service definition and development for Patients and Citizen. 1 Introduction ..... the custom applications installed in the mobile devices or PC. The user can send ...
Innovative Healthcare Services for Nomadic Users Marcello Melgara1 , Luigi Romano1 , Fabio Rocca1 , Alberto Sanna2 , Daniela Marino2 , and Riccardo Serafin2 1
2
Atos Origin Italia S.p.A., 11026 Pont-Saint-Martin (AO), Italy {marcello.melgara,luigi.romano,fabio.rocca}@atosorigin.com San Raffaele Scientific Institute, e-Services for Life and Health, 20132 Milano, Italy {alberto.sanna,daniela.marino,riccardo.serafin}@hsr.it
Abstract. Mobile users require location based, situation aware services, especially when healthcare is involved. Within “Nomadic Media” an Eureka-ITEA International projects, Ontology based semantic web service discovery and orchestration have been studied and applied to provide mobile users with innovative healthcare services. The designed system identifies, orchestrates and customises the services, according to the health status of the user and his usage location, usage conditions and environmental situations. The paper will describe the developed technologies and the implemented services. Similar approach was followed in PIPS: “Personalised Information Platform for Life & Health Services”. However in PIPS more emphasis has been applied in the definition and the development of Use Cases and service definition and development for Patients and Citizen.
1
Introduction
The PIPS Project3 is an ongoing Integrated Project within the IST 6th Framework Programme, coordinated by Fondazione Centro San Raffaele del Monte Tabor. Together with FCSR and Atos Origin, other 15 Companies and Universities from Europe, Israel, Canada and China cooperate to it. The PIPS project objective is to encompass the entire set of business processes, professional practices, and products, applied to the analysis and preservation of the citizen’s well-being, using the latest innovations in ICT. The project joins healthcare (HC) suppliers, citizens, public organizations, food/drug industry and services, researchers, and health related policy makers. These actors create a dynamic knowledge environment that feeds the system and gives added value feedback for personalized contextual knowledge and services to improve the European public’s wellbeing. In the PIPS context each actor is a supplier and receiver of personalized knowledge. This includes both explicit and implicit knowledge management based on 3
IST 2004 507019 PIPS: Personalised Information Platform for Life & Health Services, www.pips.eu.org.
traditional and new approaches to knowledge discovery out from current medical practice, evidence based medicine, and disparate knowledge sources from health/nutrition domains. The Nomadic Media4 consortium was composed by different European companies such as Nokia, VTT, Philips Digital Systems, Cefriel and Atos Origin. The Nomadic Media Project aimed to enhance consumer flexibility in the use of services and contents in the places they wish, and to enable the movement of content between their preferred devices according to their needs and circumstances. Implicit in this vision is also the need for consumers to configure services and content in the ways that suit their particular circumstances and thus enjoy the benefits of a personalized environment.
