Furnari, Chris wrote:

> As a programmer I see a need to know
>
> 1)  How should the data be stored/accessed
> 2)  How should the data be manipulated/updated

well..... actually, if you have a component available, then all you need
is the
API for the component. Unless you are building the component itself. So,
it
depends on what you are trying to do.

Most people agree that it is a good idea to build one or a small number
of
high-quality kernel components and reuse these ubiquitously. Then
application
programmers can get on with their applications, without reinventing GEHR

(CEN/CorbaMed, etc) implementations all over the place.

> If the data happens to be in SQL data table format (or similar) I as a

> programmer would need to make a 'user interface'  that would allow
access to

even better, you don't want to know anything about how it is stored,
just how to
access it (i.e. via an API)

> the data and display it in a easy to read manner (item number 1).  I
would

this is where the API comes in

> also see a need to update demographic and insurance information as
well as

Both of these are likely to reside on their own servers. GEHR does not
atttempt
to define a full demographic model by any means - it just uses the
minimum info
needed to make sense of transactions, and leaves complex demographic
models to
local demographic databases (which tend to be all different). It records
the
local unique id of demographic entities in the record if needed.

> add on records to the patient history (item number 2).   We, as a
group,
> need to address what standards, system designs, methods of storing and

> passing data are needed to accomplish all of this.  For two examples,
SQL
> tables may be the way to store the information, while HL7 may be the
way to

you probably mean relational tables here. SQL is a way of querying, and
is
something you are quite likely to want to do. But that does not mean
that you
want to store things in relational table format.

> pass data between cooperating systems.  Maybe we should be looking at
what
> standards will implement a portion of the picture best.  When we hav

HL7 will certainly move data around, but HL7 2.x is problematic. Have a
look at
the HL7 v3 work to see what they are doing now  - it is a model-based
architecture, same as GEHR. Actually, there is a wealth of experience we
can
learn from in HL7, and we are certainly studying their work in order to
improve
our own. We intend to be HL7 interoperable in the medium term.

Hope this helps.

regards,

- thomas beale






Reply via email to