents, etc in the life sciences. There are other vocabularies designed for other uses. These three schemes all have hierarchical and/or synthetic codes in the background (with ICD-10, the codes are actually in the foreground with the vocab in the background) that lend themselves towards relating concepts in an internationalized database of resources in many languages -- SNOMED specifically has an international version. There are good tools for these schemes (you'll even find ones that assign terms based on textual analysis), and it's easy to get the data so you can create your own tools if needed. If the idea is that that the values of Target Populations are kept separate from Technical Area and Expertise Area? If so, you might consider using more than one vocabulary or even to use your own hierarchy. For example, you may find that no vocabulary expresses Target Population the way you need and there probably few enough terms that it would be feasible to maintain your own hierarchy for that area. kyle On Tue, Mar 15, 2016 at 6:51 AM, Jacob Ratliff <[log in to unmask]> wrote: > All good questions, and most of which we are still in the process of > determining. The types of documents are generally project related > documentation (reports, plans, technical information, etc.), but those have > not yet been standardized. We are also in the process of doing business > analysis and user testing to determine the types and amount of metadata we > need. I just know for sure that we will need "subject" or "topic" > descriptors for all of the content. This has currently been narrowed down > to "Technical Area" (i.e. the actual health issue), target populations > (e.g. Newborns), and Expertise Areas (e.g. Monitoring and evaluation), but > all of that is subject to change based on research and user testing. > > As for the process to assign terms and the structure of the terms, all of > that is up in the air right now as we have no systems (business or > technological) to implement any of this. In an ideal world flush with > resources, I would be able to get a metadata/vocabulary management tool, as > well as a robust document/content management system that can work together, > as well as integrate with our other data and business intelligence systems > that are being created. There are currently a number of evaluative/business > analysis workstreams moving forward to try and answer some of these > questions. In reality, there is a very good chance a lot of this will need > to be managed through excel and good governance. > > The systems and vocabularies will also have to useful worldwide, as 80% of > our employees are located outside of the US (we aren't even talking about > multiple languages right now; that's a problem that is not even worth > considering at this point). > > The suggestions everyone has given so far are very helpful and are putting > me on the right track. My hope is to be able to use one of them, or at > least part(s) of them to get to where we need to go. > > Hopefully in a few months I will have a good update (and more questions) on > where we are! > > Thanks, > > Jacob > > On Mon, Mar 14, 2016 at 4:55 PM, Kyle Banerjee <[log in to unmask]> > wrote: > > > Could you say a bit more about the documents you need to manage, the > level > > of specificity you need, how they'll be used, and what process you > envision > > to assign terms? If your documents are mostly clinical in nature, SNOMED > > strikes me a good choice, but if you want terminology that could take you > > to related articles in PubMed or your needs aren't mostly clinical, MeSH > > might work better. > > > > It's possible to crosswalk across vocabularies, but the different > > vocabularies are optimized to support different needs so you'll want to > > pick one that's appropriate for your use. > > > > kyle > > > > On Mon, Mar 14, 2016 at 11:22 AM, Jacob Ratliff <[log in to unmask]> > > wrote: > > > > > Hi all, > > > > > > I currently work in an International public health non-profit, and we > are > > > setting up enterprise wide document management for dealing with > Knowledge > > > Management and Information Management issues. Lots of moving pieces, > but > > I > > > wanted to get some input on metadata specific to the medical/health > > world. > > > I am looking for some metadata guidance specifically related to the > > > medical/health world. Is anyone using any standard controlled > > vocabularies? > > > Should I be looking into Linked Data? I'm starting off the research > phase > > > for all of the metadata, so links to resources and case studies is > > greatly > > > helpful! > > > > > > Bonus points to anything that is international in scope, as over 75% of > > > the employees at my company are non-US based (most of them in Africa). > > > > > > Thanks, > > > > > > Jacob Ratliff > > > Information Architect / UX Specialsit > > > Management Sciences for Health > > > [log in to unmask] > > > > > >