Bika OpenMRS interface
lemoene
I was contacted off-list by one of Carl's colleagues about to experiment
(correct me Carl) with interfacing LIMS to OpenMRS via XML and have
since forwarded Darryl's (www.bluebird.co.za) spec used by their well
established system
From a LIMS analyst point of view i want to make sure that all possible
data elements be included and i attach a completely over the top list
which could definitely be reduced. Please comment
Bests
l
Bill Lober
I'll cc a couple of people because I'm not sure they're on this
list. I'll also forward your email to them so they get the attachments.
We've been talking about an OpenELIS - OpenMRS interface w/ Darius
since the AMIA meeting in the fall, but have not written any code
yet. We have finally (last week!) gathered what we think are the
right standards documents, though.
It sure makes sense to take the same approach w/ Bika, and with our
Haiti EMR - iSante. It would be nice to have a two-EMR, two-LIS
world, all open source, and all using the same interface specs.
When we talked w/ Paul, Burke, and Darius, my understanding was that
they wanted to stick w/ HL7 2.x messaging, according to the following
general specs.
In brief:
Information Interaction
================
We wanted to support both a standalone Lab and connected Lab-EMR
installation.
- LIS would be able to query for a patient, and use those
demographics if a match was found, or choose demographics to refine
query if multiple matches were found
- EMR would be able to transmit orders to LIS, regardless of whether
a patient was known to be in the LIS.
- LIS would transmit results to EMR for any patient the LIS knew to
be associated with the EMR.
Some points left unclear
- if a patient in the LIS were matched w/ an EMR patient after
results were entered in the LIS, would historical results also be
sent as a message.
- would historical orders in the EMR be sent under any conditions
- should intermittent connections (queuing) be handled in the
application, or at the interface layer
Message-based Standards
===================
Demographic query - use PIX-PDQ implementation guide from IHE to let
LIS look up demographics in a master patient index, or an EMR serving as an MPI
Orders - use HL7 2.6 orders (most likely)
Results - use ELINCS 3 (HL7 2.5.1) for results
Document-based (XML)
=================
I had favored a document based approach, using CDA documents,
constrained according to the IHE XDS-Lab spec, for lab results and
perhaps orders/status. I'd use a document based query approach,
based on the demographic part of XDS-Lab, if it were up to me.
But, I got talked out of it by what I understood to be the messaging
orientation of OpenMRS.
I'd be interested in comments from Darius, Paul, Burke, etc., either
on or off list.
Bill
lemoene
> I'll cc a couple of people because I'm not sure they're on this
> list. I'll also forward your email to them so they get the attachments.
>
I reply to all here and also cc Carl and Ryan who are investigating the
OpenMRS / LIMS interface locally
> We've been talking about an OpenELIS - OpenMRS interface w/ Darius
> since the AMIA meeting in the fall, but have not written any code
> yet. We have finally (last week!) gathered what we think are the
> right standards documents, though.
> It sure makes sense to take the same approach w/ Bika, and with our
> Haiti EMR - iSante. It would be nice to have a two-EMR, two-LIS
> world, all open source, and all using the same interface specs.
>
> When we talked w/ Paul, Burke, and Darius, my understanding was that
> they wanted to stick w/ HL7 2.x messaging, according to the following
> general specs.
>
I have contact with the Malaysian group (Lee Seldon, Alvin Marcelo,
Molly Cheah) including IndivoHealth, myCare2x, World Vista, PrimaCare,
OSCAR, OpenMRS, MoCa, CHITS, ffEHR, investigating not only HL7, but also
XML standards such as CCR, CDA and CCD. They seem to favour 'modern'
XML standards with the provision that these need to talk HL7 too and
mention the Australian Health Messaging Laboratory for testing HL7 messages.
I do not have capacity to research these at the moment - in fact, please
excuse my ignorance and enlighten us where i stray. As LIMS designer i
look from too high up, see my plain text list of data elements;-) and
did not realise that HL7 and XML are mutually exclusive
At this stage though it does not look as if the Malaysian group has a
LIMS/LIS in the mix yet - if OpenElis and Bika are using the same open
standard it would help consolidate matters
I'll be forwarding this mail to Lee, Marcello and Molly for comment. I
could not establish whether they have an open forum where their
discussions can be tracked. I think they are just starting out after
their interoperability sessions in Penang November 2008
I am not sure what Ryan and Carl are planning for their project either.
Carl/Ryan would you mind posting a summary here please?
At Bika all our development is carried out to (client) sponsor
requirements. At this stage our government grant is too limited to
include EMR interfaces in Bika Health 1 development but i'd love to
specify it in any case, using an open standard
In South Africa Bluebird has big penetration amongst labs, hospitals and
general practitioners, including the NHLS, and we cannot ignore that.
