HEALTHINF 2023 Abstracts

Full Papers Short Papers
Full Papers
Paper Nr: 9
Title:

Context Discovery and Cost Prediction for Detection of Anomalous Medical Claims, with Ontology Structure Providing Domain Knowledge
Authors:

James Kemp, Chris Barker, Norm Good and Michael Bain
Abstract: Medical fraud and waste is a costly problem for health insurers. Growing volumes and complexity of data add challenges for detection, which data mining and machine learning may solve. We introduce a framework for incorporating domain knowledge (through the use of the claim ontology), learning claim contexts and provider roles (through topic modelling), and estimating repeated, costly behaviours (by comparison of provider costs to expected costs in each discovered context). When applied to orthopaedic surgery claims, our models highlighted both known and novel patterns of anomalous behaviour. Costly behaviours were ranked highly, which is useful for effective allocation of resources when recovering potentially fraudulent or wasteful claims. Further work on incorporating context discovery and domain knowledge into fraud detection algorithms on medical insurance claim data could improve results in this field.
Download
Paper Nr: 15
Title:

Wearable EEG-Based Cognitive Load Classification by Personalized and Generalized Model Using Brain Asymmetry
Authors:

Sidratul Moontaha, Arpita M. Kappattanavar, Pascal Hecker and Bert Arnrich
Abstract: EEG measures have become prominent with the increasing popularity of non-invasive, portable EEG sensors for neuro-physiological measures to assess cognitive load. In this paper, utilizing a four-channel wearable EEG device, the brain activity data from eleven participants were recorded while watching a relaxation video and performing three cognitive load tasks. The data was pre-processed using outlier rejection based on a movement filter, spectral filtering, common average referencing, and normalization. Four frequency-domain feature sets were extracted from 30-second windows encompassing the power of δ, θ, α, β and γ frequency bands, the respective ratios, and the asymmetry features of each band. A personalized and generalized model was built for the binary classification between the relaxation and cognitive load tasks and self-reported labels. The asymmetry feature set outperformed the band ratio feature sets with a mean classification accuracy of 81.7% for the personalized model and 78% for the generalized model. A similar result for the models from the self-reported labels necessitates utilizing asymmetry features for cognitive load classification. Extracting high-level features from asymmetry features in the future may surpass the performance. Moreover, the better performance of the personalized model leads to future work to update pre-trained generalized models on personal data.
Download
Paper Nr: 16
Title:

Predicting Comorbidities in Diabetic Patients and Visualizing Data for Improved Healthcare
Authors:

Giridhar Krishnan and Waqar Haque
Abstract: Diabetes is one of the most common chronic diseases in the world with patients being more susceptible to develop additional comorbidities over time. In this research, we have used clinical data collected over six years to perform predictive and visual analytics which enables healthcare professionals gain valuable insight into early identification of the risk of developing comorbidities thereby resulting in effective diabetes management and reduced burden on healthcare system. We first present predictive models developed to forecast the likelihood of one of the three common comorbidities for diabetic patients – Benign Hypertension, Congestive Heart Failure, and Acute Renal Failure. The models use advanced data mining algorithms such as Logistic Regression, Neural Network, CHAID, Bayesian Network, Random Forest and Ensemble. Results from these models are incorporated into an interactive assessment tool that can take user input and predict the likelihood of developing one of these comorbidities. In addition, an interactive diabetes dashboard presents aggregated data using visually appealing charts, graphs, and tables. The dashboard also provides drilldown capabilities to allow navigation at finer granularities of various metrics.
Download
Paper Nr: 20
Title:

Leveraging Out-of-the-Box Retrieval Models to Improve Mental Health Support
Authors:

Theo Rummer-Downing and Julie Weeds
Abstract: This work compares the performance of several information retrieval (IR) models in the search for relevant mental health documents based on relevance to forum post queries from a fully-moderated online mental health service. Three different architectures are assessed: a sparse lexical model, BM25, is used as a baseline, alongside two neural SBERT-based architectures - the bi-encoder and the cross-encoder. We highlight the credibility of using pretrained language models (PLMs) out-of-the-box, without an additional fine-tuning stage, to achieve high retrieval quality across a limited set of resources. Error analysis of the ranking results suggested PLMs make errors on documents which contain so called red-herrings - words which are semantically related but irrelevant to the query - whereas human judgements were found to suffer when queries are vague and present no clear information need. Further, we show that bias towards an author’s writing style within a PLM affects retrieval quality and, therefore, can impact on the success of mental health support if left unaddressed.
Download
Paper Nr: 22
Title:

Recommender System for Alarm Thresholds in Medical Patient Monitors
Authors:

Denise Schmidt, Jonas Chromik and Bert Arnrich
Abstract: Intensive care unit staff relies on patient monitors to identify critical conditions. The monitors trigger alarms as soon as the patient’s vital parameters deviate from predefined threshold ranges. However, these ranges are usually not adapted to the individual patient. High numbers of false alarms burden clinical staff and pose a major risk to patient safety. We propose a recommender system for threshold values to enable a patient-centered monitoring system. This can reduce false alarms caused by default monitoring settings. We employ CatBoost – a gradient boosting algorithm – to predict blood pressure and heart rate thresholds. We use SHAP values to evaluate the importance of different patient characteristics, diagnoses, or medications. Several patient characteristics show an impact on the model output: Diagnoses, first care unit, vital parameter measurements, and the amount of general anaesthetics are the most important features in all threshold models. The recommendations of our system deviate from the actual thresholds by approximately 3.5 bpm for the heart rate and 4.9 mmHg for the blood pressure thresholds. Blood pressure thresholds have a higher variance which leads to larger errors. However, the underlying data is not very patient-centered and we require better alarm data to further improve threshold recommendation.
Download
Paper Nr: 24
Title:

Predicting the Socio Economic Status of end Users of a Maternal Health App by Machine Learning
Authors:

Rajanikant Ghate, Sumiti Saharan and Rahee Walambe
Abstract: Digital technologies posit an immense opportunity to provide scalable solutions for narrowing the health equity gap and proving affordable access to quality healthcare in low resource settings. A key step towards harnessing the power of digital health is developing a scalable mechanism for identifying the socioeconomic profile of end users. Socio-economic status (SES) of individuals has been classically estimated through standard questionnaires. This methodology is not scalable and prone to immense bias if implemented digitally as a self-report questionnaire. Together for Her (TFH) is a digital app for pregnancy that aims to provide equitable access to quality pregnancy information and support to pregnant women in India. To assess our reach to users from low socio-economic settings, we developed a machine learning model that leverages digital indices for estimated SES. We propose this approach holds immense value for digital health interventions, both as a mechanism for gaining insight on the socio-economic profile of users being reached and as an evaluation metric for interventions aimed at driving health equity.
Download
Paper Nr: 28
Title:

