Ebook: Telehealth Ecosystems in Practice

The 2023 European Federation for Medical Informatics (EFMI) Special Topic Conference (STC) was held in Turin, Italy, on October 25–27, 2023. The Scientific Programme Committee (SPC) was co-chaired by Mauro Giacomini, Head of the Health Informatics Laboratory at the University of Genoa, Italy, and Chair of the EFMI Translational Health Informatics Working Group, and by Lăcrămioara Stoicu-Tivadar, University Politehnica Timisoara, Romania.

The theme of STC 2023 was Telehealth Ecosystems in Practice.

Telemedicine concerns all health services provided remotely, and is seen as an innovative medical practice in contrast to traditional face-to-face interactions. It allows for the breaking down of geographical distance and aims to equalise access to care by using information and communication technologies (ICTs), thereby enabling the secure transmission and sharing of medical data and information for the monitoring and controlling of patients’ clinical status. Throughout the last two decades, the European Community has been supporting telemedicine through the funding of several research projects powered by technological development and this has prompted an increased interest in telemedicine. But despite the opportunities and benefits related to telemedicine services, their large-scale spread has mainly been slowed down to date by the high cost of technologies, the absence or inadequacy of laws related to eHealth and privacy, the poor capability of elderly patients to use ICT, the frequently unpredictable evolutionary speed of patient status, and sometimes also by the lack of qualified actors. In order to support the recovery and resilience of Member States, the European Union has approved the Next Generation EU programme, which allocates €750 billion. One of the aims of this ambitious project, which will end by 2026, was to improve the digital transition of healthcare systems. All beneficiary countries have defined their priorities but in many cases, have also allocated substantial resources to strengthen their national health systems, enhancing the protection against environmental and climate-change-related health risks with the aim of better responding to the needs of communities with regard to local care and assistance. Local healthcare assistance is, in many cases, fragmented and subject to regional disparities, which can result in different levels of healthcare provision and health outcomes across regions. The provision of integrated home-care services is considered to be low, and the different healthcare and social service providers are considered to be only weakly integrated in many national plans. One of the answers most relied upon to solve these problems is telehealth. For this reason, we have decided to dedicate this conference to an international exchange of views on this topic, focusing both on the theoretical content, but also on the perspectives of real applications of innovations derived from scientific and technological research.

These proceedings present full papers and short communications covering a broad range of topics and methods in telemedicine, as well as a smaller selection from the broader sub-domains of biomedical informatics. The proceedings are published online with open access and indexed in the major bibliographic databases such as Medline and Scopus to ensure visibility to the wider scientific community. We would like to thank all members of the SPC for their hard work and commitment in managing the submissions and the programme: Gabriella Balestra, Arriel Benis, Stefano Bonacina, Alessio Bottrighi, Thomas Deserno, Parisis Gallos, Lenka Lhotska, Sara Marceglia, Alejandro C. Pazos Sierra, Samanta Rosati, and Lucia Sacchi. We also thank all of our authors for submitting their work, our peer reviewers for volunteering their time and expertise to ensure the quality of the programme, and, once again Gabriella Balestra and Samanta Rosati for leading the work of the local organising committee, which handled all the practical arrangements for the conference.

Mauro Giacomini and Lăcrămioara Stoicu-Tivadar

SPC Co-chairs

Abortion remains a highly controversial topic in many countries, particularly in the United States. As the COVID-19 pandemic introduced new challenges and restrictions, society saw a marked increase in demand for self-managed care. Likewise, the utilization of abortion care via telemedicine sparked interest, especially in communities with high infection rates. However, as unregulated online forums became an outlet for the discussion of sensitive health-related topics, the spread of false and misleading information markedly increased. As patients continue seeking reliable health-related information, personalized solutions are needed to provide accurate, evidence-based insight. As a pillar in digital precision health, precision health promotion via Personal Health Library (PHL) could aid in equipping patients with the necessary information to support informed health decision-making. In previous works, we have proposed the utilization of a PHL for the self-management of disease and health promotion/education. Herein, we introduce our work-in-progress in implementing the PHL-Enabled Abortion Care and Education (PEACE) platform for facilitating and supporting reliable access to informative reproductive care, such as abortion via telemedicine.

Patient involvement in research has been highlighted as a major requirement for the development of products and services that cover actual patients’ needs. However, there has not been an agreement on a commonly used standard for patient involvement in research, at least not in the EU, partially because of lack of common terminology and implementation methodology. Within the standardization activities of “LifeChamps: A Collective Intelligent Platform To Support Cancer Champions”, this qualitative study was developed to discover patients’ views for their engagement in research. This is an ongoing qualitative study of semi-structured interviews of cancer survivors aged over 65 years of age, exiting the feasibility studies of the LifeChamps project in Stockholm and Thessaloniki. Findings from the thematic analysis of this study are expected to indicate requirements for involvement of patients in research studies as participants.