2
eHealth Services
In the following chapters we describe the approaches in the multimodal and pervasive eHealth services adopted in the two research projects Nomadic Media and PIPS. 2.1
The PIPS Service Oriented Approach
PIPS implemented solutions enable: – HC Professionals to deliver just-in-time personalized and prevention-focused HC services compliant with the Citizen’s personal health state, preferences and ambient conditions – Citizens to make informed decisions about therapies and nutrition at any time and place according to the real-time evaluation of their health state – HC Authorities to improve risk management of HC systems The current PIPS project implementation foresees three scenarios: Diabetes, Hearth Failure, and Nutritional. Figure 1-A depicts an example of possible functionalities that could be implemented for the scenarios. In order to better explain which kind of services the Platforms allows, we are going to present one fictional scenario, completely supported by the current version of the PIPS Platform. In the remaining of this section we will then present the technical architecture employed by the Project, which allows to implement such scenario. Our main actor for this scenario will be John Fitzegerald, a 55 year old person who is currently under treatment for an ischemic cardiomyopathy with heart failure (HF) complication. This type of disease usually impairs the functional capacity and quality of life of affected individuals. For this kind of patient it is mandatory to monitor vital signs and the arise of symptoms indicating possible disease accentuation [1]. For this reasons, John 4
Eureka E!2023 - ITEA if02019: Nomadic Media: Entertainment at home and on leave. ITEA - Information Technology for European Advancement, http://www.iteaoffice.org
Fig. 1. PIPS Scenarios and possible services (A) and PIPS overall architecture (B)
is following a complex treatment made of different components: he has to take a quite large number of pills daily, he must measure a fixed set of vital signs every morning and he is on a diet. Therefore, the first support that PIPS provides to John is a reminder service, set with respect to all his prescriptions and integrated with his mobile phone agenda, that reminds him of all actions he must perform on a timely basis. In John’s case, he must measure his weight, blood pressure and heart rate every day before breakfast. Monitoring these signals daily is fundamental to keep his overall health status controlled and to react properly to any abnormal condition, for instance to accumulation of the body fluids. For this scenario let’s suppose that John was out of town for the weekend, exceeded a little with drinking and forgot to take some of his pills. When he receives the reminder on Monday morning he uses the measurement devices provided by PIPS and collects the needed vital signs. These devices are all connected wirelessly to John’s home network and the data collected are immediately sent to the PIPS system via a web service. In John’s case, the analysis of the collected vital signs (increased weight, low blood pressure, high heart rate) indicates a body fluids accumulation, which is confirmed by the symptoms collected using an online questionnaire (weakness, shortness of breath, reduction of diuresis, etc.) and additional vital signs (low oximetry). As a result, he is suggested to resume his diet (strictly adhering to his water intake regime) and a new therapy with an augmented dosage of diuretic is set by the system for him (updating his agenda). The diuretic therapy change is prescribed beforehand by John’s doctor who, along with the normal dosage, will provide the augmented dosage to be taken in case of suspected fluid accumulation.
John’s doctor is also warned of the situation. On her PIPS Portal home page, or cellphone if so configured, she will receive a message that John’s therapy has been changed according to her indication and she will be able to check the collected data (both vital signs and questionnaire answers). At that point she can take an informed decision and react appropriately (for example, fixing John’s therapy, contacting and reassuring him or inviting him for a checkup). This scenario demonstrates some of the key benefits that a system like PIPS can provide: constant monitoring of patient health status enabled by the integration of self-care networked measuring devices, complex decision support system and new communication technologies, just in time abnormal health condition detection for chronic patients, integrated access to patients continuity of care records, personalized advices to improve health status. Moreover, all these functionalities are personalized according to the context, the situation and the adopted communication devices. The architecture applied for PIPS system is based on different tiers as shown in Figure 1-B: Front End, Business, Decision Support System, Knowledge Management. – Front End Tier: it is the collector of the requests incoming from users that can interact with PIPS system through the Portal web application, personal mobile devices and medical devices. – Business Tier: it implements the Business Logic of the PIPS system, interacting with Front End tier and Decision Support System – Decision Support System (DSS): it is the technological core of the delivered PIPS system and is based on multi agent platform [2]. DSS processes the information incoming from users in order to monitor the Patient health and wellness status and provide suggestions that should be performed by the PIPS actors [3]. – The Knowledge Management (KM): it provides Knowledge Base information that can be used by DSS and by the final user in order to retrieves trusted e-health information [4]. The DSS is the technological core of the delivered PIPS system and is based on multi agents platform and rule engine. The analysis and design of the system was based on the notion of computational organisations, whereby the roles played by the agents are modeled on the basis of the roles in real-world organisations. For the high-level analysis and much of the design was guided using the Gaia methodology [5], in which roles are a central concept. The model is based on real-world health-care systems. Using the Gaia methodology, two main roles within the PIPS DSS were derived, personal and specialized agents, as follows: citizens have personal advisory agents, what in many organisations are “personal assistants”, which provide personal information to them, such as their diary, or their medical history; specialised agents represent the health care specialists and nutritional experts that assess the health and diet of the citizen and provide advice or information for the nutritional and medical fields in which the PIPS system specialises, such as diabetes or heart problems.
