Ebook: dHealth 2022

This book presents the proceedings of the 16th annual conference on Health Informatics meets Digital Health (dHealth 2022), held in Vienna, Austria, on May 24–25, 2022. In keeping with its interdisciplinary mission, the conference series provides a platform for researchers and decision makers, health professionals, healthcare providers, governmental and industry representatives to discuss innovative digital health solutions to improve the quality and efficiency of healthcare using digital technologies.

Students from various flavours of health informatics courses are also a very important target group for the conference, both as attendees to complement their education in the respective tertiary teaching schools and as presenters in the student competition. We consider it crucial to help to engage the next generation to become digital-health developers, architects, facilitators, pioneers, innovators, managers, entrepreneurs, and evangelists.

The 42 papers included here provide an insight into the state-of-the-art of different aspects of dHealth, including the design and evaluation of user interfaces, patient-centred solutions, electronic health/medical/patient records, machine learning in healthcare and biomedical data analytics. Therefore, the book offers the reader an interdisciplinary view of the state-of-the-art and of ongoing research activities in digital health.

In 2019, we decided to change the name of the conference from eHealth to dHealth to emphasise that the healthcare of the future will become ever more data-driven, decision-supporting, deep, or simply more digital. And indeed, since then new “digital” terms have emerged and became an integral part of the way digital health professionals speak and write. One example is “digital therapeutics”.

This is reasonable, since it supports the evolution of the field, its differentiation and specialisation. It is also a sign of maturation, and this trend will probably continue for a few more years, with a constant stream of new terms spinning off from the “digital” stem, until a new lead term emerges. Today, no one knows when and what that term will be, but we will do our best to capture and cultivate the semantic future of our field in the context of this and subsequent dHealth conferences. So please stay tuned!

Graz, Hall in Tyrol, Vienna, May 2022

Günter Schreier (AIT)

Bernhard Pfeifer (UMIT)

Martin Baumgartner (AIT)

Dieter Hayn (AIT)

Patient summaries grant healthcare providers a concise overview of a patient’s status. This paper showcases to which degree International Patient Summaries (IPS) represented in HL7 FHIR format can be generated using data from the nationwide Austrian Electronic Health Record system ELGA. A solution is presented which enables the automated software-assembled generation of an IPS using the FHIR Mapping Language. The generated document successfully validates against the IPS profiles. Our results show that all required IPS sections can be supplied from ELGA data.

Background:

Patient portals may support patient engagement, yet they may differ largely in their characteristics.

Objectives:

To compare the Austrian patient portal with 10 portals from six other countries using the TOPCOP Taxonomy.

Methods:

We described the portals using openly available information.

Results:

The Austrian patient portal shows basic functionality but lacks further functions that other portals partly offer.

Conclusion:

Comparing portals using TOPCOP is possible and shows functions to improve usefulness of portals.

The diagnosis of rare diseases is often challenging for physicians, but can be supported by Clinical Decision Support Systems. The MIRACUM consortia, which includes ten university hospitals in Germany, develops a Clinical Decision Support System to support the diagnosis of patients with rare diseases. The users are involved in different phases using a user-centred design process. This publication has the objective to summarize the results of all studies performed in context of the requirements elicitation and to derive concrete requirements for the development of the system. Several studies were performed for requirements elicitation: a cross-sectional survey, expert interviews and a focus group. Participants were experts in rare diseases of the MIRACUM locations. 32 requirements were derived and implemented in a prototype. The prototype allows similarity analyses as a decision support functionality by comparing patients without a diagnosis to patients with a rare disease. In the final evaluation, the prototype was rated with a good usability. Since the system is limited in its functionality, further work and improvements are necessary to make it ready for clinical usage.

