Global (English)
Giovanni Ferrando, Genoa, Italy, 1965.
An electronic engineer, he is Product Manager for Enterprise Imaging Line at Ebit, an Esaote Group company focusing on Medical IT. He has been working in the design of PACS systems for more than 20 years, with specific expertise in the field of radiological and cardiological imaging as well as in radiotherapy solutions. He loves sports and travel.
A data system is a bridge between patient care and societal well-being. AI intervention is essential but must be designed in such a way that it is left to human creativity to come up with solutions.
We are all producers and consumers of information that, starting with data (the raw material of knowledge), follow paths and processes that give shape to awareness, a tool for orientation in all fields. In healthcare, the new opportunities for data collection afforded by increasing technological capacity have meant that the amount of data available has grown exponentially. This opens up opportunities for a significant leap forward in terms of the well-being of the community in which the company operates but, at the same time, considerably increases complexity as a result of the introduction of new types of data (genotypic and phenotypic).
In addition, the introduction of European privacy regulations has forced every producer, manager, and consumer of data to question the quality of the long process by which information becomes shared knowledge.
Ebit, an Esaote Group company focused on information systems, has been protecting and organizing the data collected during the diagnostic process in cardiology and radiology since the 1990s. Its systems enable the collection and management of sensitive data related to health, which represent an asset for the patient.
This asset must be leveraged by ensuring a balance between confidentiality and accessibility, which helps to grow the relationship of trust between doctor and patient, based largely on respect for privacy, timeliness of intervention, and a consistent view of the clinical and personal history of those on a healthcare journey.
All the information collected is a representation that offers the physicians all the elements they need to be able to interpret the patient’s clinical picture in the best way possible, but often this information can appear at first glance as an abstract picture, as complex, dispersive, and unfocused. The objective of the people who work at Ebit is to create tools that will assist clinical practice, by enabling the simple and immediate navigation of complex information, such as that taken into account during the process of interpreting clinical signs and symptoms, known in medicine as semiotics.
As part of the design work, great care is taken to ensure that these tools preserve the centrality of the role of the physician as the main actor in the diagnostic and therapeutic decision-making processes. The ability to make data available, ensuring adequate levels of continuity of service and redundancy to patients and clinicians who follow up the patient care journey, requires a high level of complexity – something which Ebit has a long and extensive experience of managing, drawing on its leadership position in the Italian and international market.
Ebit faces this challenge by using the latest cloud and cloud storage technologies and ensuring full support and compatibility with international standards of integration in the medical field. Through the cloud distribution portal or through integrations with regional/company Electronic Health Records, individuals can access their own reports with full security and confidentiality and can decide whether they want to share them with their own doctor. Thanks to the Vendor Neutral Archive and its documents and bioimages, medical staff can access the patient’s entire medical history from a single, centralized repository.
These powerful tools provide a large amount of data, both textual (medical history data, diagnostic questions, reports, laboratory tests) and iconographic (images), which must be evaluated by clinical staff, sometimes urgently. The same data must then be made available to the patient (and in turn to the treating physician) within the report, in a format that follows a template in accordance with the recommendations of the national and European societies of radiology and cardiology.
Hence the need for simplification, which can be effectively addressed using AI techniques. Within our solutions, artificial intelligence is used to help professionals increase their confidence, facilitate decision-making, and foster dialog.
Attention is given to optimization of the workflow through prioritizing (triaging) the reading of diagnostic investigations, and to data analytics for real-time insights into clinical and administrative data.
Particularly challenging is the use of natural language processing (NLP) techniques used to extract relevant information from the texts of the reports, essential for the production of structured reports, in accordance with the standards proposed by medical societies and with their classification for scientific and research purposes. The challenges in this area are complex yet rewarding and involve the use of an innovative approach based on pre-trained neural language models. This approach has the advantage of being easily extendable to different domains and report types and has extensive reading comprehension capabilities (negation, synonyms, abstraction, etc.).
However, AI brings with it problems relating to the availability of an adequate amount of information required to train the network, in order to reach a minimum level of generalizability, and relating to legislative and ethical aspects (think, for example, of identifying where responsibility lies in the event of an incorrect prediction during the use of such tools).
Ebit’s mission in its approach to data complexity management is to keep the human being at the center of the decision-making process, providing a portfolio of solutions that simplify access to information, focusing the attention of the clinicians on what is most relevant to their work.
A further guarantee of our human-centric approach is Ebit-Esaote’s membership of the European regulatory system, the first to set out to create a system of ethical rules to safeguard democracy and privacy in the face of the algorithm. The “European way” calls for algorithms to be fed with data that reflect European values of privacy and individual freedoms, and this makes careful data design even more strategic. Another goal, in order to guarantee the involvement of the person, is to overcome algorithmic systems characterized by black boxes; the data must be entered into systems that allow an understanding, even for non-experts, of how they work and how they make decisions. In this way, data complexity will take forms that are increasingly able to simplify access to awareness.
