Have spent a bit of time looking at non NHS business process, especially open source business process engines (Activiti) and business rules engines (Drools).
In health (I believe) business rules equates to clinical decision support. It enables (non techie) rules to be created such as:
If child has attended ED 3 or more times in the last month create a task for safeguarding team
Well it’s not exactly like that but just wanted to show what I was talking about. Also I’m a techie not a clinician so I’m not creating the rules (the rules engine is meant to be simple and so non IT techie).
With more and more interfaces becoming available it’s getting easier to do this. For example the above rule by using Spine CPIS, can be changed to
If child has attended ED 3 or more times in the last month AND is not known to social services create a task for safeguarding team
Once GP interfacing becomes a possibility we can include rules against the primary care record. Its also likely more systems will expose FHIR API’s at a rapid rate.
Are we likely to go down this route? Example open source project (using drools and FHIR) http://www.opencds.org/
I’ve been thinking about this for sometime. There is certainly a place
for generic work flow tools, like those you mention, while more complex
decision support (which is not simply ruled based) requires something
different Proforma?
The challenge is that there are too many relevant standards and
supporting tools. I recently become involved in a new open source
venture to try and pull some of these together have a look at this http://synapta.org.uk/http://synapta.org.uk/We just setting up a
Synapta community with Code4Health and are keen to work with all
interested parties.
Synapta have just decided to become part of the openhealthhub family, rather than hosting their own forum. That way they benefit from joining in with the existing knowledge and community here on OHH.
I’m hoping they will be able to provide clarity and straightforward implementable standards for decision support, where none have really been able to do so thus far.
I think it’s important that they don’t get too bogged down in the technology of it all, and focus on providing clear, peer-reviewed, evidence-based decision support rules in plain English. Which can then be implemented in any technology. (Later, we may see some kind of DSL for these rules, enabling programmatic implementation, but I’d be wary of trying to get the DSL in place first, there are other technologies that have tried to do this previously and failed - eg openclinical)
Yes, an excellent move for Synapta to get involved here on OHH.
The recognition of the large and swampy territory that is CDS and pathway / workflow implementation was a big part of the decision to go open and to start gathering information and act as a focus for expertise.
I’d like to think a DSL will emerge and evolve organically from what Synapta does.
Agreed. I’m getting very vary of leading with tech (I’ve managed to cut down my use of FHIR drastically ) and computer scientists + others need to be wary of making things too complex (HL7v3 and CDA)
Look forward to learning more about Synapta.
@Ewan not looked at Proforma but thanks for the links. I’ve come at this topic from Java point of view and was aware I’m coming up with Java answers.
I guarantee that no decent computer scientist or software engineer would have come up with HL7v3 or CDA. That was/is one of the major problems with both. Even if they included some useful health informatics thinking, it was precisely the lack of IT/engineering professionals in the original developments that led to their very major difficulties.