The academic research environment is characterized by self-developed, innovative, customized solutions, which are often free to use for third parties with open-source code and open licenses. On the other hand, they are maintained only to a very limited extent after the end of project funding. The ToolPool Gesundheitsforschung addresses the problem of finding ready to use solutions by building a registry of proven and supported tools, services, concepts and consulting offers. The goal is to provide an up-to-date selection of “relevant” solutions for a given domain that are immediately usable and that are actually used by third parties, rather than aiming at a complete list of all solutions which belong to that domain. Proof of relevance and usage must be provided, for example, by concrete application scenarios, experience reports by uninvolved third parties, references in publications or workshops held. Quality assurance is carried out for new entries by an agreed list of admission criteria, for existing entries at least once a year by a special task force. Currently, 79 solutions are represented, this number is to be significantly expanded by involving of new editors from current national funding initiatives in Germany.

Background:

Need for additional methods for interdisciplinary involvement in the development of digital solutions in the health and care education.

Objectives:

Development of an interdisciplinary and age-inclusive hackathon approach to support and integrate young care students in the care digitization process.

Methods:

Iterative, interdisciplinary planning, expert-based validation, themed, age-inclusive hackathon implementation.

Results:

Concept for hackathon, concept evaluation.

Conclusion:

The hackathon is seen as a valuable addition in the health and care education and will be held as a proof-of-concept in November 2022.

Background:

Postural imbalance can be adopted for the early detection of age-related diseases or monitoring the course of the disease treatment; especially in monitoring, frequent balance measurement is crucial. This is mainly done through regular in-person examinations by a physician currently. Feedback in between examinations is often missing.

Objectives:

This paper proposes mBalance, a mobile application that uses the Romberg test to detect postural imbalance. mBalance provides a camera-based, low-cost approach to measure imbalance frequently at home using mobile devices.

Methods:

Imbalance detection accuracy and usability was evaluated in two separate studies with 31 and 30 participants, respectively.

Results:

mBalance correctly detected imbalance with a sensitivity of 80% and a specificity of 87%. The study found good usability with no significant problems.

Conclusion:

Overall, this study solves the problem of postural imbalance detection by digitizing a validated balance test into an easy-to-use mobile application.

Background:

Personal contact between radiologists and their patients is scarce due to time constraints and logistical reasons which impacts on patient knowledgeability and satisfaction, but also on examination and diagnostic quality.

Objective:

We illuminate medical history interviews from a radiologist’s perspective and discuss its impact on the diagnostic quality. Based on these insights, we develop a digital medical interview assistant (DMIA) for radiology that is intended to collect information helping in improving radiological diagnostics.

Methods:

Conditions, issues, problems in the radiological examination process are assessed to collect requirements and to specify questions for a digital medical history interview.

Results:

A DMIA with conversational user interface is developed using the scripting language RiveScript. It is accessible through a social media messenger (Telegram messenger). An initial assessment of usability demonstrates a good usability.

Conclusion:

To overcome the information gap in radiology, a DMIA can simulate an assessment interview. It is still necessary to remove existing barriers in interaction with the DMIA for example by facilitating data entry options.

Background:

Many patients suffering from post-acute COVID-19 syndrome must deal with fatigue. They need physical and psychological support, strengthening, and adaption of their individual lifestyle. The use of apps can foster fatigue management.

Objectives:

The aim of this study was to analyse experiences of patients with fatigue caused by post-acute COVID-19 syndrome after using Untire for more than two weeks.

Methods:

A qualitative research design was chosen to identify user-experience after using the app. Seven semi-structured interviews were conducted and qualitative content analysis according to Mayring was applied.

Results:

The Untire App was generally perceived as being easy to use. Patients judged the Untire App as supportive in most of the features, especially in energy measurement and relaxation exercises.

Conclusion:

Thus, though developed for cancer patients, this mobile health (mHealth) application is with some restrictions also suitable for patients with post-acute COVID-19 syndrome.

Background:

Effective communication is a key factor in healthcare, essential for improving process efficiency and quality of care. This is particularly true in new services, e.g., the 3D printing service inside the hospital.

Objectives:

A web platform, called 3DSCT, has been developed to act as an interface between the three categories of operators involved in 3D printing: physicians, radiologists and engineers.

