search

openhealthcare / elcid

elCID deployment for UCLH
http://elcid.openhealthcare.org.uk
Other
7 stars 3 forks source link

Show me what antibiotics did we treat a patient with an infection caused by organism X, resistant to antibiotic Y #164

Closed davidmiller closed 10 years ago

davidmiller commented 10 years ago

@GabPoll

michaeledwardmarks commented 10 years ago

This could get complex. It would be something like:

Investigations: Test Type = Blood Culture (Or more broadly test type = a user defined Test for which we capture the Organism and Resistant Antibiotic subfields) Investigations: For the above test: Organism Type = X Investigations: For the above test: Resistant Antibiotics = Y

Outcome of interest is: All inputs of - Antimicrobial: Drug (with Date Started and Date Stopped, Route, Dose)

An example might be: This persons blood culture on the 3rd April had Staph Aureus resistant to Flucloxacillin in it. What antibiotics did the person receive before and after that culture result.

davidmiller commented 10 years ago

The ticket is intrinsically not about a single person. If you were interested in one patient you'd just look at their episode.

The question has to be:

What antibiotics are used to treat patients with Staph Aureus resustant to Flucloxacillin

davidmiller commented 10 years ago

In what sense is this not actionable?

michaeledwardmarks commented 10 years ago

Yes I agree it is about a whole population - sorry it was just a real-language description of the problem.

I think it it is non-actionable in the sense that we might want to address this question with regard to a very large range of diagnoses/organisms/drugs resistance patterns - and that as such a 'report' won't be the correct approach. It feels like the solution to this is almost certainly going to be a query and then user analysis.

michaeledwardmarks commented 10 years ago

To clarify I mean non-actionable as an automated report - not as something we want to be able to query the database about. I think this both because of the range of situations we might want to look at this kind of question in and then we might also want to be able to look at other data e.g outcomes, other investigations for those patients - hence why I think this is something that would be addressed by a full data extraction.

davidmiller commented 10 years ago

I think this is basically done already - you can query for resistant_antibiotics Flucloxacillin and test blood culture ....

Then do your analysis offline?

GabPoll commented 10 years ago

For the moment, fine to close. Will re-visit once have clearer questions in mind.