Ebook: Public Health and Informatics

This volume contains the accepted full papers, short communication papers, and poster papers from the 31st Medical Informatics in Europe Conference (MIE2021) held virtually from 29–31 May 2021. MIE 2021 was originally due to be held in Athens, Greece, according to the EFMI Council decision, but due to the continuing gravity of the pandemic situation across the globe at the time of the Call for Papers and Proposals, it was decided, for the safety of participants, that the conference would be held as a virtual event. The Conference has been organized by the MCO Congress, and the Scientific Programme Committee was chaired by Professor John Mantas, Fellow of EFMI, and President of the Greek Biomedical and Health Informatics Association (GBHI).

For some years, both eHealth applications and digitalization have been considered fundamental paradigm changers in the new era of health informatics and public health. Massive amounts of data from molecular biology and about the environment, behaviour and lifestyle, exposition factors and personal records, coupled with the rise of unprecedented processing power in health information systems and artificial intelligence, as well as health analytics methods and tools, have empowered the medicine, nursing and healthcare sciences facing the challenges posed by public health. The current pandemic situation has emphasized the importance of health informatics for the scientific process of evidence-based reasoning and decision making at all levels of healthcare.

In this 31st annual experience, there will be special tracks with European projects, and discussions on the building of global frameworks to improve the data usability necessary to support life science research across borders, systems, and languages. Ethical and legal experts, contributing to specific tracks devoted to encryption, Blockchain and privacy-conscious data sharing, will investigate and propose practical ways to support innovations which will help to alleviate the burden of diseases such as COVID-19 and promote surveillance networks.

This volume incorporates 261 papers presented during the conference, and these proceedings are published as an e-book, with open access for ease of use and browsing without losing any of the advantages of indexing and citation, in the biggest Scientific Literature Databases, such as Medline and Scopus as part of the series Studies in Health Technology and Informatics (HTI) from IOS Press.

The Editors would like to thank the Members of the Scientific Programme Committee, namely John Mantas (chair), Lăcrămioara Stoicu-Tivadar (co-chair), Catherine Chronaki, Arie Hasman, Patrick Weber, Parisis Gallos, and Mihaela Crişan-Vida, the EFMI Executive Officer Rebecca Randell, and Christian Lovis for supporting the continuous education accreditation. But mostly, we would like to thank the reviewers who have performed a very professional service, enabling a high-quality publishing achievement for a successful scientific event.

Athens, 24.04.2021

The Editors,

John Mantas, Lăcrămioara Stoicu-Tivadar, Catherine Chronaki, Arie Hasman, Patrick Weber, Parisis Gallos, Mihaela Crişan-Vida, Emmanouil Zoulias, Oana Sorina Chirila

Reproducible information is important in science, medicine and other professional fields. Repeating the same experiment with measurement should yield the same information as the result. This original information should also be transported digitally in reproducible form, as a globally well-defined sequence of numbers. The article explains that “Domain Vectors” (DVs) with the structure “UL plus sequence of numbers” are well suited for this purpose. “UL” is an efficient link to the online definition of the sequence of numbers. DVs are globally comparable and searchable and have other important advantages. It is concluded that DVs can fill an important gap in the digital representation of information.

The aim of this study is to build an evaluation framework for the user-centric testing of the Data Curation Tool. The tool was developed in the scope of the FAIR4Health project to make health data FAIR by transforming them from legacy formats into a Common Data Model based on HL7 FHIR. The end user evaluation framework was built by following a methodology inspired from the Delphi method. We applied a series of questionnaires to a group of experts not only in different roles and skills, but also from various parts of Europe. Overall, 26 questions were formulated for 16 participants. The results showed that the users are satisfied with the capabilities and performance of the tool. The feedbacks were considered as recommendations for technical improvement and fed back into the software development cycle of the Data Curation Tool.

