Physicians use natural language to record patient data in electronic medical files but they scarcely ever use the coding tools available to them as they find them awkward. As a consequence they do not use standard terminological repositories to encode their inputs, thus failing to meet an essential requirement for the interoperability of health information systems. This hinders accessing and using patients' information for all the automated patient support assistance and medical decision making applications (e.g., aids to therapeutic strategy, contra-indications checking for prescription, patient record synthesis), with plausibly important detrimental effects on individual patients' fates as well as on Public Health indicators. The SIFaDo project intends to design and evaluate methods and tools to make immediately useful and easy -- even possibly fun -- the entry of textual or graphical data to be structured and coded in electronic medical files. It gathers the competences of nine partners : four academic partners from medical and scientific institutions (Medical Informatics, Human-Machine Interfaces, Ergonomics), four industrial partners (Medical Practice Software, Hospital Information Systems, Shared Medical Records, Health Knowledge Bases), and a medical society of general practitioners. The SIFaDo project will bring the design of original methods for the entry of structured and encoded medical data based on a combination of innovating approaches (e.g., reducing the dimension of search spaces for classification, integrating probability models and statistical data, using graphical user interfaces some of which will be based on the iconic language VCM developed by one of the partners). The proposal and the selection of methods will depend on the analysis of users' needs and will be based on the ergonomic evaluation of these methods (user-centric design). Once selected, the methods will be specified and implemented by the industrial partners. The resources needed for the implementation will be made available in a "resource box". The implemented tools will be evaluated in terms of user performance, of encoding quality, and of long term use. The selected methods and the results of their evaluation will constitute the scientific output of this project. The industrial output will consist in several commercial software products for general medical practice and hospital medicine, together with an acquired expertise on project management in the domain of shared medical records.

Activity based funding is an essential part of hospital management. Diagnosis Related Groups (DRGs) is the best-known classification system that is used in this funding model. Such model was adapted in France within the PMSI (Programme de Médicalisation des Systèmes d’information) that takes into account diseases, procedures and treatments. Coding of patient stays should be consistent with the actual care provided. We assume that cross-checks between the electronic health record (EHR) and information coded in the PMSI can improve both PMSI and the EHR. Quality control of PMSI data and quality of care are first carried out by physicians and technicians of coding units, and then verified by physicians from the social insurance in relation to the coding units, and the French High Authority for Health (HAS) through the certification process. These checks involve a fair semantic interoperability between different medical applications, from EHR to PMSI system and vice versa. The objective of the TOLBIAC project (Terminologies and Ontologies for Linking Billing Information and Accurate Clinical data) is to define, model, implement, and evaluate the automated bi-directional consistency between information from the EHR and those coded in the PMSI. In order to achieve this objective it is necessary to: 1) Formalize the rules for analyzing the relationship between the EHR including clinical data from the Health information system on the one hand, and the PMSI on the other hand. 2) Set rules that allow to check the consistency. 3) Implement this methodology in hospital information systems, EHR and PMSI medical applications to: a) ensure the relevance and accuracy of coding by comparing data from multiple sources of information (biological, clinical, administrative, ...); b) assess the financial impact and efficiency of consolidated features. The TOLBIAC consortium consists of 4 public partners (University hospital of Saint-Etienne, University hospital of Nice, CISMeF team from University hospital of Rouen and Hospital of Firminy-INSERM U707) and three private partners (VIDAL, Web100T and InterSystems). The business plan and the consortium agreement will be defined with the help of seven partners. The project lasts 36 months and is divided into seven tasks (T0: coordination and project management; T1: requirements analysis, use cases, T2: formalization of regulatory constraints of the PMSI; T3: modeling methods: representation of management rules for quality and consistency of data using statistical methods and probabilistic approaches, and semantic and linguistic analysis; T4: integration, implementation, prototyping, T5: validation of tools and impact assessment; T6: dissemination and exploitation of the results ). Solutions related to the use cases will be implemented in three pilot sites (1) University hospital of Saint-Etienne in partnership with Web100T using the CristalNet EHR data, (2) University hospital of Nice with a comparison between the medico-economic data entered in the Clinicom software provided by InterSystems and data from the Agfa HealthCare EHR, and (3) Hospital of Firminy – INSERM U707. The result of the project includes production of a set of PMSI-EHR matching rules implemented and validated in pilot sites. The purpose is to ultimately provide a tool capable of producing from clinical data described in the EHR, a transcription of the patient stay as exhaustive as possible in the PMSI using data from all information sources available.

