29 March 2007
Darius-Do we want a mechanism to have an encounter be marked as fulfilling an order and if we have a bunch of obs come back at different times but are really are a part of the same thing? In this case a drug sensitivity test. How can I group those? One thought would be to have an encounter that get edited after the fact by having more obs added to it. The other alternative would be to have an obs group that stands observations in different encounters. Any idea on how that ties into HL7?
Burke stated that the most logical would be to connect the observations to an order. HL7 would let you say this observation is related to a specifc order and then that would allow you to infer the relationship. Burke thinks he would need to look at what he has available in observation through HL7 to link to something outside of the same ORU. The way this will be handled in HL7 is that there is a placer order number and a filler order number and those track who ordered the item and who fulfilled that order. The idea would be if someone ordered a culture and that order ID, or some other tag associated with that order, would be the placer order number. Ideally, that would go through whatever process in terms of generating the observations and when we got the information back from the system that had done the culture, that message would include the placer order number. What that would allow you to do is when an HL7 handler, when youÃƒÂ¢Ã¢â€šÂ¬Ã¢â€žÂ¢ve got in a message and if there was a placer order number in there, you could go look up that particular order or look up any results that also match that order and link them. The actual HL7 is what creates those relationships and in the OBR, in that group mechanism that we are already using in HL7, there is a spot to say here is the identifier for this particular request. Burke stated that his team hasnÃƒÂ¢Ã¢â€šÂ¬Ã¢â€žÂ¢t been using that. The piece they are missing in this model is this obs group table. ItÃƒÂ¢Ã¢â€šÂ¬Ã¢â€žÂ¢s on the list of things to do but just hasnÃƒÂ¢Ã¢â€šÂ¬Ã¢â€žÂ¢t been yet and thinks the answer to linking observations is across encounters. Not all observations that belong together are necassarily going to be collected in one encounter. Some of the things Burke and his team needs to work out is what is the order ID to point back to the order and needs to think about this a little because it could be part of an observation or it could be part of that obs group. Burke then asked, ÃƒÂ¢Ã¢â€šÂ¬Ã‚ÂIf you want to link an observation to an order do you have to make a group with one element?ÃƒÂ¢Ã¢â€šÂ¬Ã‚Â
Darius said that currently, observation has order ID in it. Burke stated that you could have two observations, resulting from two separate orders, on two different encounters, grouped under the same group. However, they donÃƒÂ¢Ã¢â€šÂ¬Ã¢â€žÂ¢t know how to display it but could be flexable.
Darius went on to say that one last thing to discuss is how to to do DSTs properly, they needed obs groups or else they can hack it to use a single encounter and to have it treat an encounter as one obs group. Burke replied with, ÃƒÂ¢Ã¢â€šÂ¬Ã‚ÂCreating the obs group wonÃƒÂ¢Ã¢â€šÂ¬Ã¢â€žÂ¢t be a giant task.ÃƒÂ¢Ã¢â€šÂ¬Ã‚Â
Discussion about planning GSoC project mentor assignments. Conference call, to wrap up this discussion, to take place tomorrow at 8:00 AM.
All signatures have been obtained for the OpenMRS LLC license agreement.
Approx. 130 people or more signed up for the OpenMRS Implementers Group Meeting in Cape Town from April 22-26.