2.2
The Nomadic Media Service Approach
To reach its aim the Nomadic Media project explored different usage scenarios: “At the Airport”, “On-the-Go”, “At Home” and “Healthcare”. The Healthcare scenario was the one in which Web Services (WS) [6] were investigated. For this scenario a technological architecture was defined to be able to: – Connect a variety of related services into a coherent set and thereby improve the process of, for example, ordering prescriptions, making patient appointments, and scheduling laboratory tests by healthcare professionals – Enhance physician productivity with real-time access to information via a variety of preferred devices, adapted to the variable usage condition (situational awareness) – Allow collaborative access for different users such as insurance companies, healthcare providers, drug companies and patients. As a result of this investigation we realized that the healthcare context shares many characteristics that are common to many complex and distributed applications. This led to the following summary of the key problems: – Services should be composed at runtime, based on a multiparty business process model, using an orchestration framework. – To enhance the choices and to dynamically compose the services, advanced techniques should be used to advertise and discover them. – Content and level of services provided should be adapted in relation to context, situation and user preferences. A Service-oriented Architecture (SOA) is a possible way to solve the above mentioned problems, that is to say an architecture where functionalities are implemented, essentially, as a collection of services communicating with each other. A service is a function that is well-defined, self-contained, and does not depend on the context or state of other services. As known, SOA is not new, but is an alternative model to the more traditionally tightly-coupled object-oriented models that have emerged in the past decades. Web Services (WS) represent a set of specifications defining the details needed to implement services and interact with them. Although WS is a technology in development and standardization efforts are not completed, robust enterprise toolsets are available: industrial WS based solutions can be developed in specialized areas. One of the main areas explored in Nomadic Media was service composition. It allows complex tasks to be executed as sequences of processes written using standard specifications. In the Nomadic Media Healthcare scenario the need to exchange information between different providers (services, content and context providers) was clearly identified. WS standards were used to solve the problem of obtaining robust multiparty interaction. Standards also helped to solve the interoperability problems at a syntactic level; however, the real strength of WS technologies is the possibility to afford the heterogeneity of the systems in a semantic way too. To achieve semantic interoperability, information systems must be able to
exchange data in a way that allows ready accessibility to the precise meaning of the data and the data itself can be translated by any system into a form that it understands. This was achieved by describing, functionally and operationally, services in a formal, machine-readable way.
Fig. 2. Nomadic Media Structure
Semantic interoperability enables the automation of some procedures: – Web Service discovery; the action of matching available service descriptions to a requester’s candidate service query and returning the resulting matches, – Web Service invocation; the principle of automatically interacting with an atomic service by using the semantic description to understand how to access it, – Web Service selection and composition; the action of choosing the most suitable service(s) from a set of known services and running a composite service by invoking WS, in the correct order, overcoming syntactic, structural, semantic and process heterogeneity, and handling errors and exceptions, – Web Service execution monitoring; the principle of tracking what is happening to some described aspects of a service and its component services. One possible way to proceed toward the semantic interoperability is to provide semantics using metadata and ontologies. The two major efforts in defining a Semantic Web Services Language [7] are SWSL developed by the OWL-S (Web Ontology Language for Services) [8] committee in the USA, and WSML (Web Service Modeling Language) [9] developed by the Web Service Modeling Ontology [10] project in the EU. A
table presented in [11] summarizes the comparison between WSMO and OWLS. However the approach to semantic web services does not only engage the language problem, but also the architecture. Since the WSMO project also proposes a framework (WSMF) to solve semantic interoperability, we chose this approach in our project. Its philosophy is based on two (complementary) principles; maximum de-coupling of its components, and a scalable mediation service. One of the important points we found in the WSMO was the possibility of enabling support for static and dynamic composition. The composition approach we chose was defined as Orchestration. Orchestrator is the one who defines how the overall functionality is achieved by the cooperation of more elementary services. In the orchestration approach a workflow process invokes a number of different services in a specific order because they have data and control dependencies between each other. Infrastructure based on this approach just requires one central service that is the workflow engine that controls and executes the entire workflow process. The engine we used executes an XML-based [12] script language: BPEL4WS [13]. BPEL4WS (Business Process Execution Language for Web Services) provides a language for the formal specification of business processes and business interaction protocols. By doing so it extends the WS interaction model and enables it to support business transactions.
