On Sun, 8 Dec 2002, Thomas Beale wrote: ... > I should say about this that all it is is an attempt to see what would > be needed in the EHR to support workflow. We envisage a workflow engine > outside the EHR to perform the steps, and expect that this engine will > have a formal representation of the workflow. In our example, we have > shown it as a state machine, which I know is not generally strong > enough, but it may be strong enough just to represent what state the > workflow is in for the purposes of noting in the EHR (i,e, the "real" Wf > execution in the engine will probably have a representation in a more > sophisticated formalism).
Thomas, "Recording" just the current state of the workflow instance is not sufficient. The entire "history" of execution of all workflow instances constitute the patient's medical record. This is what I decided to implement for OIO-1.0.0. > One of the problems to consider is where a Wf engine will have a > trustworthy persistence mechanism for each patient - should it maintain > its own, should it use the EHR? Can you think of a reason why it would be useful to have a separate EHR for the workflows engine? If the workflow engine keeps its own records, it would, in essence, be maintaining its own medical record system. :-) I can't think of any reason why this is useful - so, we did not take this approach. > Do we want the Wf to be represented transparently in the EHR so > clinical users can tell what things are beiing done to/for the patient > by the workflow execution system? etc If you mean whether physicians should be able to review workflows and their enactment history and status, the answer is clearly yes. In fact, this was the first feature request after we introduced workflows to OIO users at UCLA. Best regards, Andrew --- Andrew P. Ho, M.D. OIO: Open Infrastructure for Outcomes www.TxOutcome.Org