Darryl maybe you should also summarise the Bluebird service for us please?
> In brief:
>
> Information Interaction
> ================
> We wanted to support both a standalone Lab and connected Lab-EMR
> installation.
> - LIS would be able to query for a patient, and use those
> demographics if a match was found, or choose demographics to refine
> query if multiple matches were found
>
I am not sure I understand. Am i correct that in this scenario there'll
be Patient module in both LIMS and EMR and that analysis requests/orders
can be done from both the EMR and independently?
> - EMR would be able to transmit orders to LIS, regardless of whether
> a patient was known to be in the LIS.
>
Sample ID, Sample type and requested Analyses should be enough? If the
LIMS needs to report out of range results and does have a specifications
module, patient age, gender, weight, fasting Y/N, etc. too. My feeling
has always been that results interpretation should be done in the EMR
where more information and Patient history will be available
> - LIS would transmit results to EMR for any patient the LIS knew to
> be associated with the EMR.
>
Sample ID, Analyses and their results. When the analyses were
requested/ordered via the EMR. The patient can be referenced from the
Sample ID in the EMR. Some redundancy could be considered
> Some points left unclear
> - if a patient in the LIS were matched w/ an EMR patient after
> results were entered in the LIS, would historical results also be
> sent as a message.
>
I need to be convinced of Patient info in both LIMS and EMR. I am
overlooking something, what is the use case?
> - would historical orders in the EMR be sent under any conditions
>
New EMR installation with incomplete DB?
> - should intermittent connections (queuing) be handled in the
> application, or at the interface layer
>
Interface layer is my guess
> Message-based Standards
> ===================
> Demographic query - use PIX-PDQ implementation guide from IHE to let
> LIS look up demographics in a master patient index, or an EMR serving as an MPI
>
This assumes once more that the LIMS will need patient demographics. Is
their any need for this other than results interpretation? I think this
is the important bit missing in my understanding of Health LIMS/LIS
> Orders - use HL7 2.6 orders (most likely)
>
> Results - use ELINCS 3 (HL7 2.5.1) for results
>
Handy. Ryan please note!-)
> Document-based (XML)
> =================
> I had favored a document based approach, using CDA documents,
> constrained according to the IHE XDS-Lab spec, for lab results and
> perhaps orders/status. I'd use a document based query approach,
> based on the demographic part of XDS-Lab, if it were up to me.
>
I confess ignorance...
> But, I got talked out of it by what I understood to be the messaging
> orientation of OpenMRS.
>
> I'd be interested in comments from Darius, Paul, Burke, etc., either
> on or off list.
>
None so far, if so and also from others, i'll get permission to post it
here. I feel not to be technically clued enough to contribute much more
than the data requirement posted earlier. It could be meaningful to
establish an interoperability forum where yours, the Malaysian group and
Carl/Ryan can thrash the technicalities out. I'll find out whether Lee
has something going already
Bests
lemoene
www.bikalabs.org
lemoene
Carl/Ryan would you mind posting a summary (of your project objective) here please?
We are developing an OpenMRS module that allows basic communications with a labs system through a Mirth server. The module as it is planned at the moment will only be able to request lab results for patients as well as receive and parse labs results and create an encounter in OpenMRS for those results.
We have decided to use the Blue Bird XML specification for this communication as the implementation that requires this module is here in South Africa and apparently Blue Bird is use extensively here.
I am watching this with interest though to see peoples comments on how this integration should best be done. I am trying my best to code this module generically so that different messaging formats could be added (say HL7 for example).
Regards,
Ryan
Ryan
Sorry for waking the dead, but I've now coded an OpenMRS module that allows for some basic LIS connectivity. I have been using the bluebird XML spec for receiving results but I would now want to move onto HL7.
I've been looking at the HL7 2.6 spec and so far the messages I've picked out for communication between the system could be a OML^O21 message for the lab orders and one of the OUL's for receiving results.
I remember from this conversation the following was mentioned:
Orders - use HL7 2.6 orders (most likely)
Results - use ELINCS 3 (HL7 2.5.1) for results
Why for the results would a 2.5.1 message be used?
I want to support messages that would be most useful. What messages would be best to use?
Regards,
Ryan
I am not sure what Ryan and Carl are planning for their project either. Carl/Ryan would you mind posting a summary here please?
We are developing an OpenMRS module that allows basic communications with a labs system through a Mirth server. The module as it is planned at the moment will only be able to request lab results for patients as well as receive and parse labs results and create an encounter in OpenMRS for those results.
We have decided to use the Blue Bird XML specification for this communication as the implementation that requires this module is here in South Africa and apparently Blue Bird is use extensively here.
I am watching this with interest though to see peoples comments on how this integration should best be done. I am trying my best to code this module generically so that different messaging formats could be added (say HL7 for example).
Regards,
Ryan