The aim of the workshop is to promote cooperation between local companies, health authorities and research and technology transfer actors in order to develop innovative solutions based on artificial intelligence in healthcare - especially for the management and analysis of large amounts of data. The objective is to use intelligent algorithms to identify data regularities and their hidden relationships and, consequently, to offer better services and applications. With the event Data4SmartHealth (D4SH), the Smart Data Factory of the Free University of Bozen/Bolzano also wants to promote the development of a community, centered on the NOI techpark, which is committed to e-health, i.e. the use of digital technologies in the healthcare sector. For technology transfer and networking, the laboratory and the Smart Data Factory team are available at NOI Techpark.
8:50 Invited Talk: Explainable Artificial Intelligence in Medicine: a Data-Centric Perspective (Prof. Carlo Combi, University of Verona) (slides)
9:50 Session on Rehabilitation (Chair: Dr. Daniela D'Auria - Smart Data Factory)
11:30 Panel: How will the data shape the future of healthcare and biomedicine? Promising directions and suggestions (Moderator: Dr. Paola Lecca - Smart Data Factory)
12:30 Session on Decision Support (Chair: Dr. Alessandro Mosca - Smart Data Factory)
14:30 Session on Decision Support (Chair: Dr. Alessandro Mosca - Smart Data Factory)
Prof. Carlo Combi, University of Verona
Explainable Artificial Intelligence in Medicine: a Data-Centric Perspective
I will discuss about explainability and interpretability, considering how metadata, specific data constraints, conceptual representations of (even partial) data schemata, and queries could help understand and evaluate black-box machine learning results in medical domains. As an example, it is often the case that clinical features considered in machine learning algorithms are extracted from medical records, without any further consideration of clinical database schema. But it could be that the knowledge derived from the schema and from queries on data, could integrate and allow the right interpretation of ML results.
Ing. Federico Gori - Microgate
Cognition in Motion: a new approach in assessment and rehabilitation
Neurosciences and movement analysis applied to the field of rehabilitation provide new perspectives for patients based on big data analysis. Tracking short and long term movement disorders and cognitive impairment provide a valuable KPI to monitor effectiveness of interventions and large population trends.
Dr. Floriano Zini - Smart Data Factory and Prof. Mauro Gaspari - University of Bologna
AI-supported cognitive training with MS-rehab
The talk presents a web-accessible system, called MS-rehab, developed specifically for the cognitive rehabilitation of patients suffering from multiple sclerosis, but usable also in other contexts where cognitive training is needed. It also illustrates the results of two pilot studies with MS-rehab, including one where an AI-based mechanism is used to adapt the difficulty of cognitive training exercises to the trainee's performance.
Nuove tecnologia per la riabilitazione - Software e sensori inerziali per rendere l'esercizio terapia misurabile e accessibile anche da remoto
CoRehab è un azienda di Trento che progetta e realizza dispositivi tecnologicamente avanzati finalizzati alla riabilitazione ortopedica, neurologica, sportiva e all'esercizio fisico preventivo. I sistemi sviluppati si basano sull'utilizzo di sensori inerziali indossabili validati per l'analisi del movimento e un software interattivo che guida professionisti e pazienti in maniera semplice e veloce nell'esecuzione di esercizi e valutazioni funzionali. L'obiettivo dei dispositivi di CoRehab è rendere l'esercizio terapia misurabile, motivante e accessibile anche dal punto di vista logistico. Allo scopo di rendere la riabilitazione accessibile ovunque si trovi il paziente, CoRehab ha realizzato e lanciato sul mercato una soluzione per la riabilitazione a domicilio che consente al paziente, grazie a una app e un solo sensore indossabile, di effettuare esercizi riabilitativi comodamente da casa con la supervisione da remoto del proprio professionista.