The development of Electronic Healthcare Records (EHR) Datawarehouses and the Personal Health Record leads to a new paradigm in the way the healthcare professionals (as well as patients) interact with medical data. These data, whether structured or full text, are both numerous, accurate, diverse, and immediately accessible at the patient bedside. Because of this profusion of information, the healthcare professionals may face difficulties for making diagnostic and therapeutic decision and patients may misunderstand their health status. They need to have effective tools to locate in real time relevant elements of the patient’s EHR and visualize them according to synthetic and intuitive presentation models. The RAVEL project aims at achieving this goal by performing a high profile industrial research and development programme on the EHR considering the following areas: (i) Semantic indexing, (ii) Information retrieval (iii) Data visualization. The consortium includes both researchers specialized on these topics, persons involved in the implementation of Hospitals Information Systems (HIS) and market leaders private companies. The scientific research program of the project focuses on the improvement of the state of the art on the three mentioned areas. To do so, the research works will attempt to meet the requirements to be defined by business use cases with the medical staff of two University Hospitals. The RAVEL project will lead to the implementation of a prototype sufficiently generic and loosely coupled to data sources in order to be transposed into different university HIS. This prototype will be evaluated in situations of actual use. Many benefits are expected from the RAVEL project. From the medical point of view, it will be the improvement of the decision making of the healthcare professionals and therefore the quality of patients care. From the societal point of view, the methods and tools developed will allow the patients to better take possession of their medical data in order to be proactive in their approach to healthcare. From the scientific point of view, the RAVEL project is a unique opportunity to develop Natural Language Processing (NLP) methods, approaches for the semantic integration of heterogeneous data, retrieval and multimodal representation methods applied to medical data. Finally, from the industrial point of view, the innovations resulting from the project will allow a significant functional and technological leap in a market increasingly competitive.

Prevention is central for the national health strategy. The e-health tools may be useful to change the medical practice and to make patient involved in prevention. The objective of PEPS is to develop computer based tools to make effective the prevention approach for both physician and patient. The patient will have a prevention plan made with the physician and in accordance with the content of evidence based practice guidelines and tools to manage it. The project will provide solutions to the problems related to the collection of interoperable data, construction of decision rules, building customized prevention plans, monitoring of prevention, feedback to the physician and the patient. The consortium brings together complementary skills for this project: 2 Inserm units involved in e-health (LIMICS, CIC-IT Biosensors and e-Health, Innovation and Practice), 3 industrials in the e-health sector (CompuGroup Medical Solutions , Silk Computing, Vidal), and the College of General Practice. The PEPS project is organized around 8 tasks: - a task of coordination (T1), - a task for needs analysis allowing to identify the functional specifications of the platform PEPS, the roles and preferences of the various users and the ergonomic specifications of the interfaces (T2), - a task of data capture modelling and the specification of helping tools for prevention (T3), - a task of knowledge base development containing the rules of prevention and a Web service to reach prevention initiatives from a patient profile (T4), - a task of collaborative construction of a plan of prevention personalized (PPP) from the proposed prevention initiatives and the preferences of the patient (T5), - a task of integration of the PPP in the patient record (T6), - a task of integration in the PPP of tools of documentation, generation of alerts, follow-up, re-motivation on the platform PEPS ( T7) and finally of a task of experiment and validation led together with the CMG and the CISS, Interassociative Collective on the health, so demonstrating the usability and the interest of the platform for the healthcare professional in cabinetor in multi-professional nursing homes and for patient (T8). The expected impacts from the project are important. At the scientific level, the PEPS project will be a contribution to the implementation of solutions of interoperability for the re-use of healthcare data within the framework of prevention. The project is innovative at the methodological level offering a user-centered conception of "persuasive" innovative medical tools. At the industrial level, an enrichment of the patient record will be done by the introduction of prevention services. At the societal level, the platform PEPS will give healthcare authorities the possibility of influencing the healthcare professionals and the population thanks to the availability of indicators and to the better knowledge of the practices in the field of the prevention and will allow the citizen to become a main actor of his prevention in a simple, reliable and guided way.