The early detection of Heart Disease (HD) and the prediction of Heart Failure (HF) via telemonitoring and can contribute to the reduction of patients’ mortality and morbidity as well as to the reduction of respective treatment costs. In this study we propose a novel classification model based on fuzzy logic applied in the context of HD detection and HF prediction. The proposed model considers that data can be represented by fuzzy phrases constructed from fuzzy words, which are fuzzy sets derived from data. Advantages of this approach include the robustness of data classification, as well as an intuitive way for feature selection. The accuracy of the proposed model is investigated on real home telemonitoring data and a publicly available dataset from UCI.

Metadata management is an essential condition to follow the FAIR principles. Therefore, metadata management was one asset of an accompanying project within a funding scheme for registries in health services research. The metadata of the funded projects were acquired, combined in a database compatible with the metamodel of ISO/IEC 11179 “Information technology – Metadata registries” third edition (ISO/IEC 11179-3), and analyzed in order to support the development and the operation of the registries. In the second phase of the funding scheme, six registries delivered a complete update of their metadata. The mean number of data elements increased from 245.7 to 473.5 and the mean number of values from 569.5 to 1,306.0. The conceptual core of the database had to be extended by one third to cover the new elements. The reason for this increase remained unclear. Constraints from the grant might be causal, a deviation from an evidence-based development process as well. It is questionable, whether the revealed quality of the metadata is sufficient to fulfill the FAIR principles. The extension of the metamodel of ISO/IEC 11179-3 is in agreement with the literature. However, further research is needed to find workable solutions for metadata management.

The integration of surgical knowledge into virtual planning systems plays a key role in computer-assisted surgery. The knowledge is often implicitly contained in the implemented algorithms. However, a strict separation would be desirable for reasons of maintainability, reusability and readability. Along with the Department of Oral and Maxillofacial Surgery at Heidelberg University Hospital, we are working on the development of a virtual planning system for mandibular reconstruction. In this work we describe a process for the structured acquisition and representation of surgical knowledge for mandibular reconstruction. Based on the acquired knowledge, an RDF(S) ontology was created. The ontology is connected to the virtual planning system via a SPARQL interface. The described process of knowledge acquisition can be transferred to other surgical use cases. Furthermore, the developed ontology is characterised by a reusable and easily expandable data model.

This work aims to describe how EHRs have been used to meet the needs of healthcare providers and researchers in a 1,300-beds tertiary Hospital during COVID-19 pandemic. For this purpose, essential clinical concepts were identified and standardized with LOINC and SNOMED CT. After that, these concepts were implemented in EHR systems and based on them, data tools, such as clinical alerts, dynamic patient lists and a clinical follow-up dashboard, were developed for healthcare support. In addition, these data were incorporated into standardized repositories and COVID-19 databases to improve clinical research on this new disease. In conclusion, standardized EHRs allowed implementation of useful multi- purpose data resources in a major Hospital in the course of the pandemic.

The Fast Healthcare Interoperability Resources (FHIR) contain multiple data-exchange standards that aim at optimizing healthcare information exchange. One of them, the FHIR AdverseEvent, may support post-market safety surveillance. We examined its readiness using the Food and Drug Administration’s (FDA) Adverse Event Reporting System (FAERS). Our methodology focused on mapping the public FAERS data fields to the FHIR AdverseEvent Resource elements and developing a software tool to automate this process. We mapped directly nine and indirectly two of the twenty-six FAERS elements, while six elements were assigned default values. This exploration further revealed opportunities for adding extra elements to the FHIR standard, based on critical FAERS pieces of information reviewed at the FDA. The existing version of the FHIR AdverseEvent Resource may standardize some of the FAERS information but has to be modified and extended to maximize its value in post-market safety surveillance.

Data integration is an increasing need in medical informatics projects like the EU Precise4Q project, in which multidisciplinary semantically and syntactically heterogeneous data across several institutions needs to be integrated. Besides, data sharing agreements often allow a virtual data integration only, because data cannot leave the source repository. We propose a data harmonization infrastructure in which data is virtually integrated by sharing a semantically rich common data representation that allows their homogeneous querying. This common data model integrates content from well-known biomedical ontologies like SNOMED CT by using the BTL2 upper level ontology, and is imported into a graph database. We successfully integrated three datasets and made some test queries showing the feasibility of the approach.