Dr. Bart Geerts - Healthplus.ai
Learnings getting continuous-decision support for surgeons to the clinic
As the covid pandemic painfully illustrated health care systems are being threatened as their sustainability seems limited; Growing number of patients, the complexity of patients, and a lack of highly trained staff necessitate action. We need to reduce cost, improve outcomes and experience for patients and health workers. This seems to be a challenge we can only get out of with new technology aiding us. Computer science and machine learning specifically can address a number of challenges. For instance by taking away menial and administrative tasks, but also improve speed and quality of diagnosis. Another example where ML can be of benefit is decision-support. Healthplus.ai and a dozen hospital partners focus on building PERISCOPE. As up to 35% of all patients encounter an infection in the 30-days after their surgery, we re-use existing electronic patient record data and machine-learning to predict infections before they are diagnosed. This decreases the time to diagnosis, improves allocation of monitoring and decisions to discharge or keep patients in the hospital. As the average diagnosis is on day 5, we ‘buy’ the surgical essential time. A first set of algorithms was 90% sensitive and specific in predicting infections at the end of surgery (i.e. 5 days prior to average diagnosis). In this presentation, we will look at the some of the technical results but we will focus on the hospital, validation, regulatory and medical challenges that face us when implementing decision support in the hospital.
Dr. Panagiotis Symeonidis and Prof. Markus Zanker, Free University of Bozen-Bolzano
Explainable AI for Health
Patients with complex diseases (i.e., cancer, diabetes, etc.) often follow a therapeutic that consists of multiple drugs, focusing at different human targets such as genes, proteins, etc. In this talk, we will describe our research work, which tries to provide both accurate and explainable drug recommendations. In particular, we will describe a method, which can help doctors screen candidate drugs and their possible substitutes more comprehensively, by providing also robust explanations. That is, based on previous similar patients' historical drug treatments, we can provide personalized drug recommendations along with explanations to support critical medical decisions.
Dr. Giuseppe Capasso - BPCOmedia
BPCOmedia: Telemonitoraggio predittivo per la rilevazione precoce dell’insufficienza respiratoria in pazienti affetti da BPCO e da COVID-19
Il dato clinico per stabilire l'inizio delle cure in pazienti con sindrome infettiva da COVID-19 non è la comparsa dell'insufficienza respiratoria acuta conclamata ma la presenza della desaturazione nelle fasi cliniche precoci, quando i sintomi non sono percepiti dal paziente ma è utile monitorare la saturazione emoglobinica (SpO2). BPCOMedia è un sistema di telemonitoraggio della frequenza cardiaca e della SpO2 con pulsiossimetro associato ad un algoritmo di IA che predice gli eventi acuti, sviluppato per la BPCO e adattabile a pazienti COVID-19. BPCOMedia ha una dashboard per il controllo remoto delle misurazione effettuate. Il servizio proposto rende disponibile il sistema esistente per telemonitorare i soggetti sottoposti al regime di isolamento domiciliare obbligatorio, introducendo anche il six minutes walk test per individuare più precocemente segni precoci della polmonite.
Dr. Daniela D'Auria - Smart Data Factory
reCOVeryaID: An intelligent telemonitoring application for symptomatic, asymptomatic and pre-symptomatic coronavirus patients
COVID-19 is an infectious respiratory disease caused by the virus called SARS-CoV-2 belonging to the coronavirus family. In the course of the disease, after an initial phase with a flu-like course, a very strong respiratory syndrome may arise linked to the development of bilateral interstitial pneumonia, whose symptoms lead to breathing difficulties, dyspnea, breathlessness, and increased heart rate. In fact, COVID-19 pneumonia leads to a decrease in the level of oxygen in the blood (saturation) without the patient being aware of it, until the urgency of hospitalisation occurs. Therefore, there is a need to monitor subjects at home who are under observation for COVID-19 infection, in order to control the level of saturation, so that it does not fall below the established threshold, especially in the absence of previous diseases affecting the respiratory system. In such situations, thanks to the trend of the saturation data, the medical staff will be able to understand if that patient, asymptomatic, symptomatic or pre-symptomatic and in home isolation, will have to be hospitalized or not, thus arriving at an early hospitalization before the clinical picture can worsen.
Event realized with the contribution of Microgate.
Floriano Zini is an academic technologist at Smart Data Factory, the technology transfer laboratory of the Faculty of Computer Science at the Free University of Bozen-Bolzano. Over the past 10 years, his research activity has focused on the design, development and evaluation of mobile and web-based information systems for personalized health care. Personalization is achieved through the application of Human Computer Interaction and Artificial Intelligence methodologies and algorithms, particularly Machine Learning. Floriano holds a Ph.D. in computer science and has authored about 60 scientific papers on e-health, personalized information systems, grid computing, multiagent systems and evolutionary computing. According to Google Scholar, his h-index is 21.