Acquisition to admin interface
This forum is now open for business! Thanks to the members from industry for joining up. I am sure more clinicians will join up in due course.
I should like to get the ball rolling by asking how you see the links between recording equipment and hospital admin systems developing.
- Will HL7 be the final choice or an XML variant?
- Will machines interface directly to hospital and NHS systems or through an intermediate system?
- What sorts of function will be available?
- Is there scope for industry co-operation and standards or have we clinicians got to negotiate this interface separately with each company?
Groups:
© 2000 - 2010 Dr Adrian J Fowle
Comments
Connectivity of patient data/record
The real problem is not the flow of information from our machines/service systems but, from the Trust's perpective, that we have patients that don't feature on PAS or who have multiple entries. They are desperate to reduce the duplication of records, one of the main factors being the variation in minimum data set offered up by the referring agent. This will be an even greater challenge when we get referrals from 'any willing provider'. We already have an ISTC that holds records and images that cannot be accessed from the Trust so duplication results, even with the obvious IRMER implications. Bring back Stalin!
Connectivity of patient data/record
True but if every HIS/PAS/EMR had fully implemented and used the unique NHS Number then duplicates would be severely reduced as has happened in Trusts where they have done this. Indeed the business of hospitals having their own hospital number seems to me to fly in the face of the logical value of using the NHS Number (### ### ####).
I agree - bring back Stalin or at least bring in someone with the clout and foresight to mandate some of these ideas.!!
Multiple numbers work
In relational database theory, you should not use an externally generated identifier as the key for important tables. I won't bore you with the complicated reasons - but this might dictate the use of a hospital generated number internally. How different systems communicate with each other is another matter - a shared, agreed identifier would be a great help.
I am not sure that all patients have NHS numbers. The Arab business man who flies into London just to have his EMG done by Sir So and So at a private facility might not. Of course he may require other attendances elsewhere at some point and might then need two systems to communicate without the NHS no.
I have found the argument that these are rare occurences little help when one turns up! It is better to plan for this, even if some manual intervention is required.
Dr Adrian J Fowle FRCP BSc Consultant Clinical Neurophysiologist. Editor, www.clinicalneurophysiology.org.uk
Data safety/exchangeability
My view is that in the absence of a formal mandate from literally anyone, that HL7 be adopted by the profession. XML is fine but appears to be at the behest of our masters, Microsoft whereas HL7 is a very transparent and controllable language. This is what other groups have done and not to their detriment. Providing that true HL7 results files can be produced, there is no reason why our clinical gear can't interface directly with NHS systems although there may be other good reasons for doing it via an internediate system (eg local data ownership issues). For clinicians to negotiate with each vendor every time any hardware is being replaced sounds arduous and if the professional group defined an HL7 standard then the vendors would have to meet it. I accept that some vendors would say that the UK market is too small to do a one-off development but as software is now available to convert Word .doc into HL7 that's an ether argument. By data ownership above, I mean that data accuracy is vastly higher where the departments 'own' their own database of referrals and results (as opposed to simply shunting the results thru the network to the NHS system). Furthermore, maintaining local departmental databases provides a degree of stability esp given the fact that the average life of a hospital system is seven years. I am aware of departmental systems that HL7 to clinical gear and talk to hospital systems that have been in place for 24yrs+. But also a word of caution - there is HL7 and there is HL7. I've seen outputs from supposed HL7 interfaces that don't even start to meet even v2 of the official spec (see www.hl7.org). The phrase 'hl7' seems to have become an advertising lever with little control over it's true spec. Hope this rambling isn't too long..... regards Owen
XML
I yield to no one in my distrust of Microsoft - but XML is supposed to be an open standard defined by W3C (See /www.w3.org/TR/REC-xml/). I am aware of proposals to bundle HL7 into XML. I found HL7 hard to get to grips with when I tried it several years ago. XML on the other hand seems a simple concept, even if it is a little tricky to parse sometimes.
I entirely agree with you Owen about the importance of local control and ownership.
I think we need to start with really simple things - can we agree as a group of clinicians and industry on
Regards, Adrian
Dr Adrian J Fowle FRCP BSc Consultant Clinical Neurophysiologist
West Surrey Clinical Neurophysiology, St. Peter's Hospital, Chertsey, Surrey. KT16 0PZ
Still learning
Dr Adrian J Fowle FRCP BSc Consultant Clinical Neurophysiologist. Editor, www.clinicalneurophysiology.org.uk
HL7 and all that
It's still a live subject Adrian - I've been busy over the last year or so writing HL7 outputers for MINDEX to the likes of Accenture, ICE and a few others. I've also developed a 'hacker' that pulls the text-only out of .doc or .rtf files and creates the necessary HL7 whilst retaining the colum alignment - now that was a task and a half given Billy's lack of formal documentation about table structures! There are also a few interesting tools around to check the HL7 struture and the various HISS/PAS providers now seem to be working from the same hym sheet in that they all provide the same document for their side of the interfaces. Interesting times.....
HL7 and all that
I also meant to ask - dose anyone know if the NHS Number live allocation service ever happened? I understand that it was intended to cater for those Cat-II's and/or non-resident patients who had no number. There are certainly plenty of numbers available in the range!
HL7 and all that
To keep this discussion alive (where is all of the vocal expertise out there or does apathy rule?) - and this will probably interest you Adrian, there is a very nice HL7 testing package called 'Interace Explorer' that really does test .hl7 message/result files. It's only a few quid but quite frankly puts most providers to shame when run! You may already have heard of it.