Re: Help finding natural keys

From: Bernard Peek <bap_at_shrdlu.com>
Date: Mon, 20 Jan 2003 11:27:59 +0000
Message-ID: <mptkuc1$09K+EwHi_at_diamond9.demon.co.uk>


In message <2MZV9.300$I55.18392_at_news2.east.cox.net>, Alan Gutierrez <ajglist_at_izzy.net> writes
>Ray Cassick (Work) wrote:
>
>>
>> "Alan Gutierrez" <ajglist_at_izzy.net> wrote
>>>
>>> The organizations are small, and even in our large organizations the
>>> charting is done by small facilities. This is charting, not billing.
>>> Patientnumbers are not in use here. A good thing. I'd say that a doctor
>>> that referred to me as Q123.456 has got to work on her bed-side mannor.
>>
>> From a patient perspective I can agree here, but from a developer
>> perspecitve This is very limiting.
>>
>> People have to uderstand that thier data is stored in computers, and to
>> computers, numbers are easy.
>>
>> Just because the patient has a number is no reason for a small company to
>> not interact with them via thier name anyway.
>
>Of course. I am making the point that there is no existing patient number
>for most of my customers since, so far, none of them have electronic
>charting. Therefore, there is no existing numeric key. So I made one up.

OK. That's perfectly reasonable. Just watch out for the problems that this introduces.

If a Mr. J. Smith is admitted to have his left leg amputated can you always distinguish him from Mr J. Smith who has been admitted to another ward to have his right leg amputated?

What happens if the hospital reorganises and swaps the two patients over so each of them occupies the bed that the other one was using?

What happens if a patient transfers between one hospital and another. If both of the above patients move to other hospitals can you guarantee that the right records will go with them? Is this your problem or the hospital's?

>
>I am asking the group if there is some natural key for patient that can be
>used as a primary key. One that I might have missed. People asked how
>patients are indentified currenlty, as a source of clues. I am noting that
>there is no existing numeric indentifier that is brandied about by the
>nurses as they care for their patients. (Er, maybe room numbers)

The important thing about a primary key is that it is reliable. There must be no chance that two patients will have the same number. How much do you trust the hospitals to always give you the right number?

I'll give you a clue. When I was in hospital they put a wrist-band on me which had my patient ID, my name and date of birth written on it. Each time I had any medication the nurse would check the wrist-band then ask me my name. They didn't even trust an ID number that was physically attached to the patient.

-- 
Bernard Peek
bap_at_shrdlu.com
www.diversebooks.com: SF & Computing book reviews and more.....

In search of cognoscenti
Received on Mon Jan 20 2003 - 12:27:59 CET

Original text of this message