Functional specifications of Data Processing and Decision Support Services

Colantonio S., Martinelli M., Moroni D., Asirelli P., Barcaro U., Bozzi E., Chimenti M., Pieri G., Salvetti O, Chiarugi F., Sakkalis V., Candelieri A., Conforti D., Costanzo D., Guido R., Lagani V., Perticone F., Sciacqua A., Gamberger D., Prcela M.

Decision Support Services 

HEARTFAID aims at defining efficient and effective health care delivery organization and management models for an "optimal" patients' management in the field of cardiovascular diseases. An informative and decision support platform of services will be devised for improving all the processes related to diagnosis, prognosis, treatment and personalization of health care of the Heart Failure (HF) in elderly population. To this end, innovative results on computational modelling, knowledge discovery methodologies, visualization and imaging techniques, and the medical knowledge of the relevant domain will be opportunely integrated to design and develop an effective and reliable Clinical Decision Support System (CDSS): the HEARTFAID CDSS, corresponding to the core of HEARTFAID intelligence. This system will be able to process clinical knowledge and patient-related information, intelligently filtered, processed and presented at appropriate times, to enhance patient care. It will be devised as a service of the HEARTFAID platform for providing an effective support to the daily practice of the clinicians, by implementing adequate data processing algorithms, by providing guidelines to medical protocols as well as access to the knowledge base, by sending alarms in case of critical situations, and by supplying diagnostic suggestions. Two peculiar issues are involved in the development of the HEARTFAID CDSS for supporting medical decision making, i.e. innovative approaches for biomedical signal and image processing; robust and reliable reasoning approaches, based on Machine Learning and inference methodologies on declarative and procedural domain knowledge. The first, fundamental step towards the design and development of the HEARTFAID CDSS consists in analysing the requirements and specifying the functional details of both data processing and decision support services. These topics are the focus of this deliverable, which collects the results of Task T5.1, "Identification of representation features for signals and images processing", and Task T5.3, "Requirements and functional specification of the Decision Support System", of Work Package WP 5. A deep investigation into issues inherent in (i) data typologies and their characterization, (ii) reasoning methodologies, (iii) HEARTFAID CDSS logical- functional vision, and (iv) current available technologies are herein provided, with the final aim of supplying the guidelines for the development of the HEARTFAID data processing and decision support services. The document layout has two main "souls", addressing, from one side, all the methodological and technological aspects of signal and images processing, and from the other, all the topics related to decision support. Moreover, it is ideally divided into two parts: in the first part (Chapters 2, 3, and 4) the discussion consists of a revision of the State of the Art; whereas in the second part (Chapters 5, 6, 7, and 8) the attention is focused on the HEARTFAID specificity, regarding the final goal of the deliverable. More in detail, the document is organized as follows. At first, an introduction supplies an overview of the problems to face for reaching the HEARTFAID Grand Vision about the intelligent support to the medical decision making. In Chapter 2, the available diagnostic resources for signals and images acquisition are introduced, analysing their peculiarities, relevance to the HEARTFAID clinical domain, and open-problems. An overview of the principles of signal and image processing and of the mathematical models for features extraction is also provided. Afterwards, in the third chapter, a careful survey of the methodological foundations of decision support systems (DSS) is reported, discussing the main aspects of decision theory and typical DSS models and structures. According to HEARTFAID definition as a knowledge-based platform of services, medical domain knowledge plays a fundamental role in HEARTFAID decision support development. Thus, particular attention is reserved to the category of Expert Systems or Knowledge-based DSS, and related issues, such as knowledge representation and inference engine modelling. Clinical applications and topics inherent to CDSS, e.g. guidelines modelling, are widely addressed as well, for better understanding the design implications involved. The chapter ends with a discussion about the important issues regarding DSS design and success factors, which should be taken into account when devising the HEARTFAID CDSS. Chapter 4 reports a critical description and analysis of the available up-to-date technologies for both data processing and decision support services. In Chapter 5, the biomedical parameters, relevant to HF and selected by the medical partners, are critically reviewed in the perspective of data processing. From this analysis, the most relevant problem in data processing w.r.t. HEARTFAID platform is selected and the functional specifications of the planned signals and image processing are described. The HEARTFAID clinical decision making problems are deeply investigated in Chapter 6, in order to identify the overall objectives of the HEARTFAID CDSS, i.e. what it should accomplish and why, specifying the element of the knowledge to be formalized. The system requirements are defined responding to HF clinicians' needs, as also stated in the deliverable D5, in order to devise an efficient and usable system for supporting the medical personnel in their daily activity.

Source: Project report, HEARTFAID, Deliverable D15, 2007