An Easy-to-Use and Robust Approach for the Differentially Private De-Identification of Clinical Textual Documents
Authors:

Yakini Tchouka, Jean-François Couchot and David Laiymani
Abstract: Unstructured textual data is at the heart of healthcare systems. For obvious privacy reasons, these documents are not accessible to researchers as long as they contain personally identifiable information. One way to share this data while respecting the legislative framework (notably GDPR or HIPAA) is, within the medical structures, to de-identify it, i.e. to detect the personal information of a person through a Named Entity Recognition (NER) system and then replacing it to make it very difficult to associate the document with the person. The challenge is having reliable NER and substitution tools without compromising confidentiality and consistency in the document. Most of the conducted research focuses on English medical documents with coarse substitutions by not benefiting from advances in privacy. This paper shows how an efficient and differentially private de-identification approach can be achieved by strengthening the less robust de-identification method and by adapting state-of-the-art differentially private mechanisms for substitution purposes. The result is an approach for de-identifying clinical documents in French language, but also generalizable to other languages and whose robustness is mathematically proven.
Download
Paper Nr: 30
Title:

Novel Distributed Informatics Platform to Support Machine Learning Discovery of Metabolic Biomarkers in Hypoxia Predisposition
Authors:

Anthony Stell, Vedant Chauhan, Sandra Amador, Felix Beuschlein, Judith Favier, David Gil, Philip Greenwood, Ronald de Krijger, Matthias Kroiss, Samanta Ortuno, Attila Patocs and Axel Walch
Abstract: To realise the scientific and clinical benefits of machine learning (ML) in a multi-centre research collaboration, a common issue is the need to bring high-volume data, complex analytical algorithms, and large-scale processing power, all together into one place. This paper describes the detailed architecture of a novel platform that combines these features, in the context of a proposed new clinical/bioinformatics project, Hypox-PD. Hypox-PD uses ML methods to identify new metabolic biomarkers, through the analysis of high-volume data including mass spectrometry and imaging morphology of biobank tissue. The platform features three components: a content delivery network (CDN); a standardised orchestration application; and high-specification processing power and storage. The central innovation of this platform is a distributed application that simultaneously manages the workflow between these components, provides a virtual mapping of the domain data dictionary, and presents the project data/metadata in a FAIR-compliant external interface. This paper presents the detailed design specifications of this platform, as well as initial test results in establishing the benchmark challenge of current direct transfer times without any specialised support. An initial costing of CDN usage is also presented, which indicates that significant performance improvement may be achievable at a reasonable cost to research budgets.
Download
Paper Nr: 32
Title:

Development, Implementation and Validation of a Stochastic Prediction Model of UICC Stages for Missing Values in Large Data Sets in a Hospital Cancer Registry
Authors:

Sebastian Appelbaum, Daniel Krüerke, Stephan Baumgartner, Marianne Schenker and Thomas Ostermann
Abstract: Cancer is still a fatal disease in many cases, despite intensive research into prevention, treatment and follow-up. In this context, an important parameter is the stage of the cancer. The TNM/UICC classification is an important method to describe a cancer. It dates back to the surgeon Pierre Denoix and is an important prognostic factor for patient survival. Unfortunately, despite its importance, the TNM/UICC classification is often poorly documented in cancer registries. The aim of this work is to investigate the possibility of predicting UICC stages using statistical learning methods based on cancer registry data. Data from the Cancer Registry Clinic Arlesheim (CRCA) were used for this analysis. It contains a total of 5,305 records of which 1,539 cases were eligible for data analysis. For prediction classification and regression trees, random forests, gradient tree boosting and logistic regression are used as statistical methods for the problem at hand. As performance measures Mean misclassification error (mmce), area under the receiver operating curve (AUC) and Cohen’s kappa are applied. Misclassification rates were in the range of 28.0% to 30.4%. AUCs ranged between 0.73 and 0.80 and Cohen kappa showed values between 0.39 and 0.44 which only show a moderate predictive performance. However, with only 1,539 records, the data set considered here was significantly lower than those of larger cancer registries, so that the results found here should be interpreted with caution.
Download
Paper Nr: 41
Title:

Predicting Hospital Length of Stay of Patients Leaving the Emergency Department
Authors:

Alexander Winter, Mattis Hartwig and Toralf Kirsten
Abstract: In this paper, we aim to predict the patient’s length of stay (LOS) after they are dismissed from the emergency department and transferred to the next hospital unit. An accurate prediction has positive effects for patients, doctors and hospital administrators. We extract a dataset of 181,797 patients from the United States and perform a set of feature engineering steps. For the prediction we use a CatBoost regression architecture with a specifically implemented loss function. The results are compared with baseline models and results from related work on other use cases. With an average absolute error of 2.36 days in the newly defined use case of post ED LOS prediction, we outperform baseline models achieve comparable results to use cases from intensive care unit LOS prediction. The approach can be used as a new baseline for further improvements of the prediction.
Download
Paper Nr: 45
Title:

A Question of Trust: Old and New Metrics for the Reliable Assessment of Trustworthy AI
Authors:

Andrea Campagner, Riccardo Angius and Federico Cabitza
Abstract: This work contributes to the evaluation of the quality of decision support systems constructed with Machine Learning (ML) techniques in Medical Artificial Intelligence (MAI). In particular, we propose and discuss metrics that complement and go beyond traditional assessment practices based on the evaluation of accuracy, by focusing on two different dimensions related to the trustworthiness of a MAI system: reputation/ability, which relates to the accuracy or predictive ability of the system itself; and expertise/source reliability, which relates instead to the trustworthiness of the data which have been used to construct the MAI system. Then, we will discuss some previous, but so far mostly neglected, proposals as well novel metrics, visualizations and procedures for the sound evaluation of a MAI system’s trustworthiness, by focusing on six different concepts: advice accuracy, advice reliability, pragmatic utility, advice value, decision benefit and potential robustness. Finally, we will illustrate the application of the proposed concepts through two realistic medical case studies.
Download
Paper Nr: 51
Title:

Analysis of Virtual Reality Therapy Game Prototype for Persons Living with Dementia in the Philippines
Authors:

Veeda M. Anlacan, Roland G. Jamora, Angelo F. Panganiban, Isabel O. Salido, Romuel Z. Apuya, Bryan C. Galecio, Michael L. Tee, Maria R. Aguila, Cherica A. Tee and Jaime L. Caro
Abstract: Immersive technologies such as virtual reality (VR) had rapidly gained interest as part of the technological revolution in healthcare. Because of its inspirational affectation, VR had been considered a potential intervention for individuals with impaired memory such as dementia. This paper had two goals that were achieved in collaboration with healthcare professionals and scientists from the Philippines: (a) to create an improved VR therapy game for persons with behavioral and psychological symptoms of dementia (BPSD), and (b) to acquire professional insights and recommendations about the game. With this, a VR game prototype was developed and tested among five health and four game design professionals. A focus group discussion (FGD) was then held to discuss the participants’ experience with the game. The results of the FGD provided an in-depth analysis regarding the game’s architecture and overall design for the use of persons with BPSD in the Philippines. Critical points discussed in this paper may be adopted in future VR studies for general healthcare applications personalized for a specific demographic.
Download
Paper Nr: 52
Title:

Lessons Learned from mHealth Monitoring in the Wild
Authors:

Pedro Almir M. Oliveira, Rossana C. Andrade and Pedro A. Santos Neto
Abstract: In the modern world, it is no overstatement to say that “our devices know us better than we know ourselves”. In this sense, the vast amount of data generated by wearables, mobile devices, and environmental sensors has enabled the development of increasingly personalized and intelligent services. Among them, there is a growing interest in the delivery of medical practice using mobile devices (i.e., mobile health or mHealth). mHealth makes it possible to optimize healthcare systems based on continuous and transparent health monitoring, aiming to detect the emergence of diseases. However, mHealth monitoring in the real world (i.e., uncontrolled environment or, as labeled in this paper, “in the wild”) has many challenges. Therefore, this practical report discusses ten lessons learned from the Quality of Life (QoL) monitoring of twenty-one volunteers over three months. The main objective of this QoL monitoring was to collect data capable of training Machine Learning algorithms to infer users’ Quality of Life using the WHOQOL-BREF as a reference. During this period, our research team systematically recorded the problems faced and the strategies to overcome them. Such lessons can support researchers and practitioners in planning future studies to avoid or mitigate similar issues. In addition, we present strategies for dealing with each challenge using the 5W1H model.
Download
Paper Nr: 54
Title:

On the Use of Generative Adversarial Networks to Predict Health Status Among Chronic Patients
Authors:

María T. Jurado-Camino, David Chushig-Muzo, Cristina Soguero-Ruiz, Pablo M. Bohoyo and Inmaculada Mora-Jiménez
Abstract: Chronic diseases (CD) are the leading cause of death worldwide, presenting higher mortality rates and economic burden (both in the health and social context) as the complexity of the CD increases. The use of Electronic Health Records (EHRs) and Machine Learning (ML) contribute to significant progress in health domain research, supporting identifying the patient's health status for early interventions. Despite these achievements, the class imbalance can limit the generalization capability of many ML models and data augmentation techniques are proposed to face this limitation. In this work, a Generative Adversarial Network named medWGAN is used to generate synthetic patients considering clinical data collected from EHRs linked to the University Hospital of Fuenlabrada. Data are associated with patients diagnosed with both simple CD (diabetes, hypertension, congestive heart failure, chronic obstructive pulmonary disease) and multiple CD. Experimental work using decision trees as predictors to determine the patient's health status showed the ability of medWGAN for preserving the underlying (high-dimensional and sparse) clinical patterns. Our results indicate that the identification of patients with multiple CD may benefit from the use of medWGAN as long as the data used for its training is diverse enough, contributing to supporting clinical decision-making in complex scenarios with many features.
Download
Paper Nr: 81
Title:

How Different Elements of Audio Affect the Word Error Rate of Transcripts in Automated Medical Reporting
Authors:

Emma Kwint, Anna Zoet, Katsiaryna Labunets and Sjaak Brinkkemper
Abstract: Automated Speech Recognition software is implemented in different fields. One of them is healthcare in which it can be used for automated medical reporting, the field of focus of this research. For the first step of automated medical reporting, audio files of consultations need to be transcribed. This research contributes to the investigation of the optimization of the generated transcriptions, focusing on categorizing audio files on specific characteristics before analyzing them. The literature research within this study shows that specific elements of speech signals and audio, such as accent, voice frequency and noise, can have influence on the quality of a transcription an Automated Speech Recognition system carries out. By analyzing existing medical audio data and conducting an pilot experiment, the influence of those elements is established. This is done by calculating the Word Error Rate of the transcriptions, a useful percentage that shows the accuracy. Results of the analysis of the existing data show that noise is an element that carries out significant differences. However the data of the experiment did not show significant differences. This was mainly due to having not enough participants to reason with significance. Further research into the effect of noise, language and different Automated Speech Recognition technologies should be done based on the outcomes of this research.
Download
Paper Nr: 82
Title:

Using Balancing Methods to Improve Glycaemia-Based Data Mining
Authors:

Diogo Machado, Vítor S. Costa and Pedro Brandão
Abstract: Imbalanced data sets pose a complex problem in data mining. Health related data sets, where the positive class is connected to the existence of an anomaly, are prone to be imbalanced. Data related to diabetes management follows this trend. In the case of diabetes, patients avoid situations of hypo/hyperglycaemia, which is the anomaly we want to detect. The use of balancing methods can provide more examples of the minority class, and assist the classifier by clearing the decision boundary. Nevertheless, each over-sampling and under-sampling method can affect the data set uniquely, which will influence the classifier’s performance. In this work, the authors studied the impact of the most known data-balancing methods applied to the Ohio and St. Louis diabetes related data sets. The best and most robust approach was the use of ENN with SMOTE. This hybrid method produced significant performance gains on all the performed tests. ENN in particular had a meaningful impact on all the tests. Given the limited volume of glycaemia-based data available for diabetes management, over-sampling methods would be expected to have a greater role in improving the classifier’s performance. In our experiments, the clearing of noise values by the under-sampling methods, produced better results.
Download
Paper Nr: 83
Title:

