Article Preview
TopIntroduction
The integration of pervasive computing in healthcare has enabled patients to take profit from high quality health care services anytime and anywhere. For this reason, providing continuous remote monitoring especially for elderly patients into the comfort of their home will be certainly less expensive than the cost of in-clinic recovery and more efficient. Furthermore, the advancement in mobile devices and in context-aware technologies has contributed to enhance the quality of healthcare services and the users’ mobility from one place to another.
Pervasive healthcare has the potential to reduce medical costs and improve quality of services; however it faces many technical difficulties related to the integration of the different medical information systems interoperating under the same healthcare application and maintaining periodically changeable schemas. In front of this challenging problem, middleware is gaining interest by insulating different systems from proprietary interfaces and providing layers of abstraction giving hence the illusion of using a homogeneous system. In many cases, physicians face problematic situations threatening patients' lives and needing fast and firm decisions; nonetheless classical medical middleware systems used to force physicians to submit traditional queries in which they must inflexibly precise their crisp conditions. This inconvenience limits the potential of the healthcare application used by physicians having few times to save patients’ lives. In fact, they might have enough time to fix carefully their queries and can't wait any longer for needed answers. For this reason, we put forward a pervasive healthcare application based on an efficient virtual integration system in order to allow physicians to express freely their preferences using natural language and vague terms and consequently quickly to obtain their desirable answers anywhere and anytime while treating patients in the shortest time and therefore free medical resources for eventually other urgent requests.
In this paper, we propose the Top-N rewritings algorithm, a novel and efficient algorithm for generating ordered top-N query rewritings which are likely to return the top k-tuples fitting with user preferences. This algorithm is the first one to generate the N best rewritings without computing all possible ones while it discards useless rewritings from the query rewritings searching step and hence, generates only the best likely satisfactory ones ordered according to their satisfaction degrees of the user query. We discuss also two different solutions on how exploiting the output of the Top-N rewritings algorithm in order to recover the Top-K tuples considered as best answers. Finally, we envision an implicit query cache memory in the middleware to avoid useless query processing when it is already available in the cache.
In the next section, a brief summary about pervasive healthcare is presented. Section 3 shows the contribution of middleware to help integrating different medical information systems interoperating under the same healthcare application. In section 4, we define a fuzzy medical conjunctive query. Section 5 is devoted to describe the two proposed algorithms. This section describes also how to exploit the Top-N rewritings to determine the Top-K tuples. Section 6 explains when and how using the query memory cache. In section 7 we discuss the feasibility of the study andthe final section concludes the paper.