Decision support, audit, hinting engines and several other tasks are too large for any individual or a single project to do all of them on their own.
One approach to handling some of these is to provide a means by which additional programs running on the same or a separet networked computer as the primary clinical program can inspect the data as it is saved, and take action when their algorithms tell them to. This is in my view better than the primary programs calling out for all additional functions, although it does not rule that out, and the advantage is not present for all cases. There is scope for a daemon or set of daemons or specification for a general case daemon which observes the data being recorded, accepts a list of conditions (eg "addition of Read Code G20|diagnosis of HT to the narrative record" and shells a program from a corresponding list to perform certain actions. The reason for attacking the general problem in this way is that it allows many people to work on the range of problems, rather than requiring a large project to handle several or all of them. Bodies of knowledge and work that are available for application to this area include the Arden Syntax, and the idea of Medical Logic Modules, and assorted triage and expert system systems where the code can be disclosed or there is a documented command line interface or API that can be used. This is also a good point to build in knowledge services, but that depends upon symptoms or findings being recorded in real time, which is not the case for all doctors or for all record systems - some don't commit notes until the patient is out of the door- but clearly there are a number of ways around that. I suggest the underlying daemon is a suitable concept for an Open Source project to take on, and that the promotion of its use and securing further components for it to use is a reasonable task for OSHCA and other health OSS groups. I invite comments off list or if they are likely to be of general interest, on list. -- From one of the Linux desktops of Dr Adrian Midgley http://www.defoam.net/