HydReminder-W: A Bottle Cap that Listens to Your Heart to Remind You to Drink!
Authors:

Nishiki Motokawa, Anna Yokokubo and Guillaume Lopez
Abstract: Hydration is essential for maintaining life. Consequently, a lack of awareness of the amount of water required and the actual amount of water intake has been issued for all generations and may lead to diseases that result in death. However, existing systems still have limited capabilities for tracking hydration reminder systems in unconstrained, realistic environments. Therefore, this study employs personalized information based on the user’s biometric information and environmental information. We proposed and developed a smart bottle cap system that includes an environmental sensor, a barometric pressure sensor, an infrared sensor, the user’s heart rate from a smartwatch, and a cap-shaped shell to create a compact system. Our proposed system, ‘HydReminder-W,’ promotes proper hydration using the individual’s biometric information, which teaches individuals about the lack of hydration. A comparison of hydration with HydReminder-W showed that the amount of water intake increased with the use of HydReminder-W. The average System Usability Scale (SUS) score of all the subjects was 78.9 points, confirming that HydReminder-W has excellent usability and is helpful as a hydration promotion system.
Download
Short Papers
Paper Nr: 2
Title:

A Systematic Review and Recommendation of Software Architectures for SARS-CoV-2 Monitoring
Authors:

Kay Smarsly, Yousuf Al-Hakim, Patricia Peralta, Silvio Beier and Claudia Klümper
Abstract: The coronavirus disease 2019 (COVID-19) is a highly infectious respiratory disease caused by the severe acute respiratory syndrome coronavirus 2 (SARS‑CoV‑2), which has led to the ongoing global pandemic with more than half a billion cases worldwide. Response measures to COVID-19 outbreaks suffer from lagging between detecting and reporting COVID-19 cases and by underreporting. By contrast, using wastewater allows detecting SARS-CoV-2 ribonucleic acid (RNA) in human feces, serving as a timely and reliable basis for devising effective measures to prevent and control COVID-19 outbreaks. As a technological basis, software systems for monitoring SARS-COV-2 RNA in wastewater are required, which are capable of (i) interlinking COVID-19-related data from different sources, (ii) providing user interfaces with remote access, (iii) implementing software design concepts that are well-established, and (iv) deploying on-demand SARS-CoV-2 data analysis. To ensure reliable operation, it is crucial to set up SARS-COV-2 monitoring systems based on sound software architectures. This paper systematically reviews and categorizes software architectures for SARS-CoV-2 monitoring systems, considering journals, book series, and conference proceedings indexed in the Scopus database. Then, a software architecture for SARS-CoV-2 monitoring systems is proposed. In future work, the proposed software architecture may be implemented and validated for SARS-CoV-2 monitoring.
Download
Paper Nr: 5
Title:

Fall Prediction Amongst the Elderly Using Data from an Ambient Assisted Living System
Authors:

Philip Branch, Divya B. Sridharam, Andre Ferretto and Tim Carroll
Abstract: Falls amongst the elderly are life threatening. Being able to predict falls means steps could be taken to reduce fall likelihood or severity. In this paper we report on our work using data generated by HalleyAssist, an advanced Ambient Assisted Living System, to predict falls amongst the elderly. HalleyAssist unobtrusively monitors older people using sensors to provide services to help them with their day-to-day activities. We conducted a three-month trial of the HalleyAssist system with six households of older people primarily to gauge acceptance and utility of the system. During the trial we also asked participants to keep a ’falls diary’ in which they recorded the date, time and location of any falls. After the initial trial we continued monitoring one of the participants (with her consent) who was susceptible to falls, for an additional seven months. Over the ten months of the trial she fell 32 times on 28 days. None of the other participants fell during the trial. We analysed data from the sensors and correlated it with whether she fell later in the day. Using techniques from machine learning we were able to identify features that enabled a fall to be predicted with 64.9 % accuracy.
Download
Paper Nr: 6
Title:

Designing Personalised Gamification of mHealth Survey Applications
Authors:

Paulina Adamczyk, Sylwia Marek, Ryszard Pr˛ecikowski, Maciej Kuś, Michał Grzeszczyk, Maciej Malawski and Aneta Lisowska
Abstract: To monitor patients’ well-being and evaluate the efficacy of digital health intervention, patients are required to regularly respond to standardised surveys. Responding to a large number of questionnaires is effortful and may discourage mHealth app users from engaging with the intervention. Gamification might reduce the burden of self-reporting. However, researchers have adopted various approaches to the personalisation of gamification design: ranking of game elements by the user, Hexad Gamification User Types classification (G) and selection of preferred design mockups (MU) . In this paper we report on a small population study involving 54 healthy participants aged 17 to 60, and investigate if these alternative approaches lead to the same design choices. We find that different evaluation approaches lead to different choices of gamification elements. We suggest to use game element ranking in combination with mockup selection. Hexad player classification might be less useful in the context of mHealth applications design.
Download
Paper Nr: 10
Title:

Daily Pain Prediction in Workplace Using Gaussian Processes
Authors:

Chetanya Puri, Stijn Keyaerts, Maxwell Szymanski, Lode Godderis, Katrien Verbert, Stijn Luca and Bart Vanrumste
Abstract: Work-related Musculoskeletal disorders (MSDs) account for 60% of sickness-related absences and even permanent inability to work in the Europe. Long term impacts of MSDs include “Pain chronification” which is the transition of temporary pain into persistent pain. Preventive pain management can lower the risk of chronic pain. It is therefore important to appropriately assess pain in advance, which can assist a person in improving their fear of returning to work. In this study, we analysed pain data acquired over time by a smartphone application from a number of participants. We attempt to forecast a person’s future pain levels based on his or her prior pain data. Due to the self-reported nature of the data, modelling daily pain is challenging due to the large number of missing values. For pain prediction modelling of a test subject, we employ a subset selection strategy that dynamically selects a closest subset of individuals from the training data. The similarity between the test subject and the training subjects is determined via dynamic time warping-based dissimilarity measure based on the time limited historical data until a given point in time. The pain trends of these selected subset subjects is more similar to that of the individual of interest. Then, we employ a Gaussian processes regression model for modelling the pain. We empirically test our model using a leave-one-subject-out cross validation to attain 20% improvement over state-of-the-art results in early prediction of pain.
Download
Paper Nr: 14
Title:

Benchmarking Disease Modeling Techniques on the Philippines’ COVID-19 Dataset
Authors:

Christian Pulmano and Proceso Fernandez
Abstract: The COVID-19 pandemic has emphasized the importance of timely and accurate prediction of disease outbreaks. Mathematical disease models can help simulate the trajectory of diseases and guide policymakers in identifying priorities and gaps in current policies. This study evaluates the performance, on various metrics, of three different parameter estimation algorithms in compartmental models, i.e., Nelder-Mead, Simulated Annealing, and L-BFGS-B, together with the ARIMA time-series modeling, in modeling COVID-19 cases. Using the daily number of confirmed cases of COVID-19 in the Philippines as the dataset, the models were trained on 90 different periods, with each period having 30 days of case data. After training, the models were used to predict the cases up to 30 days later. The Negative Log Likelihood (NLL), time spent, iterations per second, and memory allocation were all measured. The results show that ARIMA performed better in terms of accuracy, time, and space efficiency than each of the other algorithms. This suggests that ARIMA should be preferred for predicting the number of cases. However, policymaking sometimes requires scenario-based modeling, which ARIMA is unable to provide. For such requirements, any of the three compartmental models may be preferred, as each performed generally very well, too.
Download
Paper Nr: 18
Title:

IT-Structures and Algorithms for Quality Assurance in the Medical Advisory Service Institutions in Germany. Step 2: To err is Human. Consensus-Conferences
Authors:

Vera Ries, Klaus-Peter Thiele, Bernhard van Treeck, Sarah Schroeer, Christina Witt and Reinhard Schuster
Abstract: 16 Regional Medical Advisory Service Institutions perform medical expertise assessments upon German in- and out-patient care. Assessments have to accomplish a nationwide quality assurance plan with mandatory public reporting. We developed strategies to resolve conflicting quality measurement evaluations in the same item by different peers without unveiling the identity of the criticised medical expert or peer in the processes. All workflows are completely digitalized using mathematical IT-based procedures for randomized sampling and for an equal distribution of the medical expertise assessments to be reviewed. We even allow for smaller sample sizes, so regional heterogeneity and the heterogeneity of the types of medical expertise assessment pose a constraint satisfaction problem. We discuss models addressing this kind of problem type and present possible solutions. Our technical framework for peer review distribution, data collection and final result analysis includes a completely IT-based workflow not only masking the origin of the medical expertise assessments discussed, but routing the peer review processes in a way that independent and impartial review sheets are produced by peers that were previously not yet involved in the reviewing process. Finally, the statistical distribution and outcomes of the review results are analysed.
Download
Paper Nr: 29
Title:

Facilitating SNOMED-CT Template Creation by Targeting Stopwords
Authors:

Rashmi Burse, Michela Bertolotto and Gavin McArdle
Abstract: Quality Assurance (QA) of biomedical ontologies is a major challenge in the health-informatics domain. One of the preliminary ways in which we can maintain the quality of a biomedical ontology is by ensuring consistency in the modelling styles of biomedical concepts. Maintaining consistency in the lexical, structural and ontological modelling of biomedical concepts reduces a concept’s susceptibility to errors. SNOMED-CT, which is one of the most widely adopted biomedical ontologies, strives to achieve this consistency by creating templates for logical definitions based on the description of biomedical concept names. The work presented here in based on the observation that the majority of the SNOMED-CT templates contain stopwords (non-medical terms) in their description that indicate a relationship between two medical concepts. We hypothesize that the process of creating SNOMED-CT templates can be automated to a large extent by targeting stopwords. In this work, we present a method that exploits stopwords in concept names to create templates for the structural and logical modelling of lexically and semantically similar biomedical concepts. The results have shown promising potential by extracting a multitude of SNOMED-CT templates, exhibiting more than 200 templates for the stopword of. Given the high demand for QA of biomedical ontologies, these results are highly beneficial in automating the existing mechanisms employed in maintaining consistency in the modeling of SNOMED-CT concepts. The presented method can be used as a complementary process to mitigate the manual efforts of SNOMED-CT curators. Furthermore, auditing potentially incomplete definitions of SNOMED-CT concepts using the extracted templates has identified 49-87% inconsistent concepts for the stopwords of and in in the biomedical ontology.
Download
Paper Nr: 31
Title:

(ε, k)-Randomized Anonymization: ε-Differentially Private Data Sharing with k-Anonymity
Authors:

Akito Yamamoto, Eizen Kimura and Tetsuo Shibuya
Abstract: As the amount of biomedical and healthcare data increases, data mining for medicine becomes more and more important for health improvement. At the same time, privacy concerns in data utilization have also been growing. The key concepts for privacy protection are k-anonymity and differential privacy, but k-anonymity alone cannot protect personal presence information, and differential privacy alone would leak the identity. To promote data sharing throughout the world, universal methods to release the entire data while satisfying both concepts are required, but such a method does not yet exist. Therefore, we propose a novel privacy-preserving method, (ε, k)-Randomized Anonymization. In this paper, we first present two methods that compose the Randomized Anonymization method. They perform k-anonymization and randomized response in sequence and have adequate randomness and high privacy guarantees, respectively. Then, we show the algorithm for (ε, k)-Randomized Anonymization, which can provide highly accurate outputs with both k-anonymity and differential privacy. In addition, we describe the analysis procedures for each method using an inverse matrix and expectation-maximization (EM) algorithm. In the experiments, we used real data to evaluate our methods’ anonymity, privacy level, and accuracy. Furthermore, we show several examples of analysis results to demonstrate high utility of the proposed methods.
Download
Paper Nr: 34
Title:

An Android App for Posture Analysis Using OWAS
Authors:

Christian Lins, Franziska D. Quang, Rica Schulze, Stefanie Lins, Andreas Hein and Sebastian Fudickar
Abstract: In this paper the APA, an App for Posture Analysis, that incorporates the OWAS method (Ovako Working Posture Analysis System) for assessing potential harmful postures of physically hard working employees, is presented and evaluated. The app is intended as a tool for occupational safety experts in assessing the individual postures of workers to identify and prevent harmful working situations in regards to musculoskeletal hazards. For this, APA incorporates a digitized assessment sheet for the OWAS method and timing support that helps occupational safety experts structure and simplify the assessment workflow. To investigate whether the protocol sheet can be replaced with the app, a study was conducted in which the inter-rater reliability among the app and the paper protocol sheet (as control group) was evaluated. In addition, the usability of the app was determined via the User Experience Questionnaire (UEQ). In the study, the app achieved higher inter-rater reliability for watching a video recording of postures than the control group. The chi-square test revealed differences in the use of the app and paper only for leg postures. The UEQ indicated overall above average results, which indicates a sufficient usability, which was also confirmed by free-textual comments of the 13 study participants. The results suggest that it would be possible to replace the paper protocol sheet in the future with an app.
Download
Paper Nr: 35
Title:

