[FOSS_health] Re: interoperability

[David Forslund]

Tim C wrote:
> On 04/06/07, *David Forslund* <forslund at mail.com 
> <mailto:forslund at mail.com>> wrote:
>
>     Precisely.   I have a simple analogy to explain this.  Just because a
>     group of people all might use MS Access does not mean that they
>     have any
>     real ability to exchange data between these systems.  They are all
>     running the same software, but have probably created their own data
>     dictionaries (tables, etc). 
>
>
> That's not an ideal example, because MS-Access is an application 
> development environment, not a specific application per se. generally, 
> all instances and implementations of a specific applications share the 
> same core tables and schema, although many applications allow a lot of 
> site-specific customisation of the schema, data capture screens, 
> reports etc around the edges. That is a problem, as is the semantic 
> encoding: the use of different code sets at each site, even if each 
> site uses the same software. If there is some overarching or shared 
> governance between sites, then a shared or common "data dictionary" 
> (that is, meta-metadata, codesets etc) can be formulated, but often 
> the only thing in common between different sites is the fact that they 
> are using software from the same vendor (or open-source provider) - 
> and that, as we have been discussing, does *not* mean that 
> interoperability will ensue.
I understand.  I was only trying to give an analogy and I've seen this 
problem first hand..  I know that it easy to change the dictionaries in 
VistA, which is equivalent to changing the schema in a typical 
database.  Your example is, obviously, much better.   And since people 
don't try to keep the semantics well documented or well organized, the 
issues you raise are always present.  As I said, the DoD, didn't have 
enough "governance" to regulate this process even between sites.  The 
issue for the DoD was, in my opinion, mismanagement by the company that 
was doing this (it made them more money).

Dave
>
> Of course, there are only very weak "business drivers" for a hospital 
> in Tegucigalpa and a hospital in Sydney to be able to share or 
> interoperate with respect to clinical information (research 
> information, maybe), but there would seem to be strong business 
> drivers for a hospital in San Salvador or Managua to be able to 
> interoperate at the semantic and syntactic levels with a hospital in 
> Tegucigalpa?
>
> And business drivers for interoperability between US military health 
> information systems? Um, what is that old saying about military 
> intelligence?
>
> Tim Churches
>
>       The DoD have all used the same software,
>     but weren't required to use a common dictionary or to maintain a
>     common
>     dictionary.  This may have changed more recently, but it certainly was
>     the case around the year 2000.   I think the VA has done a better
>     job in
>     this regard.
>
>     Dave
>     Joseph Dal Molin wrote:
>     > In both open and proprietary systems users of the same application
>     > across multiple organizations usually have the power to add new
>     > terminology etc. without some centralized "terminology"
>     > standardization process.
>     >
>     > Joseph
>     >
>     > Nandalal Gunaratne wrote:
>     >> David,
>     >>
>     >> "Just running the same system at multiple locations is
>     >> no guarantee of semantic interoperability."
>     >>
>     >> Please explain why this is so, in brief.
>     >>
>     >> Thanks
>     >>
>     >> Nanda
>     >>
>     >
>