Methods:

The 3DSCT platform has been designed using Microsoft Visual Studio Code, enclosing .js scripts and HTML pages with the relative CSS formats.

Results:

When applied to a real 3D printing service, the 3DSCT platform provided an effective solution that streamlined the process of designing and manufacturing 3D-printed artifacts, from physician’s request through development to printing.

Conclusion:

By incorporating the platform into the hospital management system, it will be possible to reduce the overall lead time and decrease the waste of time for the operators involved in 3D printing inside the hospital.

Decision-making based on so-called medical guidelines supported by semantic AI solutions is an essential and significant task for medical personnel in both a pre-clinical setting and an inner-clinical environment. Semantic representations of medical guidelines and Fast Healthcare Interoperability Resources (FHIR) using Semantic Web technologies, i.e., Resource Description Framework (RDF), rules (RuleML and Prova), and Shape Constraint Language (SHACL), provide a semantic knowledge base for the decision-making process and ease technical implementation and automation tasks. Current medical decision support systems lack Semantic Web integration using FHIR-RDF representations as a data source. In this paper, we implement a particular medical guideline using two different approaches: Prova [8] and SHACL [13]. We generate a series of raw FHIR-data for a selected guideline, the ABCDE approach, and compare the implemented two programs’ (Prova and SHACL) results. Both approaches deliver the same results in terms of content. Both may be used within a distributed medical environment depending on the need of organizations.

SNOMED CT has an enormous number of clinical concepts and mapping to SNOMED CT is considered as the foundation to achieve semantic interoperability in healthcare. Manual mapping is time-consuming and error-prone thus making this crucial step challenging. Terminology Servers provide an interface, which can be used to automate the process of retrieving data. Snowstorm is a terminology server developed by SNOMED International. In this work, the feasibility of using Snowstorm to automate the data retrieval and mapping has been discussed.

Background:

From 2022, the “Outpatient Clinic Report” and the “Telehealth Note” will complement the existing e-Reports in the Austrian Electronic Health Record system ELGA.

Objectives:

The specification of two harmonized implementation guides with standardized structure for all types of outpatient clinics in hospitals on the one hand and for telemonitoring treatments on the other hand.

Methods:

With the participation of expert groups, the contents were harmonized, and a data model was created. Template specifications were modelled in ART-DECOR and approved in the course of an HL7 Austria ballot.

Results:

Two sets of freely selectable building blocks and machine-readable content were created.

Conclusion:

The “Outpatient Clinic Report” and the “Telehealth Note” are currently being implemented. The use of these documents will be evaluated as well as if additional machine-readable content is needed.

Semantic interoperability is the centerpiece of successful communication between information systems within the health care system. As such, a terminology server provides information systems with the concepts needed for coding information accordingly. As the current Austrian terminology server does not meet all requirements anymore, a new terminology server is necessary. Consequently, after an unsuccessful call for bids, the Semantic Competence Center (SCC) of the ELGA GmbH set up a technologically more advanced solution for a new terminology server. The result is a Git based architecture in combination with the HL7^® FHIR^® IG publisher complemented by a FHIR^® server and some in-house developments. The Git repository is hosted on GitLab.com, hence, allowing the use of GitLab’s CI/CD functionalities for publishing the contents. Since the 1 January 2022 the new terminology server “TerminoloGit” is operational and provides 171 code lists and 141 value sets. Using open-source components only and making all parts of the system publicly available results in a cost-effective solution which may also be improved by the open-source community.

In many developing countries like India, there is a widespread lack of general awareness about the importance of good oral health, which causes dental patients to neglect their oral hygiene, thus precipitating many long-term ailments. We developed an application that promotes the significance of regular dental checkups and oral health care by explaining to patients how these are intrinsic to overall health. Our application, in essence, extracts relevant health information from published scientific studies according to a patient’s medical history and shares it with the patient at the discretion of the supervising dentist, thereby empowering patients to make more informed decisions. We present a detailed overview of our semi- autonomous machine learning-based solution, along with the complex challenges involved in the design, development, and real-world deployment of our application. Finally, we conducted a randomized parallel-group study in India with 224 dental patients over two years to assess the utility of our proposed solution. Results show our application improved the patient recall rate from 21.1% to 37.8% (p-value = 0.024).