Hospital-acquired infections, particularly in ICU, are becoming more frequent in recent years, with the most serious of them being Gram-negative bacterial infections. Among them, Acinetobacter baumannii, Klebsiella pneumoniae, and Pseudomonas aeruginosa are considered the most resistant bacteria encountered in ICU and other wards. Given the fact that about 24 hours are usually required to perform common antibiotic resistance tests after the bacteria identification, the use of machine learning techniques could be an additional decision support tool in selecting empirical antibiotic treatment based on the sample type, bacteria, and patient’s basic characteristics. In this article, five machine learning (ML) models were evaluated to predict antimicrobial resistance of Acinetobacter baumannii, Klebsiella pneumoniae, and Pseudomonas aeruginosa. We suggest implementing ML techniques to forecast antibiotic resistance using data from the clinical microbiology laboratory, available in the Laboratory Information System (LIS).

Chatbots potentially address deficits in availability of the traditional health workforce and could help to stem concerning rates of youth mental health issues including high suicide rates. While chatbots have shown some positive results in helping people cope with mental health issues, there are yet deep concerns regarding such chatbots in terms of their ability to identify emergency situations and act accordingly. Risk of suicide/self-harm is one such concern which we have addressed in this project. A chatbot decides its response based on the text input from the user and must correctly recognize the significance of a given input. We have designed a self-harm classifier which could use the user’s response to the chatbot and predict whether the response indicates intent for self-harm. With the difficulty to access confidential counselling data, we looked for alternate data sources and found Twitter and Reddit to provide data similar to what we would expect to get from a chatbot user. We trained a sentiment analysis classifier on Twitter data and a self-harm classifier on the Reddit data. We combined the results of the two models to improve the model performance. We got the best results from a LSTM-RNN classifier using BERT encoding. The best model accuracy achieved was 92.13%. We tested the model on new data from Reddit and got an impressive result with an accuracy of 97%. Such a model is promising for future embedding in mental health chatbots to improve their safety through accurate detection of self-harm talk by users.

Access to hospitals has been dramatically restricted during the COVID 19 pandemic. Indeed, due to the high risk of contamination by patients and by visitors, only essential visits and medical appointments have been authorized. Restricting hospital access to authorized visitors was an important logistic challenge. To deal with this challenge, our institution developed the ExpectingU app to facilitate patient authorization for medical appointments and for visitors to enter the hospital. This article analyzes different trends regarding medical appointments, visitors’ invitations, support staff hired and COVID hospitalizations to demonstrate how the ExpectingU system has helped the hospital to maintain accessibility to the hospital. Results shows that our system has allowed us to maintain the hospital open for medical appointments and visits without creating bottlenecks.

Clinical trials are carried out to prove the safety and effectiveness of new interventions and therapies. As diseases and their causes continue to become more specific, so do inclusion and exclusion criteria for trials. Patient recruitment has always been a challenge, but with medical progress, it becomes increasingly difficult to achieve the necessary number of cases. In Germany, the Medical Informatics Initiative is planning to use the central application and registration office to conduct feasibility analyses at an early stage and thus to identify suitable project partners. This approach aims to technically adapt/integrate the envisioned infrastructure in such a way that it can be used for trial case number estimation for the planning of multicenter clinical trials. We have developed a fully automated solution called APERITIF that can identify the number of eligible patients based on free-text eligibility criteria, taking into account the MII core data set and based on the FHIR standard. The evaluation showed a precision of 62.64 % for inclusion criteria and a precision of 66.45 % for exclusion criteria.

