Sunday, December 24, 2006

Ubuntu Medical: Centralised patient databases in the African context

Andy eventually got me to sit down and give some prelim views. The trickiest thing, if you REALLY want to make a difference in South Africa is planning. There are a couple of lovely third world things that get in the way:


*Politicans
*Egos of management
*Distribution
*Concentrating on PATIENT MANAGEMENT, not financial management.
*Training
*Training
*Training


For instance, if the open source community can help with the situation I'm about to describe, and just this, you have no idea how much of a difference it will make to primary healthcare in this country. Build the "system" so it can be grown, sure. But little steps is the only way you'll get anywhere in a third world country.

Here's the situation. Primary healthcare patients (GP-based needs, meant to go to clinics, both rural and metro BEFORE they get referred up to secondary and tertiary centres - where they would get seen by specialists and clog up already clogged hospitals) HAVE to carry around a green piece of cardboard - their clinic card.

This card links to a manually filed medical history inside the clinic archives. To get to a hospital, you're meant to get referred - thus creating the link. If you don't, hospital doctors are practicing blind, because there is absolutely no way to retrieve paper-based files from clinics around the country.

And most patients believe they can just pitch up at hospitals with primary care problems (translation: trivial from a hospital's point of view). Small wonder we're understaffed? Especially Baragwanath.

Hospital doctor's don't need every little bit of previous scrawl - but there are some critical pieces of information that should be kept centrally on a patient database - and easily accessed at critical times. Such as: Patient diagnoses (incl the controversial HIV + - status, in code), current medication / dosages, allergies and relevant major tests (eg. CT SCAN, ECG, major blood work).

Solve that problem and you've taken a huge leap forward.

7 comments:

Sebastian Hilbert said...

*Concentrating on PATIENT MANAGEMENT, not financial management.

--> GNUmed could help you with that

... - but there are some critical pieces of information that should be kept centrally on a patient database - and easily accessed at critical times. Such as: Patient diagnoses (incl the controversial HIV + - status, in code), current medication / dosages, allergies and relevant major tests (eg. CT SCAN, ECG, major blood work).

--> GNUmed is up to that. Take a look at the roadmap at http://wiki.gnumed.de

Sebastian Hilbert said...

Hello again from the GNUmed docs.

Thanks to the Debian guys there are Ubuntu packages to GNUmed.

http://packages.ubuntulinux.org/edgy/misc/gnumed-client

We just released version 0.2.3 which is not yet available for Ubuntu I guess.

Stay tuned and take a look at what GNUmed can do for you today.

http://wiki.gnumed.de/bin/view/Gnumed/WhatCanIActuallyDOWithGNUmedToday

Christopher Marshall said...

If this project were to materialize I would be very interested to hear what happens; it is an area of the open source world which I find surprisingly neglected.

A couple of points to make or questions to ask:

Sebastian: Is GNUmed actually used anywhere in a live environment? I couldn't find anything on the wiki to suggest that is is, but obviously that is not proof!

Does GNUmed manage messaging lab results? If so, does it use standard technologies (e.g. HL7)

Is it capable of clinical coding; by this, I mean the more clinically useful terminology like SNOMED CT rather than ICD10.

SA Doc: How computerized are hospitals in SA? I understand that at least some labs use commercial systems, but I have not heard of an administrative or clinical system being used. Secondly, in the community would it be safe for people to use such a tool, or would it expose them to too much risk from theft?

Beyond that, I've subscribed to your blog and will be watching with interest, good luck.

Andy Hadfield said...

Sebastian: Thanks for the comment. I'm continually astonished since this post went out at the amount of activity in the open source environment, medical field. Am very interested though in practical applications? Anywhere I can go to look at an actual documented installation?

It may be short-sighted - but I think the biggest barrier to open source would be the support. If it works, perfect, but if it breaks, you can't call in your 12 year old to reboot/fix the windows registry :) Linux as I've seen gets a bit more complicated. Stability will be key if an already existing package gets used... Thoughts?

Andy Hadfield said...

Christopher: Very interesting stuff and share the sentiment. Lab messaging, as we said in the post - is extremely important. It's that real world gap between the paper based filing systems of clinics, the paper based filing systems of hospitals, the e-filing used by labs and SOME way to get everyone to talk to everyone... Will go check out those links asap.

SA Doc said...

hey Christopher

There are computers in most of the hospitals and clinics in SA. The problem is that most of them are ancient and you're lucky to find a functioning one between 5 wards.

They are currently not used for patient databases but for admission stats ( they are supposed to communicate with the admin dept) and lab results.

The NHLS (labs) use a system called DISALAB to run all results on but they are only specific for the lab that the results are sent to (Bara patients at Bara Lab). Thus if you wanted results from a different lab (the patients previous admission at JHB Gen Hospital) you cannot access those.

The risk of theft is always going to be an issue. But in most clinics and hospitals, the computers are welded into place with huge steel cage structures, that unless the thief brings his angle-grinder, cannot be removed.

Hope that helps...

Christopher Marshall said...

This does ring very true, I worked on a clinical app a while ago which was little more than a result view; this on its own was got a very good response to the end users.

Regarding the age and usefulness of hardware I think you might be interested in the "Ndiyo" project; despite its name it isn't a South African project (sorry Andy!) but based in Cambridge, UK. However, it has been used in South Africa as a means of delivering computer power to end users very cheaply. I'm quite convinced that it would be capable of providing the sort of power required to show some lab results and basic patient information!

http://www.ndiyo.org/

- Christopher