COVID-19 impact on population mental health has been reported around the world. Statistics Canada has conducted a survey among Canadian population to gauge mental health challenges they experienced, specifically in terms of anxiety. We create a machine learning model to predict anxiety symptoms as measured by the General Anxiety Scale among the sample of 45,989 respondents to the survey. Eight algorithms including Logistic Regression, Random Forest, Naive Bayes, K Nearest Neighbours, Adaptive boost, Multi linear perceptron, XGBoost and LightBoost. LightBoost provided the highest performing model AUC score (AUC=87.45%). In addition, the features “perception of mental health compared to before physical distancing”, “perceived life stress”, and “perceived mental health” were found to be the most important three features to predict anxiety. A limitation of this study is that the sample is not representative of the Canadian population. Preparing for virtual care interventions during a crisis need to take into considerations these factors.

Major Depressive Disorder (MDD) has a significant impact on the daily lives of those affected. This concept paper presents a project that aims at addressing MDD challenges through innovative therapy systems. The project consists of two use cases: a multimodal neurofeedback (NFB) therapy and an AI-based virtual therapy assistant (VTA). The multimodal NFB integrates EEG and fNIRS to comprehensively assess brain function. The goal is to develop an open-source NFB toolbox for EEG-fNIRS integration, augmented by the VTA for optimized efficacy. The VTA will be able to collect behavioral data, provide personalized feedback and support MDD patients in their daily lives. This project aims to improve depression treatment by bringing together digital therapy, AI and mobile apps to potentially improve outcomes and accessibility for people living with depression.

Background:

Artificial intelligence (AI) can potentially increase the quality of telemonitoring in chronic obstructive pulmonary disease (COPD). However, the output from AI is often difficult for clinicians to understand due to the complexity. This challenge may be accommodated by visualizing the AI results, however it hasn’t been studied how this could be done specifically, i.e., considering which visual elements to include.

Aim:

To investigate how complex results from a predictive algorithm for patients with COPD can be translated into easily understandable data for the clinicians.

Methods:

Semi-structured interviews were conducted to explore clinicians’ needs when visualizing the results of a predictive algorithm. This formed a basis for creating a prototype of an updated user interface. The user interface was evaluated using usability tests through the “Think aloud” method.

Results:

The clinicians pointed out the need for visualization of exacerbation alerts and the development in patients’ data. Furthermore, they wanted the system to provide more information about what caused exacerbation alerts. Elements such as color and icons were described as particularly useful. The usability of the prototype was primarily assessed as easily understandable and advantageous in connection to the functions of the predictive algorithm.

Conclusion:

Predictive algorithm use in telemonitoring of COPD can be optimized by clearly visualizing the algorithm’s alerts, clarifying the reasons for algorithm output, and by providing a clear overview of the development in the patient’s data. This can contribute to clarity when the clinicians should act and why they should act on alerts from predictive algorithms.

The COVID-19 pandemic underlined that communities are key in sharing trusted, timely and relevant information especially during a health emergency where the overabundance of information makes it difficult to make decisions to protect one’s health. The WHO Hive project grew out of the desire to create a community-centered solution with the potential to change the way credible health information is shared, adapted, understood and used for health-related decision making before, during and after an epidemic or pandemic. The Hive online platform provides a safe space for knowledge-sharing, discussion, and collaboration, including access to timely scientific information through direct engagement with WHO subject matter experts, and the true innovation lies within the platform’s ability to leverage the power of communities to crowdsource solutions to community concerns and needs. The platform is equipped with a set of synchronous and asynchronous features and tools to encourage coworking and facilitate cross-sectorial collaboration. The Hive seeks to leverage the expert communities to share resources and knowledge for epidemic and pandemic preparedness and provide an environment that is able to respond to the challenges faced in a complex information ecosystem.

In this study, we automated the diagnostic procedure of autism spectrum disorder (ASD) with the help of anatomical alterations found in structural magnetic resonance imaging (sMRI) data of the ASD brain and machine learning tools. Initially, the sMRI data was preprocessed using the FreeSurfer toolbox. Further, the brain regions were segmented into 148 regions of interest using the Destrieux atlas. Features such as volume, thickness, surface area, and mean curvature were extracted for each brain region, and the morphological connectivity was computed using Pearson correlation. These morphological connections were fed to XGBoost for feature reduction and to build the diagnostic model. The results showed an average accuracy of 94.16% for the top 18 features. The frontal and limbic regions contributed more features to the classification model. Our proposed method is thus effective for the classification of ASD and can also be useful for the screening of other similar neurological disorders.

