There are many different types of application systems in a health care environment, such as the hospital information system, serving the clinic as a whole, telemedicine tools, and lower level systems, like Picture Archival Communication Systems (PACS) and radiology systems. They all are typically driven by data. The alternative way for driving these tools is indicated by the concept of work flow which has a strong relationship with standards such as HL7 and CORBA. Health care can be modelled as a business through a work flow system, consisting of processes and resources which may integrate staff, organisation units, roles and levels. In turn, all these resources are included into activities amounting to special services, as Dr. Lemke explained.
For the modelling of those services, a series of standards has been provided by the Work flow Management Coalition, which defines the process as being broken up into a number of activities that are linked accordingly and which draw on a variety of resources, including application programmes like HIS, RIS, PACS, or future independent functions that probably will work without PACS. In different clinical departments, like radiology or surgery, the various processes are modelled according to this work flow architecture. The old way of looking at processes is sequential, taking many steps to be implemented. Work flow modelling in turn enables the hospital management to parallelise the process and support it with parallel computing, according to Dr. Lemke.
Work flow is the automation of the business process, broken down into a set of activities. In radiology, these different types of activities may consist of the pre-loading of images or menu-loading, seen as one logical step in a process, as a basic building block for many other services. The run time system can be consulted to get instances of the models in order to operate or control the environment. The run time work flow server has access to the applications, conventions, and related functions. Medical staff has the possibility to look at the various activities in which they want to engage via client workstations and the run time system will provide them with the right frameworks.
Dr. Lemke briefly introduced the standards' issue with HL7 and CORBAmed to show how they integrate into CORBAmed activities. The HL7 consortium has issued version 3 which provides an information content based on XML, and some sort of patient record architecture. All this can be easily interfaced with the different hospital departments. The Electronic Health care Record can be transformed into certain types of coded entries or templates, in order to be communicated to lab systems. HL7 equally has a link to CORBAmed. Both consortia have come together to discuss the electronic patient record as well as the middleware.
The aim is to develop a commercial bible for the health care industry and to make the systems interoperable. To this purpose, the working groups have established contacts with specialised companies to bring standard proposals to the market. Although CORBA forms the largest software standardisation group, no European companies have been selected to participate in it, so Dr. Lemke fears that whatever comes out of CORBA will be very much targeted to the American standard proposal market. As such, the European Union or CEN will have little choice but to adopt the results. However, there might be still a chance to participate, if we return to the work flow issue, as Dr. Lemke points out.
Within the CORBA set-up, the middleware consists of three levels. The upper or business level where the services are being identified can be modelled with the work flow modelling concepts. CORBA is not so strong in this field, while the Work flow Management Coalition is leading the dance at this particular level where the interactions, objects, and interfaces have to be defined, using appropriate standards. If Europe is fast enough, it can play a key role in this specific computational area. A few European institutions already have gained some experience in work flow modelling. CORBA is principally concentrating on the so-called bottom level of the middleware, which is the engineering and technological support.
Dr. Lemke expressed a personal proposal to the MediMedia audience to look into these middleware upper level aspects in the future. None of the existing standards are satisfactory for the middleware levels that are covered only in a very limited and narrow way. There still remains a lot of space to fill in with suitable standards. In the application context, functions are becoming more and more important from the computational modelling point of view. The 3 different activities of work flow, HL7, and CORBAmed provide an interesting background for future ongoing work supported by the European Community. In conclusion, Dr. Lemke stated that the software architecture of the existing application systems will be substantially affected by this innovative type of middleware.