Background:

Various machine learning (ML) models have been developed for the prediction of clinical outcomes, but there is missing evidence on their performance in clinical routine and external validation.

Objectives:

Our aim was to deploy and prospectively evaluate an already developed delirium prediction software in clinical routine of an external hospital.

Methods:

We compared updated ML models of the software and models re-trained with the external hospital’s data. The best models were deployed in clinical routine for one month, and risk predictions for all admitted patients were compared to the risk ratings of a senior physician. After using the software, clinicians completed a questionnaire assessing technology acceptance.

Results:

Re-trained models achieved a high discriminative performance (AUROC > 0.92). Compared to clinical risk ratings, the software achieved a sensitivity of 100.0% and a specificity of 90.6%. Usefulness, ease of use and output quality were rated positively by the users.

Conclusion:

A ML based delirium prediction software achieved a high discriminative performance and high technology acceptance at an external hospital using re-trained ML models.

Background:

One of the most important tasks within wound management and the basis of successful wound care is a clear and accurate nursing documentation which can be created in various ways. However, this documentation is only partially standardized mainly due to the interdisciplinarity of this task.

Objectives:

The goal is to overcome interdisciplinary boundaries in home care between health professionals, and to increase the patient’s motivation to actively participate in wound healing by providing consistent structure and guidance. In this paper, the development of a mobile wound documentation application is described.

Methods:

Comprehensive literature reviews were conducted on state-of-the-art wound documentation as well as on machine learning frameworks, databases and mobile app development technologies.

Results:

As a result, a mobile wound application with Flutter, SQLite and Tensorflow Lite was developed. During the development process interoperability was considered to enable future extensibility.

Conclusion:

Further development in respect of backend and API is planned, involving user field tests.

Background:

The European health care industry faces massive changes which impose new challenges on its stakeholders.

Objective:

In this paper, we present the results of a market-based analysis for upcoming changes in the European health care industry and what this specifically means for issues corresponding to data.

Method:

Scenarios are a common tool to explain and analyze future changes in business environments. This method was used in a series of workshops together with an interdisciplinary group of experts.

Results:

Ten individual scenarios represent potential futures with distinctive subsets of data landscapes. Their assessment shows that the expected future of health data is still rather unclear, while desired changes are quite distinct.

Conclusion:

The Health Data Scenarios offer a comprehensive framework for analyzing future data-driven developments in the health care industry.

Background:

In recent years, there has been a rising interest in the application of deep neural networks (DNN) for the delineation of the electrocardiogram (ECG).

Objectives:

A variety of DNN architectures has been investigated in a 5-fold cross-validation approach.

Results:

The best performing network achieved 100% sensitivity and >97% positive predictive value for all ECG waves.

Conclusion:

Our DNN could achieve similar classification performance as other DNN approaches described in the literature at a reduced computational cost.

People with physical limitations face significant challenges when using existing toilets. User requirements work shows the wide range of user needs and confirms the high demand for innovative toilets, enabling people to leave home more often and participate more in societal life. The Toilet For Me too (T4ME2) project aims to implement and test a new ICT-based toilet system capable of physically supporting users, allowing autonomous and safe use outside the home.

Background:

The number of software products in the field of health and medicine increases excessively. Self-tracking, fitness, dose calculation, and analysis of physiological data – apps are popular and commonly used. For young entrepreneurs, it is difficult to recognize and understand the distinction between software as medical product and health software products. A product-related orientation guide was developed to help start-ups to understand the difference and to find the right strategy for placing their product on the market.

Objectives:

An initial evaluation of this orientation guide to improve the comprehensibility, the simplicity, and to validate the benefit.