The automation of medical documentation is a highly desirable process, especially as it could avert significant temporal and monetary expenses in healthcare. With the help of complex modelling and high computational capability, Automatic Speech Recognition (ASR) and deep learning have made several promising attempts to this end. However, a factor that significantly determines the efficiency of these systems is the volume of speech that is processed in each medical examination. In the course of this study, we found that over half of the speech, recorded during follow-up examinations of patients treated with Intra-Vitreal Injections, was not relevant for medical documentation. In this paper, we evaluate the application of Convolutional and Long Short-Term Memory (LSTM) neural networks for the development of a speech classification module aimed at identifying speech relevant for medical report generation. In this regard, various topology parameters are tested and the effect of the model performance on different speaker attributes is analyzed. The results indicate that Convolutional Neural Networks (CNNs) are more successful than LSTM networks, and achieve a validation accuracy of 92.41%. Furthermore, on evaluation of the robustness of the model to gender, accent and unknown speakers, the neural network generalized satisfactorily.

The current movement in Medical Informatics towards comprehensive Electronic Health Records (EHRs) has enabled a wide range of secondary use cases for this data. However, due to a number of well-justified concerns and barriers, especially with regards to information privacy, access to real medical records by researchers is often not possible, and indeed not always required. An appealing alternative to the use of real patient data is the employment of a generator for realistic, yet synthetic, EHRs. However, we have identified a number of shortcomings in prior works, especially with regards to the adaptability of the projects to the requirements of the German healthcare system. Based on three case studies, we define a non-exhaustive list of requirements for an ideal generator project that can be used in a wide range of localities and settings, to address and enable future work in this regard.

Against the background of increasing numbers of indications for Cochlea implants (CIs), there is an increasing need for a CI outcome prediction tool to assist the process of deciding on the best possible treatment solution for each individual patient prior to intervention. The hearing outcome depends on several features in cochlear structure, the influence of which is not entirely known as yet. In preparation for surgical planning a preoperative CT scan is recorded. The overall goal is the feature extraction and prediction of the hearing outcome only based on this conventional CT data. Therefore, the aim of our research work for this paper is the preprocessing of the conventional CT data and a following segmentation of the human cochlea. The great challenge is the very small size of the cochlea in combination with a fairly bad resolution. For a better distinction between cochlea and surrounding tissue, the data has to be rotated in a way the typical cochlea shape is observable. Afterwards, a segmentation can be performed which enables a feature detection. We can show the effectiveness of our method compared to results in literature which were based on CT data with a much higher resolution. A further study with a much larger amount of data is planned.

During the current COVID-19 pandemic, the rapid availability of profound information is crucial in order to derive information about diagnosis, disease trajectory, treatment or to adapt the rules of conduct in public. The increased importance of preprints for COVID-19 research initiated the design of the preprint search engine preVIEW. Conceptually, it is a lightweight semantic search engine focusing on easy inclusion of specialized COVID-19 textual collections and provides a user friendly web interface for semantic information retrieval. In order to support semantic search functionality, we integrated a text mining workflow for indexing with relevant terminologies. Currently, diseases, human genes and SARS-CoV-2 proteins are annotated, and more will be added in future. The system integrates collections from several different preprint servers that are used in the biomedical domain to publish non-peer-reviewed work, thereby enabling one central access point for the users. In addition, our service offers facet searching, export functionality and an API access. COVID-19 preVIEW is publicly available at https://preview.zbmed.de.

Automated text classification is a natural language processing (NLP) technology that could significantly facilitate scientific literature selection. A specific topical dataset of 630 article abstracts was obtained from the PubMed database. We proposed 27 parametrized options of PubMedBERT model and 4 ensemble models to solve a binary classification task on that dataset. Three hundred tests with resamples were performed in each classification approach. The best PubMedBERT model demonstrated F1-score = 0.857 while the best ensemble model reached F1-score = 0.853. We concluded that the short scientific texts classification quality might be improved using the latest state-of-art approaches.

