In the history of medical information technology, mainframes and database management systems (DBMS) running on COBOL and mainly used for laboratory applications, were the top of the bill in the eighties. The focus was merely put on data. The nineties witnessed a shift towards client/server systems with object oriented programming languages and relational C++. In this context, relevant information instead of unstructured data became more important. Today, Dr. Van de Velde observes a growing trend to apply three tier systems implemented with Corba and Java, resulting in pragmatic, open systems for multi-functional use. There is a big discrepancy though between the rapidly evolving technologies and the applications that only change every 10 years.
To overcome this obstacle, we need a new type of HIS architecture which is both user and component driven, suitable for enterprise-wide applications, with a great mobility and intelligent features, according to Dr. Van de Velde. The client/server systems, which serve as repositories of knowledge, enable clinical users to communicate with remote systems. However, only 12% of medical knowledge is electronic-based. In addition, the number of medical expressions has accumulated to an exponential growth of 10 factor 7 during recent years. Because of this, we have to cope with very chaotic and complex database structures.
Quoting the famous Steve Jobs, Dr. Van de Velde made a plea for turning to knowledge and components. The challenges which lie ahead for better object use are productivity, reuse, scalability, and plug & play possibilities. In the HIS anatomy, the electronic patient record is the basic element. The care process has a central position, which in turn generates administration tasks and general services, including the lab and medical examination procedures. All these activities involve various work flow components, such as authorised agents, resources, outcome, and knowledge. With an ideally integrated HIS, it is possible to computise all areas within the hospital by following the logic of health care.
Polymorphism characterises this enterprise-wide model in which each single act always results in management. Electronic medical record (EMR) models not only are static but also dynamic, thus generating four possible acts that are fact, command, possibility, and wish. In addition, the HIS equally has to integrate within the work flow automation the two possibilities of temporary interruption and exceptional termination of the care process. In this way, the EMR models are turned into active objects, providing time management for patient and intervention referrals, messaging, task management per service in the hospital and patient lists, allowing physicians and nurses to view the patient situation on their individual workstation.
Dr. Van de Velde stressed the fact that a functional integration also implies the relationship with other components and applications, as well as the use of a common security model. The trinity of security measures is defined by the usability, safety, and cost. Modern HIS implementation is comprised of intelligent layers for medical decision making both in diagnosis and actions for treatment, where in fact there is much more support electronically for treatment than for diagnosis. Dr. Van de Velde described the four essential tactics to avoid errors and adverse events as a reduction of the complexity, the optimisation of the information processing, the use of constraints, and the competence to wisely automate.
Built-in defences, like alerts and reminders, serve as barriers to safeguard a patient from wrong treatment. Further more, physicians and nursing staff are guided by problem solving facilities and clinical pathways, describing the sequence of tasks. In turn, we see a growing importance of evidence-based medicine due to socio-economical reasons. The increasing role of information technology allows doctors to see more patients in less time, whereas the Web has introduced the means to consult the growing literature on best practice and to empower the patient.
The innovative HIS architecture also has to be user driven to transform the user needs into formal models. Dr. Van de Velde explained how this results in menu-based clinical notes where unstructured data is being classified into relevant information. The impact on the usability is enormous since it allows clinical staff to deal with knowledge instead of loose data. In contrast with the reusability in purpose oriented systems, the consequences of the current data entry methods consist in a severe loss of information, both in quantity and quality. Dr. Van de Velde indicated natural language processing as the solution to this problem. The patient-physician relationship is also expected to be positively influenced if the HIS can produce a temporal representation of the patient history, enabling the doctor to view the patient at a glance.
The mobility of the system has to be guaranteed via multiple access points in and outside the hospital. Remote access features will open the gateway to a telemedical approach including applications in tele-consultation as well as in tele-education, with a twilight zone in between these two areas. Whereas in 1993, we were still faced with proprietary point-to-point systems, at present plug & play systems are rapidly conquering the field. Nonetheless, some difficult to take hurdles still need to be overcome. In this regard, Dr. Van de Velde was forced to mention important factors such as the lack of vision and strategy, the lack of resources, uncertainty and a typical fear of change, the mismatch between applications and needs, the fragmented medical market, and the lack of standardisation.