Methods:

The evaluation was performed using a questionnaire. In total, 15 employees from start-ups or those in foundation phase and in regulatory affairs positions or comparable were interviewed.

Results:

the orientation guide was rated as very helpful and comprehensible. 80 % would highly recommend it to others.

Conclusion:

The orientation guide was positively evaluated and can be used in field. Nevertheless, further investigations must be performed, and a post-market surveillance will be necessary.

Background:

Digital health solutions have been omnipresent in policy agendas. However, we still need to better understand how citizens experience these developments and, more specifically, how citizens would ideally want such solutions to look like.

Objective:

We explore the needs and concerns citizens expressed in different phases of the co-creation process for a prototype of a citizen-centred health data platform within a large-scale European project: Smart4Health.

Method:

We follow a qualitative approach in our analysis of 9 discussion groups in addition to a diverse set of 49 qualitative interviews with citizens and health care professionals.

Results:

We show how citizens identify the positive potential of health data infrastructures and how they relate digital health to wider developments in contemporary societies. We then outline citizens’ concerns that potentially prevent them from becoming users and thus destabilize the policy vision of digital health.

Conclusion:

Four preconditions need to be met for citizens to find their place within a digital health data environment: transparency/trust, infrastructural literacy, digital justice, and a careful consideration of the distribution of responsibilities.

Background:

Process mining is a promising field of data analytics that is yet to be applied broadly in healthcare. It can streamline the care process, leading to a higher quality of care, increased patient safety and lower costs.

Objectives:

To get deeper insights into the emergence and detectability of delirium in a gerontopsychiatric setting.

Methods:

We use process mining to create process models from routinely collected, anonymised nursing data from two gerontopsychiatric wards. We analyse these models to get a longitudinal view of the care processes.

Results:

The process models comprise all activities during patients’ stays but are too extensive and challenging to interpret due to the wide variation in care paths. Although the models give insight into frequent paths and activities, they are insufficient to explain the emergence of delirium meaningfully. No apparent difference between stays with or without delirium could be detected.

Conclusion:

Conducting process mining on routinely collected data is easy, but the interpretation of the results was a challenge. We identified four limitations associated with using this data and gave recommendations on adapting it for further analysis.

Background:

Tele-rehabilitation is gaining importance due to the increasing need for objectiveness in the evaluation of patients with impaired motor functions. Low-cost marker-less motion capture systems are becoming key enabling technologies as support in the treatment of musculoskeletal diseases.

Objectives:

The goal of this work is to investigate the use of the Microsoft Azure Kinect device to develop a tele-rehabilitation platform for shoulder motor function recovery. The platform comprehends a set of serious games, which are fundamental to increase the patients’ engagement in shoulder rehabilitation.

Methods:

Starting from a set of functionalities identified together with the medical personnel of an Italian hospital, the Azure Kinect device has been used as motion capture system to interact with the serious games. Mobile applications for patients and physicians have been developed to manage the rehabilitation process.

Results:

The solution has been tested by the involved medical personnel. It has been considered interesting and promising. Further improvements in the design of the virtual environment of the serious games are required.

Conclusion:

The presented platform is a starting point to develop a complete IT solution for the daily shoulder rehabilitation.

Background:

Patient safety classifications are used to collect, classify and analyze patient safety data.

Objective:

This review was conducted to identify and compare the subject and coverage of existing patient safety classifications for Health Information Technology (HIT) and medical devices in which HIT may cover.

Methods:

All studies in patient safety that developed or extended any type of classification in HIT and medical devices were included. We identified and classified the covered concepts in these systems.

Results:

We identified 7 articles that met all of the inclusion criteria, resulting in 6 classifications. The most common patient safety subjects included adverse events and medication errors. Incident types and contributing factors/hazards were the most frequently covered concepts.

Conclusions:

Patient safety classifications in HIT cover more concepts and classes than medical device classifications. It is therefore recommended to improve existing classification systems in terms of covered concepts and classes.