2.3
Multimodality and Content Adaptation
Multimodality is the technology whose purpose is to enhance the user experience by enabling service providers to combine different ways to build-up intuitive and powerful applications. For users, multimodality represents an efficient way to interact with a device. For network operators, the combination of audible and visual functions represents the future of communications. Soon, applications such as the mobile ones, will take advantage of multiple simultaneous channels of communication, leading to a new wave of service offerings. Some new standards are focused on multimodality technology [14] [15]. Multimodality and adaptation of contents cover an important role in Health services. Multimodality allows the possibility to transfer information through different channels, while adaptation can be read according to a more general meaning, depending not only on the features requested by the device (hardware and software), but also on particular contexts in which the request is made. A generic eHealth framework should manage situations in which the user cannot receive complete information, but just an abstract of them, or he would like to temporally freeze and save this information. In particular, when user’s presence is revealed, his state is communicated to the system, together with his preferences and his possibility to make use of nomadic services; for example it is possible that a user would not like to be disturbed by unexpected communications.
Besides, the user physical collocation makes easier the service interaction: information related to the structure in which the user is located, are immediately provided (i.e. pharmacies, hospitals) so he can be better assisted in his choices. The PIPS system offers to final users different access mode in order to acquire information data, related to vital signs user data, food and drug information. The user can provide data to the system manually, through web Browser UI, or mobile application (e.g. transcript the vital signs values from traditional device to the web form provided by system). The same type of data can be directly sent to PIPS system if are used wireless medical devices connected to the custom applications installed in the mobile devices or PC. The user can send to the system food and drug information data interacting in different mode: optical mode, RFID technology. For example using the camera of user mobile phone, the picture of the product bar code is processed by an OCR, in order to send the food information data directly to the PIPS system. The same information could be sent to the system by filling the web form provided by the mobile application with the barcode number. Also the RFID technology is used if the food or drug product data are store in tag RFID. The user can also provide inputs to the system, writing on a paper, using an optical pen connected via Bluetooth to the mobile device. Also in this case the data can be provided automatically to PIPS system. In Nomadic Media the solution approach is based on an adaptation engine service: it uses the client features to dynamically support the device context. The page rendering is performed at run time, whenever possible, depending on wireless device multimedia capabilities. The Nomadic Media Framework, using database devices service [16], is able to display images and adapted content. It supports, also, Voice Interaction [17] used for manage healthcare vocal application.
3
Conclusions
The technologies studied and applied in PIPS and Nomadic Media have shown their powerfulness to build Healthcare and Wellness Services for Citizen and Patients. This technological enablers have allowed to define innovative complex Service models, derived and sustained by a strong Business Strategy. In order to make the services more and more intelligent and automated, machine processable data have been adopted, to deal with dynamic content and services. In order to enable the interoperability among systems, a framework using machine processable data, rather than more human oriented data, should be developed. Process ontologies, rule based Multi-Agent environments, and ontology bridging are the most promising avenues for integrating automation and orchestration. According to the practical experience, the technology required to make eHealth services more “intelligent” (meaning that they are tailored on user) should be explored based on the work in the field of semantic web and multi agent platform.
Taking into account the different environments in which eHealth and Wellness solutions may be exploited, the Service should automatically adapt their behavior to the current environment (Situational Awareness) and allow to use the most suitable communication channels and mode (Multi-Channel Multi-Modal Interaction). The aforementioned development lines should allow to concretely achieve the condition that Tim Berners-Lee described as the “Next Generation Web” [18].
