Bl@@dy NHS IT... data matching help needed

[David Martin]

Point of order!

Access is NOT a RDBMS, it's an unholy mess.

Access is a GUI to an RDBMS  - it has an in-built database engine (JET) but is fine when pointed at more robust data stores. I've used it as a reporting front end and it does just fine.

Any dat set is as secure as the access given to it. An Excel spreadsheet can be secure - it just has to be on a machine that has limited access for users. Preferably not networked, secure login with physical verification and so on.  Describing Excel as a database is a bit of a joke though.

[tiermat]

Exactly my point, David, Access is not a RDBMS, no where near, you can use it, as you say, to be the front end to one, but a default setup of Access is a flat file DB, so not really any different to using Excel.

[Phil W]

Exactly my point, David, Access is not a RDBMS, no where near, you can use it, as you say, to be the front end to one, but a default setup of Access is a flat file DB, so not really any different to using Excel.

All rdbms have flat files behind the scenes. It just that Access has a SINGLE flat file. Well it can be two flat files if you seperate the front end elements from the data.  Access would be fine for what is wanted though. 3000 records is a tiny tiny amount.

Anyway, back on topic, does our OP have an answer now?

[David Martin]

Exactly my point, David, Access is not a RDBMS, no where near, you can use it, as you say, to be the front end to one, but a default setup of Access is a flat file DB, so not really any different to using Excel.

Access is an RDBMS by concept. Implementation is a minor detail. Excel is not relational. SQLite is an RDBMS that runs on a single file. 

As for the original query, as has been said, posting the data structure and the questions to ask will elicit appropriate queries.

[Jaded]

Access is a front end. That's where its name came from. It may have developed a bit since then.

[Chris S]

YACF - Answers to your questions, and SO much more!

Whether you wanted it or not.

[closetleftie]

Exactly my point, David, Access is not a RDBMS, no where near, you can use it, as you say, to be the front end to one, but a default setup of Access is a flat file DB, so not really any different to using Excel.

All rdbms have flat files behind the scenes. It just that Access has a SINGLE flat file. Well it can be two flat files if you seperate the front end elements from the data.  Access would be fine for what is wanted though. 3000 records is a tiny tiny amount.

Anyway, back on topic, does our OP have an answer now?

I have oh so many answers now.  Just none I can understand....

This may actually be the kick up the behind that I need. I may just need to get familiar enough with data science to be able to do simple stuff myself. It seems that, much like the ability to write or ride a bike, it's an underlying skill necessary for modern life.

SO: Where do i start? Assume reasonable intelligence but not a mind that works like that of an IT geek. Point me to a textbook that starts with "First, find the ON/OFF switch to that computer thingy".

[vorsprung]

I have oh so many answers now.  Just none I can understand....

Background:
Patient data in my NHS Trust sits on different systems. There's:

    The Patient Administration System (PAS) whci records stuff like clinic appointments, admission date, discharge date etc.
    The results system, containing e.g. blood test results
    Various theatre databases, with structured data on what happens when we cut people


Incredible as it may seem to those of you in jobs with proper IT, in my Trust (as in many others) these are in no way linked. I have a cohort of patients for whom i need data from each of these systems, pertaining to a defined time period for each patient (90 days either side of an operation which is recorded in the theatre database).

My Trust IT people metaphorically sucked their teeth and said something to the effect of "maybe the year after next, if ever" when I asked them for help.

Is there commercially available software that can link data like this? I don't need this to be automated - it can just be done once for the cohort.   Getting the source data copied from the databases into excel sheets or similar format won't be an issue. Though the copies will contain much that's redundant, of course.

Any advice appreciated.

1) they are linked by the patient's NHS number
2) any database system will tie this data up

If you want the yacf IT massive to actually suggest a "recipe" for doing the tie up and printing out some kind of report you need to

a) make a set of fake data which is like the real data but with all personal, sensitive information replaced with placeholders.  So "James Andrews" becomes "John Abcdeg".  The data should contain at least 3 relevant items per set and 3 irrelevant items.  So perhaps 3 fake patients who are like the ones you are interested in and 3 who you are not interested in
b) make a mock up of the report or report(s) that you need out of the system as output
c) present this data to the yacf tech brains, leave, stir and hope for the best

[Morat]

Just use salesforce like everyone else


(Only kidding)