Re: [Gnumed-devel] demographics.sql

[Jim Busser]

On Mar 4, 2004, at 11:51 PM, Ian Haywood wrote:
> Richard Terry's contacts panel has a hierarchial structure, roughly
> Organisations (hospitals, path. companies, clinics)
>         -> Divisions (such as hospital departments)
>                 -> Branches (such as collecting centres for path. companies)
>                         -> Individual people.

So far so good - and I like the above

> Its complex, as individuals may be associated with several entries...
> My current thinking:
> 	- table org, with a reference back to itself to represent divisions. 
> Toplevel organisations have NULL here.
>         - tables lnk_org2comm_channel, lnk_org2ext_id, lnk_org2address
> 	- other tables regularised in naming: lnk_identity2comm_channel, 
> lnk_identity2ext_id,  and so on,  so we can use the
> same client business code for both. ("identity" is the central table 
> we use for individuals at present)
> 	- no specific table for branches. Instead, we conceptualise branches 
> as organisation X at address Y.
> Similarily, individuals who are also have address Y as one of their 
> addresses are "members" of that branch.

If Mary is one of the clinic's staff, and Mary has also gotten care 
within the clinic, would Mary as Individual X at her home address Y 
define her as a patient, whereas Mary as Individual X at clinic address 
Y would define her as staff? Would it be in the lnk_identity2address 
table that the individual's "context / role / position" is defined and 
stored (e.g. patient, staff, president). Further, in the case of 
specifying a doctor, would we here define the individual simply as 
"doctor" with the requirement to use a second field in the same 
lnk_identity2address table for the detail (e.g. psychiatrist, 
pediatrician)?