In multiple publications over 3 decades, most recently in The Book of Why, Judea Pearl has led what he regards as the ‘causal revolution’. His central contention is that, prior to it, no discipline had produced a rigorous ‘scientific’ way of making the causal inferences from observational data necessary for policy and decision making. The concentration on the statistical processing of data, outputting frequencies or probabilities, had proceeded without adequately acknowledging that this statistical processing is operating, not only on a particular set of data, but on a set of causal assumptions about that data, often unarticulated and unanalysed. He argues that the arrival of the directed acyclic graph (DAG), a ‘language of causation’ has enabled this fundamental weakness to be remedied. We outline the DAG approach to the extent necessary to make the key point, captured in this paper’s title regarding DAG’s potential contribution to improved decision or policy making.

Transformer models have been successfully applied to various natural language processing and machine translation tasks in recent years, e.g. automatic language understanding. With the advent of more efficient and reliable models (e.g. GPT-3), there is a growing potential for automating time-consuming tasks that could be of particular benefit in healthcare to improve clinical outcomes. This paper aims at summarizing potential use cases of transformer models for future healthcare applications. Precisely, we conducted a survey asking experts on their ideas and reflections for future use cases. We received 28 responses, analyzed using an adapted thematic analysis. Overall, 8 use case categories were identified including documentation and clinical coding, workflow and healthcare services, decision support, knowledge management, interaction support, patient education, health management, and public health monitoring. Future research should consider developing and testing the application of transformer models for such use cases.

The application of Natural Language Processing (NLP) to medical data has revolutionized different aspects of health care. The benefits obtained from the implementation of this technique spill over into several areas, including in the implementation of chatbots, which can provide medical assistance remotely. Every possible application of NLP depends on one first main step: the pre-processing of the corpus retrieved. The raw data must be prepared with the aim to be used efficiently for further analysis. Considerable progress has been made in this direction for the English language but for other languages, such as Italian, the state of the art is not equivalently advanced, especially for texts containing technical medical terms. The aim of this work is to identify and develop a preprocessing pipeline suitable for medical data written in Italian. The pipeline has been developed in Python environment, employing Enchant, ntlk modules and Hugging Face’s BERT and BART-based models. Then, it has been tested on real conversations typed between patients and physicians regarding medical questions. The algorithm has been developed within the MULTI-SITA project of the Italian Society of Anti-Infective Therapy (SITA), but shows a flexible structure that can adapt to a large variety of data.

Numerous classification systems have been developed over the years, systems which not only provide assistance to dermatologists, but also enable individuals, especially those living in areas with low medical access, to get a diagnosis. In this paper, a Machine Learning model, which performs a binary classification, and, which for the remainder of this paper will be abbreviated as ML model, is trained and tested, so as to evaluate its effectiveness in giving the right diagnosis, as well as to point out the limitations of the given method, which include, but are not limited to, the quality of smartphone images, and the lack of FAIR image datasets for model training. The results indicate that there are many measures to be taken and improvements to be made, if such a system were to become a reliable tool in real-life circumstances.

Hand and joint mobility recovery involve performing a set of exercises. Gestures are often used in the hand mobility recovery process. This paper discusses the selection and the use of 2 models of neural networks for the classification of data that describe Leap Motion gestures. The gestures are: the hand opening and closing gesture and the palm rotation gesture. The purpose is the optimal selection of the neural network model to be used in the classification of the data describing the recovery gestures. The models chosen for the classification of the two gestures were: Linear Discriminant Analysis (LDA) and K-neighbors Classifier (KNN). The accuracies achieved in the classification of the gestures for each model are: 0.91 – LDA and 0.98 – KNN.

Rheumatoid arthritis is a common disease which affects the joints of the wrist, fingers, feet and in the end the daily activities. Nowadays, gestures and virtual reality are used in many activities supporting recovery, games, learning as technology is present more and more in different fields. This paper presents results related to the grip movement detected by a Leap Motion device using binary classification and machine learning algorithms. We used 2 models to compare the results: Naïve Bayes and Random Forest Classifier. The metrics for comparison were: accuracy, precision, recall and f1-score. Also, we create a confusion matrix for a clear visualization of the results. We used 5000 data to train the algorithm and 1500 data to test. The accuracy and the precision were bigger than 97% in all the cases.

This paper proposes to create an RPA(robotic process automation) based software robot that can digitalize and extract data from handwritten medical forms. The RPA robot uses a taxonomy that is specific for the medical form and associates the extracted data with the taxonomy. This is accomplished using UiPath studio to create the robot, Google Cloud Vision OCR(optical character recognition) to create the DOM (digital object model) file and UiPath machine learning (ML) API to extract the data from the medical form. Due to the fact that the medical form is in a non-standard format a data extraction template had to be applied. After the extraction process the data can be saved into databases or into a spreadsheets.

