 |
 |
At 22:29 +0000 on 29-12-1996, Peter Johnson wrote: > At 00:03 29/12/96 +0100, Gerard wrote: > > > >Nothing is wrong with EDIFACT. It works, there are standards! > > > >And the general free text standard for EDIFACT is a perfect wrapper for > >SGML enhanced messages. No ugly duck thing at all! > > > >But.. > >Did you ever thought how difficult it was to get at an EDIFACT standard? > >How difficult it is to change it? > >How many dialects there will be? (In Holland there are to many. One area > >cannot exchange information with the next one, because they use a slightly > >different one. In general we are very disapointed about the general > >usefullness of EDIFACT in medicine. It will work perfectly in small very > >restricted communities for very restricted purposes, but not generaly) > >How unforgiving EDIFACT is? > >How rigid EDIFACT is? > > Have you considered that if the two messages differ, it may be because there > is a real difference between the semantics of the two messages? If the two > communities haven't got together and thrashed out a common standard, then > there may be assumptions in one of the messages which would make it useless > in the other community? > > Only when the two have got together and agreed a common message that is > valid in both communities may you be sure that such possibilities have been > considered. Standards have never been easy! > > So I would argue this aspect of EDIFACT use as a positive benefit! > > There are no short cuts to defining standards - however you implement them. > The fact that nothing stops one system interpreting the Quantity tag in your > SGML prescription as the number of items to dispense, and another interprets > it as the quantity of the drug in each dose isn't important of course, > because at last we can send those on line prescriptions. Until someone dies > taking 50mg of warfarin. > Nice try. The problem is that the standard of EDIFACT has to be hardcoded into the software on both ends of the line. The nice thing of SGML is that it is not hardcoded. The standard can be derived by indirection so all can use the same version. So standards are more easily addapted if two parties decide to change things a little. And if they report it, make available to others, things might change faster. Look at the developments/evolution of HTML in the WEB community and you see how things might work out. And then the remark of the Tag for Quantity which could be interpreted as an number of tablets or dosage per tablet. This example is on the level of talking about horses and the other thinking its about flowers. If we agree that we need a general medical language (Med Speak) we all must learn it. This restricted language with a general set of concepts must be developed. The details, the refinements must be using specialised coding systems (READ, ICD, SNOMED) If we set out for the result that we must be able to use the Patient File for all purposes, the course is bound to fail. Each document can serve a few functions only. This true for paper systems. But using computers a little bit more can be achieved. Even with highly structured systems and databases it was impossible to answer all questions using the data recorded for the purpose of treating patients. Coding is a very subjective process. I happen to think that systems for treatment are for treament only! And perhaps the legal aspects. For research we will need separate systems, specialised coding, extensive trainng of the coders to code more or less uniformly. The separate systems will interact with the treatment systems offcourse. Gerard Freriks,huisarts, MD C. Sterrenburgstr 54 3151JG Hoek van Holland the Netherlands Telephone: (+31) (0)174-384296/ Fax: -386249 Mobile : (+31) (0)6-54792800 ARS LONGA, VITA BREVIS %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%