به سوی بهره برداری هوشمند از منابع ناهمگن و توزیعی در محیط های مشترک سلامت الکترونیک

The objective of this research work is to empower healthcare information systems to deliver high quality information anytime and anywhere and to present distributed and heterogeneous resources access solutions which perfectly meet user requirements in different contexts. To reach this objective, we propose in this paper a new framework, called ONtology Oriented Framework for Pervasive Applications and Services (ONOF-PAS), and mainly based on: 1) interrelated ontological models representing the main entities in pervasive computing, such as Organization, Actor, Task, Service, Process, Object, Resources, etc.; 2) rule-based reasoning to infer the available and capable resources required by each task. The main functionality of the ONOF-PAS framework is to capture and to process the acquired knowledge about the user's context, the tasks they perform, the availability and capabilities of the required resources, and the organizations that own or manage these resources. Our framework allows healthcare information systems to infer the needed resources by linking each user-task with the required, available and capable resources taking into account the clinical conditions of the patients. The ONOF-PAS framework has been successfully applied in the telemedicine domain in order to provide practical and efficient resources access solutions that better meet the contexts and the expectations of the eHealth actors.

The main limitations of current pervasive healthcare information systems are the lack of personalization and adaptation to user profiles, to user needs and to different contexts of use [1] and [2]. In general, the most important factors that determine the effectiveness of pervasive services, are the relevance, the quality and availability, where and when needed, of the provided information. However, it is a challenge to provide anywhere and anytime the same quality level of information because of the great variety of scenarios. Some of these scenarios are well known and based on pre-defined protocols and standards, e.g. flight tickets reservation, hotel room booking, etc. But, others are contextual and more complex, e.g. patient tele-assistance or orientation in hostile environments such as geographically critical and isolated areas [3]. In such scenarios, it is extremely difficult to standardize the tasks and processes that meet the users’ needs. In addition, the decisions that should be made are usually subjective and depend on the aptitude and skills of the actors who are involved in the business processes. Therefore, to face this challenge of decision making in non predefined scenarios and to meet the user's needs diversity in different contextual situations, there is an essential need to design models able to capture knowledge about: • the actors who perform multiple tasks in a given domain and in different contexts; • the heterogeneous types of resources required for performing the tasks; • the knowledge related to the organizations that own or manage these resources. It is however a critical issue to design effective and synergetic models that are able to realize knowledge acquisition from actors and other main entities in a given domain. Moreover, to achieve successful knowledge modeling, we should focus on well-defined objectives and exploit the existing technologies, standards, and protocols. To effectively support tasks and processes execution within a pervasive application and to improve information quality in both simple and complex scenarios, we firstly propose a knowledge meta-model based on ontologies that describes the different entities from a given domain. We then construct an architectural framework, called ONtology Oriented Framework for Pervasive Applications and Services (ONOF-PAS) that can be extended and implemented in different specific domains such as eHealth [4]. The proposed knowledge meta-model will have the ability to handle different contexts of use taking into account the availability and the capabilities of the required material, communication, and human resources. The task execution process is based on pre-defined rules that will be applied by an inference engine. The rule base includes logical statements that specify how to handle the contextual situation by linking the context elements such as actor profile, task type, concerned object, its status, the required resources, etc. Meanwhile, the rules shall allow optimizing the management of the priority of the messages exchanged among the different actors. This paper is organized as follows. Section 2 provides a synthesis of research works related to the Task concept in knowledge modeling. In Section 3, we present our methodology to build ONOF-PAS. Section 4 describes the overall knowledge meta-model including its formalization. Section 5 examines how ONOF-PAS can be automatically extended to the telemedicine domain taking as an example a “Patient Transfer” scenario, for supporting various healthcare providers in performing decision processes in telemedicine related tasks. The technical architecture and the empirical results are discussed in Section 6.

The key feature of the proposed ONOF-PAS framework resides in its capacity to perform reasoning on a given knowledge domain taking into consideration the availability and capability of different resources required to perform various tasks and processes, thus enabling an intelligent management of the tasks, processes, resources, and of the messages exchanged among the different actors. It contributes to enhance the quality of the exchanged information in context dependent situations. In our case study, the patient's medical data may be required by different tasks, e.g., tele-expertise or tele-consultation. If the patient's profile is not locally available, then data from the patient electronic health record (EHR) can be retrieved by using web services technology by means of the patient ID or his social security number and the Uniform Resource Identifiers (URIs) of the services provided by the EHR hosts [28]. If need be, ontology mappings can be used to handle the semantic heterogeneity problems [29]. Since ONOF-PAS is ontology-oriented, it is scalable to meet the large amount of information that each of the mentioned dimensions could contain. Additionally, it supports the capability to aggregate information at different levels of abstraction. Furthermore, the ONOF-PAS knowledge meta-model provides a common and generic repository structure that facilitates the exchange of data currently contained within individual tool models of the existing software solutions. The results achieved so far confirm that the approach we propose in this paper opens new perspectives in the eHealth domain, where the extended telemedicine ontological model will enable healthcare actors to carry out different tasks in complex scenarios where the decision making processes are not necessary standardized because of the dynamic context of the resources, the changeability of the status of the objects treated by the tasks, the user's tacit knowledge and behavior, as well as the environment conditions.