Studies investigating the suitability of SNOMED CT in COVID-19 datasets are still scarce. The purpose of this study was to evaluate the suitability of SNOMED CT for structured searches of COVID-19 studies, using the German Corona Consensus Dataset (GECCO) as example. Suitability of the international standard SNOMED CT was measured with the scoring system ISO/TS 21564, and intercoder reliability of two independent mapping specialists was evaluated. The resulting analysis showed that the majority of data items had either a complete or partial equivalent in SNOMED CT (complete equivalent: 141 items; partial equivalent: 63 items; no equivalent: 1 item). Intercoder reliability was moderate, possibly due to non-establishment of mapping rules and high percentage (74%) of different but similar concepts among the 86 non-equal chosen concepts. The study shows that SNOMED CT can be utilized for COVID-19 cohort browsing. However, further studies investigating mapping rules and further international terminologies are necessary.

One of the important questions in the research on neural coding is how the preceding axonal activity affects the signal propagation speed of the following one. We present an approach to solving this problem by introducing a multi-level spike count for activity quantification and fitting a family of linear regression models to the data. The best-achieved score is R²=0.89 and the comparison of different models indicates the importance of long and very short nerve fiber memory. Further studies are required to understand the complex axonal mechanisms responsible for the discovered phenomena.

In this paper efforts have been made to record the actual, real cost of health care services in a Neonatal Intensive Care Unit (N.I.C.U.) of a public hospital. It is well known that, in recent years, the hospitals have been reimbursed with the system of Diagnosis-Related Groups (D.R.G.’s). The purpose of this study is to determine whether the costs according with D.R.G.’s correspond to the actual-real cost, as this is recorded in the N.I.C.U. This cost is called direct cost. Here is a case study of a premature neonate in the intensive care unit (N.I.C.U.). From the outset, the age of pregnancy, the birth weight, the duration of hospitalization in N.I.C.U. and the needs of the newborn in oxygen, medication, as well as nutrition are defined which are very important in shaping the cost. Then, the cost is calculated according to the D.R.G.’s system. By setting three basic diagnoses (I.C.D.-10), we find the D.R.G. which better describes the case, as well as the associated costs. Then, we calculate the direct cost and list all the consumables, exams, staff costs, overheads. Comparing the two results we find that the cost of D.R.G. does not meet the direct cost of hospitalization. There is a significant deviation from the actual real cost, which proves the under-costing of the health services. The D.R.G.’s system leads hospitals to increase their financial deficits and provide degraded quality health services. It is necessary to readjust the D.R.G.’s according to the reality and the redefinition of the hospital’s reimbursement system to meet the direct – real cost of the health services offered.

Acute kidney injury (AKI) is a common and potentially life-threatening condition, which often occurs in the intensive care unit. We propose a machine learning model based on recurrent neural networks to continuously predict AKI. We internally validated its predictive performance, both in terms of discrimination and calibration, and assessed its interpretability. Our model achieved good discrimination (AUC 0.80-0.94). Such a continuous model can support clinicians to promptly recognize and treat AKI patients and may improve their outcomes.

International Organizations are seriously concerned about the fake news phenomenon. UNESCO has defined the term of misinformation/disinformation, which are the two faces of fake news. European Commission has conducted a survey about “Fake News” through EU citizens to estimate the awareness and people behaviour related to the appearance of fake news and disinformation on electronic. The findings are quite worrying, since about 40% come across fake news daily and 85% evaluate fake news as a problem. The aim of this work is to introduce an Artificial Intelligence approach, the Decision Trees algorithm to identify fake news on the COVID-19.

The FAIR Principles are a set of recommendations that aim to underpin knowledge discovery and integration by making the research outcomes Findable, Accessible, Interoperable and Reusable. These guidelines encourage the accurate recording and exchange of data, coupled with contextual information about their creation, expressed in domain-specific standards and machine-readable formats. This paper analyses the potential support to FAIRness of the openEHR specifications and reference implementation, by theoretically assessing their compliance with each of the 15 FAIR principles. Our study highlights how the openEHR approach, thanks to its computable semantics-oriented design, is inherently FAIR-enabling and is a promising implementation strategy for creating FAIR-compliant Clinical Data Repositories (CDRs).