This paper describes the latest development in the classification stage of our Speech Sound Disorder (SSD) Screening algorithm and presents the results achieved by using two classifier models: the Classification and Regression Tree (CART)-based model versus the Single Decision Hyperplane-based Linear Support Vector Machine (SVM) model. For every single speech sound in medial position, 10 features extracted from the audio samples along with an 11th feature representing the validation of the (mis)pronunciation by the Speech Language Pathologist (SLP) were fed into the 2 classifiers to compare and discuss their performance. The accuracy achieved by the two classifiers on a data test size of 30% of the analyzed samples was 98.2% for the Linear SVM classifier, and 100% for the Decision Tree classifier, which are optimal results that encourage our quest for a sound rationale.

Clinical texts are written with acronyms, abbreviations and medical jargon expressions to save time. This hinders full comprehension not just for medical experts but also laypeople. This paper attempts to disambiguate acronyms with their given context by comparing a web mining approach via the search engine BING and a conversational agent approach using ChatGPT with the aim to see, if these methods can supply a viable resolution for the input acronym. Both approaches are automated via application programming interfaces. Possible term candidates are extracted using natural language processing-oriented functionality. The conversational agent approach surpasses the baseline for web mining without plausibility thresholds in precision, recall and F1-measure, while scoring similarly only in precision for high threshold values.

In the context of global warming and increasing exposure to UV radiation, skin diseases are becoming more prevalent. Some of the most widespread skin conditions are solar lentigo and actinic keratosis. In this paper, we propose a technical approach related to the use of Azure Custom Vision services to classify these two conditions. The main advantage of using this service is the computational power offered by Azure. Additionally, generating a convolutional neural network model does not require a large dataset to achieve a good performance. For training the model, we used a dataset of 600 images from the ISIC database. The limitations of these approaches are imposed by the manual image labeling part that needs to be performed. As a result, we provide a trained model on a series of images that can be used for classifying images related to these two conditions. The performance of our neural network on the pre-trained images is 94%.

Research in the field of maternal-fetal medicine brings a new approach, by involving several fields: genetics, informatics, teratology, imaging, obstetric diagnosis, maternal-fetal physiology, endocrinology, and aims to determine the relationships that appear between the maternal medical pathology and the fetal one. In this article, we present an application for monitoring and calculating risk in Trisomy-21 for pregnant women. To calculate the risk, we used 2 methods, one mathematical and one using neural networks to investigate which one offers higher precision. Following the experimental results, due to the use of several variables that increase the risk for Trisomy-21, the conclusion is that the method using neural networks is better, having an accuracy of 95%.

The Moroccan healthcare system is facing several challenges in ensuring equitable access to quality services and reducing or at least controlling their rising cost. Telemedicine can address these two needs by optimizing the use of existing human and material resources through telecommunications. Today, the gradual increase in the population’s healthcare needs poses a major challenge to the Moroccan healthcare system, given the shortage of personnel in healthcare facilities and the persistent difficulties in accessing certain regions. In this regard, Morocco has established a regulatory framework defining the rules for the practice of telemedicine. Several initiatives have been launched, particularly in the public sector, aiming to cover 80% of medical deserts in Morocco by 2025.

Overcrowding in EDs has been viewed globally as a chronic health challenge. It is directly related to the increased use of EDs for non-urgent issues, leading to increased complications, long waiting times, a higher death rate, or delayed intervention of those more acutely ill. This study aims to develop Machine Learning models to differentiate immediate medical needs from unnecessary ED visits. A Decision Tree, Random Forest, AdaBoost, and XGBoost models were built and evaluated on real-life data. XGBoost achieved the best accuracy and F1-score.

In this paper, we describe Neonatal Resuscitation Training Simulator (NRTS), an Android mobile app designed to support medical experts to input, transmit and record data during a High-Fidelity Simulation course for neonatal resuscitation. This mobile app allows one to automatically send all the recorded data from the Neonatal Intensive Care Unit (NICU) of Casale Monferrato Children’s Hospital, (Italy) to a server in the cloud managed by the University of Piemonte Orientale (Italy). The medical instructor can then view statistics on simulation exercises, that may be used during the debriefing phase for the evaluation of multidisciplinary teams involved in the simulation scenarios.

Translating the proposed European Health Data Space (EHDS) regulations and requirements into reality is a challenging task. In this work, we provide a roadmap for aligning the EHDS requirement into the cardiovascular (CV) digital health domain in Austria. To achieve that, we first examined the current eHealth infrastructure and initiatives in Austria. Then, we created a CV-connected health model and addressed the challenges facing cardiac telerehabilitation in Austria. Finally, we mapped the European CV strategies to the Austrian context for EHDS implementation. Accordingly, we were able to provide an Enterprise Architecture (EA) framework for aligning CV digital health with the Austrian EHDS context. The created EA model can be also used as a guiding framework for aligning other medical domains in Austria with EHDS.