Grahame Grieve wrote: > hi David > >> If it's webservices it implies an IT infrastructure beyond the >> capability of small to medium sized general and specialist practice. > > > I think this is unfair, at this point. NEHTA have focused on web > services for precisely the opposite reason, at least in theory, which > is that anybody can get into the game, you don't need any non-free > infrastructure, and - at least in theory - the skills are widely > available. > The devil is in the fine print - webservices is all well and good but > by the time you pile all these extra things on top of webservices, > such as reliable messaging and the encryption stuff, then it's not so > easy anymore. NEHTA haven't indicated how far along that route > they will mandate, and that's what we have to pay attention to. > > Whether or not you need a "server", which *is* beyond a small GP etc, > is out of scope of the current round of documents, and also something > we need to watch. You can use webservices to > deliver content without requiring a server at the end point - the > IHE XDS shows a good way to use webservices like this. > However NEHTA do have this in the web services profile (hidden in > the techie document "Web Services Standards Profile", where it might > be argued to be easier to miss): > >> NEHTA recommends the use of a Service Oriented Architecture >> (SOA) approach for the design of health applications. >> >> Health applications should be designed using a service oriented >> architecture, where appropriate. The SOA approach is an architectural >> style where services are designed to perform self-contained units of >> work. In this context, the term “service” refers to the functionality >> provided by a piece of software. In SOA, services have well defined >> and documented interfaces. The services are invoked by sending >> messages to the service interfaces, and applications are created from >> a set of interacting services. > > > What this means all depends on what "where appropriate" means, and > who defines what appropriate is. SOA is certainly something that > requires an infrastructure, and probably a rewrite for most > applications - it's a fairly new idea in terms of application > development cycles. > > I think that NEHTA are trying to say that NEHTA thinks that NEHTA > specifications should be written to describe an SOA world - that would > be fine. If that's not what it means, then we should ask NEHTA > to rewrite it or remove it. We've had public comment on the Shared > EHR stuff, I hope that we will have public comment on this stuff too, > but I haven't received any notification of that. > > Grahame > CTO Jiva Medical / Kestral Computing. > co-chair HL7 Infrastructure & Messaging. > > [David, can you please forward this - or at least your response > if you do - to gpcg, I'm not on that]
Thanks Grahame, shall do. (Sorry for the delay. My day job got in the road.) Web services are useful if you have something to share. As the aggregators, filterers and interpreters of patient medical information, GPs are the logical choice to provide these services. Hospitals could also do this since their data will have some relevance for a window period after discharge, but hospital clinical systems are so pathetic this concept is risible. If we agree primary care is the right entity to manage this data, we have to work out how to do it. IHE XDS is not really a solution in my opinion. I know David Rowed says the Americans say everything can be reduced to a document but I am not convinced. IHE reflects work flows and not patient status. Modern Australian EHRs are quite reasonable at the latter and this is where the action is as far as clinicians are concerned. However, I agree that until such time as someone develops a SOA EHR, this is all just academic. It's all new, as you point out, but proprietary medical software development is glacially slow. Regards. David -- SIP [EMAIL PROTECTED] NodePhone +61 7 31290168 Jabber [EMAIL PROTECTED] _______________________________________________ Gpcg_talk mailing list [email protected] http://ozdocit.org/cgi-bin/mailman/listinfo/gpcg_talk