References 1. Hunt, S., Abraham, W., Chin, M., Feldman, A., Francis, G., Ganiats, T., Jessup, M., Konstam, M., Mancini, D., Michl, K., Oates, J., Rahko, P., Silver, M., Stevenson, L., Yancy, C.: ACC/AHA2005 guideline update for the diagnosis and management of chronic heart failure in the adult: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Update the 2001 Guidelines for the Evaluation and Management of Heart Failure). (American College of Cardiology Web Site. Available at: http://www.acc.org/clinical/guidelines/failure//index.pdf) 2. Wooldridge, M.J., Jennings, N.R.: Intelligent agents: Theory and practice. Knowledge Engineering Review 10(2) (1995) 115–152 3. Dominguez, D., Grasso, F., Miller, T., Serafin, R.: PIPS: An Integrated Environment for Health Care Delivery and Healthy Lifestyle Support. In: 4th Workshop on Agent applied in Healthcare, ECAI2006, Riva Del Garda (2006) 4. Goczyla, K., Grabowska, T., Waloszek, W., Zawadzki, M.: The knowledge cartography - a new approach to reasoning over description logics ontologies. [19] 293–302 5. Zambonelli, F., Jennings, N.R., Wooldridge, M.: Developing Multiagent Systems: The Gaia Methodology. ACM Transactions on Software Engineering Methodology 12(3) (2003) 317–370 6. Booth, D., Haas, H., McCabe, F.: Web Services Architecture. http://www.w3.org/TR/ws-arch/ (2005) 7. Battle, S., Bernstein, A., Boley, H., Grosof, B., Gruninger, M., Hull, R., Kifer, M., Martin, D., McIlraith, S., McGuinness, D., Su, J., Tabet, S.: Semantic Web Services Language (SWSL). http://www.w3.org/Submission/SWSF-SWSL/ (2005) 8. OWL-S: OWL Web Service Ontology. (http://www.daml.org/services/owl-s/) 9. WSML: Web Service Modeling Language. (http://www.wsml.org/) 10. WSMO: Web Service Modeling Ontology. (http://www.wsmo.org/) 11. Roman, D., Keller, U., Lausen, H., de Bruijn, J., Lara, R., Stollberg, M., Polleres, A., Feier, C., Bussler, C., Fensel, D.: Web Service Modeling Ontology. Applied Ontology 1(2) (2005) 77 – 106 12. Bray, T., Paoli, J., Sperberg-McQueen, C.M., Maler, E., Yergeau, F.: Extensible markup language (xml) 1.0 (third edition). (W3C Recommendation, http://www.w3.org/TR/2004/REC-xml-20040204/) 13. BPEL4WS: Business process execution language for web services. (http://www128.ibm.com/developerworks/library/specification/ws-bpel/) 14. Larson, J.A., Raman, T., Raggett, D.: W3C Multimodal Interaction Framework. http://www.w3.org/TR/mmi-framework/, http://www.w3.org/TR/mmiarch/ (2003)
15. Barnett, J., Bodell, M., Raggett, D., Wahbe, A.: Multimodal Architecture and Interfaces. http://www.w3.org/TR/mmi-arch/ (2006) 16. WURFL: Wireless Universal Resource File. (http://wurfl.sourceforge.net/index.php) 17. VoiceXML: Voice eXtensible Markup Language. (http://www.voicexml.org/) 18. Berners-Lee, T.: Xml and next generation web publishing. http://www.w3.org/Talks/1998/0318-Seybold-timbl/overview.htm (1998) 19. Wiedermann, J., Tel, G., Pokorn´ y, J., Bielikov´ a, M., Stuller, J., eds.: SOFSEM 2006: Theory and Practice of Computer Science, 32nd Conference on Current Trends in Theory and Practice of Computer Science, Mer´ın, Czech Republic, January 21-27, 2006, Proceedings. In Wiedermann, J., Tel, G., Pokorn´ y, J., Bielikov´ a, M., Stuller, J., eds.: SOFSEM. Volume 3831 of Lecture Notes in Computer Science., Springer (2006)