A Mobile Application for Milano Ventilatore Meccanico: A First Prototype
Authors:

Silvia Bonfanti, Angelo Gargantini and Luca Novelli
Abstract: In hospitals the need to have devices connected and accessible remotely is increasing, in order to continuously monitor patients. This need also arose for mechanical ventilators. In this paper, we introduce the first prototype of a mobile application to connect remotely to Milano Ventilatore Meccanico, a mechanical ventilator developed during COVID-19 pandemic. We show the process adopted to design and develop the first prototype in Android.
Download
Paper Nr: 36
Title:

Evaluating the Effects of a Priori Deep Learning Image Synthesis on Multi-Modal MR-to-CT Image Registration Performance
Authors:

Nils Frohwitter, Alessa Hering, Ralf Möller and Mattis Hartwig
Abstract: Radiation therapy often requires a computed tomography (CT) for treatment planning and an additional magnetic resonance (MR) imaging prior to the treatment for adaptation. With two different images from the same scene, multi-modal image registration is needed to align areas of interest in both images. One idea to improve the registration process is to perform an image synthesis that converts one image mode into another mode prior to the registration. In this paper, we address the research needed to perform a thorough evaluation of the synthesis step on overall registration performance using different well-known registration methods of the Advanced Normalization Tools (ANTs) framework. Given abdominal images, we use CycleGAN for synthesis and compare the registration performance to the one without synthesis by using four different well-known registration methods. We show that good image synthesizing results lead to an average improvement in all registration methods, biggest improvement being achieved for the ‘Symmetric Normalization’ method with 8% (measured with Dice-score). The overall best registration method with prior synthesis is ‘Symmetric Normalization and Rigid’. Furthermore, we show that the images with bad synthetic results lead to worse registration, thus suggesting the correlation between synthesizing quality and registration performance.
Download
Paper Nr: 38
Title:

On the Problem of Data Availability in Automatic Voice Disorder Detection
Authors:

Dayana Ribas, Antonio Miguel, Alfonso Ortega and Eduardo Lleida
Abstract: In order to support medical doctors in having more versatile health assistance, automatic voice disorder detection systems enable the remote diagnosis, treatment, and monitoring of voice pathologies. The main problem for developing the related technology is the availability of audio data of healthy and pathological voices manually labeled by experts. Saarbruecken Voice Database (SVD) was created in 1997, with a collection of more than 5 hours of healthy and pathologica audio data. This database has been widely used for developing voice disorder detection systems. However, it has some issues in the distribution of data and the labeling that makes it difficult to conduct conclusive studies. This paper evaluates an Automatic Voice Disorder Detection (AVDD) system using the recent Advanced Voice Function Assessment Database (AVFAD) with almost 40 hours of audio data and SVD as a reference. The system consists of a representation using spectral, prosody, and voice quality parameters followed by an SVM classifier that can obtain up to 88% accuracy in phrases and 86% in sustained vowel a. Data augmentation strategy is assessed for handling the problem of data imbalance with the SMOTE method which improves the performance of male, female, and gender-independent models without decreasing the results for scenarios with data balance. Finally, we release the system implementation for voice disorder detection including the list of train-test partitions for both databases.
Download
Paper Nr: 40
Title:

Benchmarking the BRATECA Clinical Data Collection for Prediction Tasks
Authors:

Bernardo S. Consoli, Renata Vieira and Rafael H. Bordini
Abstract: Expanding the usability of location-specific clinical datasets is an important step toward expanding research into national medical issues, rather than only attempting to generalize hypotheses from foreign data. This means that benchmarking such datasets, thus proving their usefulness for certain kinds of research, is a worthwhile task. This paper presents the first results of widely used prediction tasks from data contained within the BRATECA collection, a Brazilian tertiary care data collection, and also results for neural network architectures using these newly created test sets. The architectures use both structured and unstructured data to achieve their results. The obtained results are expected to serve as benchmarks for future tests with more advanced models based on the data available in BRATECA.
Download
Paper Nr: 42
Title:

Eating Habit Improvement System Using Dietary Sound
Authors:

Haruka Kamachi, Sae Ohkubo, Anna Yokokubo and Guillaume Lopez
Abstract: Obesity may cause lifestyle diseases such as diabetes and high blood pressure. Eating slowly and chewing well are essential to prevent obesity. This research aims to improve the consciousness of dietary behavior based on eating habits by quantifying eating behavior. It proposes “ChewReminder,” a smartphone application software that detects eating activities in real-time under a natural meal environment and gives feedback based on detected activity. ChewReminder detects four activities: chewing, swallowing, talking, and other.The smartwatch gives feedback using vibration depend on chewing count per one bite which information was linked from the smartphone. Also, the total feedback about the meal was displayed on the smartphone after finishing the meal. The chewing count for 70% subjects and chewing pace for more than half subjects was improved with using ChewReminder by the result of total chewing count, average of chewing count per bite and chewing pace. ChewReminder is effective especially people who are aware of fast eating. Also, the result of long-term experiment indicated that feedback displayed on a smartphone was effective to improve consciousness of eating activity. Therefore, the result of both experiment shows that ChewReminder is a valid system to improve consciousness of eating activity especially chewing activity.
Download
Paper Nr: 43
Title:

Extended Head Pose Estimation on Synthesized Avatars for Determining the Severity of Cervical Dystonia
Authors:

Roland Stenger, Sebastian Löns, Feline Hamami, Nele Brügge, Tobias Bäumer and Sebastian Fudickar
Abstract: We present an extended head pose estimation algorithm, which is trained exclusively on synthesized human avatars. Having five degrees of freedom to describe such head poses, this task can be regarded as being more complex than predicting the absolute rotation only with three degrees of freedom, which is commonly known as head pose estimation. Due to the lack of labeled data sets containing such complex head poses, we created a data set, consisting of renderings of avatars. With this extension, we take a step towards an algorithm that can make a qualitative assessment of cervical dystonia. Its symptomatic consists of an involuntary twisted head posture, which can be described by those five degrees of freedom. We trained an EfficientNetB2 and evaluated the results with the mean absolute error (MAE). Such estimation is possible, but the performance works differently well for the five degrees of freedom, with an MAE between 1.71° and 6.55°. By visually randomizing the domain of the avatars, the gap between real subject photos and the simulated ones might tend to be smaller and enables our algorithm being used on real photos in the future, while being trained on renderings only.
Download
Paper Nr: 44
Title:

Concept for General Improvements in the Treatment of Femoral Shaft Fractures with an Intramedullary Nail
Authors:

Finn Siegel, Christian Buj, Ralf Schwanbeck, Andreas Petersik, Ulrich Hoffmann, Jakob Kemper, Frank Hildebrand, Philipp Kobbe, Jörg Eschweiler, Johannes Greven, Ricarda Merfort, Christian Freimann, Astrid Schwaiger and Frerk M. Aschwege
Abstract: The gold standard for femoral shaft fracture treatment is intramedullary (IM) nailing. This principle has gained acceptance because of the good fracture healing rate and the rapid return to full weight-bearing of the leg. Nevertheless, a significant number of patients suffer from impairments in everyday life years after treatment. This paper discusses various causes and presents possible solutions: a) Improving the IM nailing procedure by developing a new intraoperative assistance system to precisely restore length and rotation angle of the injured femur. b) Improving rehabilitation after IM nailing treatment, through home monitoring. c) Increasing data safety, standardization, and centralization along the entire patient pathway, enabling analytics to statistically verify improvements in IM nailing treatments.
Download
Paper Nr: 58
Title:

A Gait Analysis Tool Based on Machine Learning to Support the Rehabilitation Strategy of Post-stroke Patients
Authors:

Nicoletta Balletti, Gennaro Laudato and Rocco Oliveto
Abstract: Stroke is a serious medical condition that can result in permanent brain damage and other pathological issues. Conditions suffered by survivors ranged in severity from full recovery to significant movement disability. Even though some may recover quickly, many stroke survivors require long-term support to help them achieve as much independence as they can. Thanks to a proper rehabilitation, patients who have experienced a stroke can work to regain skills that are suddenly lost when a section of their brain is injured. Due to the breakdown of neuronal networks in the motor cortex, abnormal gait patterns are a typical disability after a stroke. Therefore, gait analysis can be a powerful tool to support stroke patients during rehabilitation. In this work we propose GIULYO, a Machine Learning based tool that offers support in the assessment of video gait trials in stroke patients by providing an automatic analysis on the muscle activity of the assisted subject. GIULYO is a device-agnostic tool because it accepts motion tracking data in terms of 3d trajectories regardless of the type of instrumentation. GIULYO has been validated on the ARRA Stroke dataset and the results showed an overall accuracy of 0.74 while on a subset a patients—with common clinical assessment of mobility impairments—the accuracy increased to 0.92, therefore demonstrating the feasibility of involving a ML-based approach for the rehabilitation support of post stroke patients.
Download
Paper Nr: 60
Title:

A Correlation Network Model for Analyzing Mobility Data in Depression Related Studies
Authors:

Rama K. Thelagathoti and Hesham H. Ali
Abstract: Depression is a serious behavioural disorder that can affect the quality of life. Timely diagnosis and accurate estimation of severity are critical in supporting depression studies and starting early interventional treatment. In this study, we introduce two major contributions. First, we propose a novel computational model that can utilize non-invasive mobility data to recognize individuals suffering from depression disorders. Second, we introduce a new objective index, the Depression Severity Score Index (DSS), which can approximate the seriousness or the degree of severity of depression. The proposed approach is a data-driven model that is built on the mobility data collected from 55 subjects using wearable sensors. In the first step in our proposed approach, a graph model that represents the underlying correlation network is constructed by measuring the pair-wise correlation values between each pair of individuals. Then, we obtain the depression severity index of each subject by utilizing graph properties of the constructed network such as Intra and inter-cluster edges. Our obtained results show that the obtained correlation network model has the potential to identify participants diagnosed with depression from the control group. Moreover, the proposed Depression Severity Score (DSS) has a higher likelihood than the clinical depression score in correctly measuring the depression severity level.
Download
Paper Nr: 64
Title:

Analysis of Driving Behavior by Applying LDA Topic Model at Intersection Using VR Simulator
Authors:

Hyeokmin Lee, Hosang Moon, Jaehoon Kim, Jaeheui Lee, Eunghyuk Lee and Sungtaek Chung
Abstract: The present study aims to analyze driving style and latent driving behavior typically at intersections where various driving habits show up. To this end, 6 different scenarios were simulated and data on the gaze of the drivers were analyzed using topic modeling. Their driving styles (topics) latent in the driver’s driving behaviors (words) following a driving scenario (document) were analyzed by using the latent dirichlet allocation of topic modeling, the most frequently used in discovering latent topics in documents generally made up of words. For the study, six participants in their twenties were selected whose driver licenses were more than a year old. They were asked to drive in a virtual reality simulator, while wearing a head mounted display capable of tracking their gazes. The experimental results showed that the less experienced the drivers were, the more frequently and longer they gazed at the navigation and the speed instrument panel and repeated the start and stop. On the other hand, the more experienced the drivers were, the more they gazed briefly at the objects within the car, maintained speed after glancing at the most distant objects, and applied braking only when necessary.
Download
Paper Nr: 65
Title:

Federated Health Recommender System
Authors:

Sarah Pinon, Simon Jacquet, Colin V. Bulcke, Edouard Chatzopoulos, Xavier Lessage and Raphaël Michel
Abstract: Precision Medicine is a new and growing approach to health care. This initiative includes different patient-oriented Decision Support Systems (DSS), such as Health Recommender Systems (HRS). These patient-oriented DSS aim to increase the accuracy and personalization of health care. However, the development of these systems faces a major obstacle related to the confidential and private nature of medical data. These systems require, indeed, a large volume of data to run effectively. But medical data are dispersed among several institutions and cannot be centralized for strict confidentiality reasons. To address this issue, this position paper proposes a system’s architecture in which Federated Learning is exploited to build a HRS. Federated Learning allows exploiting the data maintained by different institutions to build the system without requiring their sharing. To demonstrate the feasibility of our proposition, we build a Federated Drug Recommender System. The goal of the system is to assist doctors in their administration of drugs by using historical disease-drug interactions and drug data. As a position paper, the objective of this use case is limited to a proof of concept realized on non-sensitive open-source data. Our ambition is then to use the architecture proposed in this paper to develop a Federated HRS on real medical data.
Download
Paper Nr: 67
Title:

Methods to Estimate Respiratory Rate Using the Photoplethysmography Signal
Authors:

Ayalon D. Moraes Filho, Guilherme Schreiber, Julio A. Sieg, Maicon D. Much, Vanessa M. Bartoski and César Marcon
Abstract: Academia and industry have devoted significant effort to the research and development of smart wearable devices applied to health monitoring. The photoplethysmography (PPG) sensor is widely used for monitoring biosignals, such as heart and respiratory rate (RR), which are influenced by the cardiovascular system. This work focuses on analyzing methods for RR estimation regarding the effect of breathing on the PPG signal variation. This work describes, implements, and analyzes four methods for estimating RR. These methods are based on capturing RR using Fast Fourier Transform, median, and extracting physiological characteristics induced by respiration in the PPG signal. The most efficient method merges three RR calculations analyzed on the same signal, achieving nearly 93% of efficacy in the best scenario. The method efficacies were calculated using PPG signals from the BIDMC and CapnoBase databases collected from patients during hospital care. The analysis allows for understanding and mitigating the RR estimation challenges and evaluating the most efficacy method for a wearable device monitoring scenario.
Download
Paper Nr: 69
Title:

Fit-Twin: A Digital Twin of a User with Wearables and Context as Input for Health Promotion
Authors:

Muhammad Sulaiman, Anne Håkansson and Randi Karlsen
Abstract: Digital health contributes to health promotion by empowering the user with the holistic view of their health. Health promotion is to enable the user to take control over their health. The availability of wearables has contributed to the shift in healthcare, that is more connected, predictive, and proactive. Proactive in healthcare is to predict and prevent a situation, beforehand. This shift in healthcare puts the user in charge of most health-related decisions. Innovative technologies like AI already contribute to the cause by applying reasoning and negotiation to the collected health data to provide timely interventions to the user. The availability of real-time data from sensors that the user wears all the time allows more opportunities with new health insights. One such prospect is the use of digital twins, which provides personalization and precision. Digital twins also allow risk-free modelling for more accurate outcomes. A user digital twin is not just a virtual replica, but it combines all the factors that can impact the user. The context of the user is a prominent factor in healthcare. The paper establishes the need for digital twins in health promotion. In this paper, a Fit-twin is presented that mimics a user with wearables and the user context as input. The Fit-twin is implemented using Azure digital twins, Fitbit charge, and local context API. This allows one-way communication between the user and the Fit-twin. The outcome is a user digital twin that can be used for health promotion by applying predictive capabilities.
Download
Paper Nr: 70
Title:

Perceptions on Telemedicine in Portugal During Sars-Cov-2 Pandemic: A Mixed-Methods Study
Authors:

Ana Dias, Sandra Duarte, Joaquim Alvarelhão and Conceição Cunha
Abstract: This study aimed to investigate how patients and professionals faced telemedicine or telehealth in Centre Region in Portugal during the Sars-Cov-2 pandemic. Mixed-methods exploratory and parallel study including data from a survey of 190 healthcare patients and seven qualitative interviews with healthcare professionals from the Centre Region of Portugal were carried out. Descriptive and multiple correspondence analysis was used for survey results evaluation while healthcare professionals' perceptions were studied using a thematic analysis approach. Although few participants (15%) experienced telemedicine before the pandemic, most (73.2%) consider the health sector prepared to provide it. The most mentioned benefits of telemedicine were the avoidance of travel, convenience, and comfort for the patient. The limitations that may exist in this modality relate to patients who do not have the necessary technological devices, the lack of adequate diagnostic tools, and limitations to the patient-doctor relationship. Younger participants (<30y) were associated with characteristics of the telemedicine operating system, like the adequacy of diagnostic tools while persons more than 50 years old were associated with the lack of preparation or predisposition of professionals to provide telemedicine.
Download
Paper Nr: 71
Title:

A Survey on Technologies Used During out of Hospital Cardiac Arrest
Authors:

Gaurav Rao, David W. Savage, Vijay Mago and Pawan Lingras
Abstract: Out of hospital cardiac arrest (OHCA) causes close to 400,000 deaths every year in North America, and it is also a leading cause of death among young athletes. OHCA is a treatable medical condition, and the patient’s survival chances can be increased if immediate treatment is provided to the patient. However, nontreatment of the patient leads to a dramatic decline in survival chances at 10% per minute. Currently, various technologies are being used, and many more are being researched to reduce the time to provide early treatment to the patient. This survey focuses on summarizing various available technologies for use during OHCA. This survey focuses on evaluating technologies used in each step of the OHCA process. In this survey, articles were searched using the term “ohca” on Google Scholar and more than 18,000 articles were found. The articles were further filtered using keywords for each stage of the OHCA process, finally, 112 articles were used in this survey. The technologies that exist today work independently and are not linked with the other steps of the OHCA process. Integration between these technologies could help in reducing time and increase the survival chances of the patient.
Download
Paper Nr: 73
Title:

Development of Learning System to Support for Passing Steps of Wheelchair
Authors:

Kotone Sakiyama, Yukie Majima and Seiko Masuda
Abstract: Recent aging of the population has led to an increased number of persons requiring assistance and a shortage of caregivers. Wheelchairs are often used for transportation by people who require assistance, but they must use appropriate operating techniques because they can easily impose burdens on caregivers when climbing over steps. Therefore, for this study, an educational system was developed based on issues elucidated by conventional educational methods and earlier research. Assistive technology evaluation in this system is performed from the perspective of a passenger’s riding comfort by measuring and analyzing the wheelchair’s degree of tilt and vibration level using a sensor. The system provides a learner with feedback for adopting an appropriate operating posture based on the evaluation results. This system can engender efficient learning by quantitatively measuring and presenting the learner’s level of proficiency and by providing immediate feedback according to the user’s proficiency level.
Download
Paper Nr: 75
Title:

Effectiveness of Reflexology for Premenstrual Syndrome (PMS) and Other Health Difficulties Specific to Women
Authors:

Ayame Inoue, Yukie Majima, Seiko Masuda and Takeshi Matsuda
Abstract: Reflexology is said to be effective for regulating physical discomfort and relieving pain, but it is also expected to be effective for improving PMS symptoms, which are common among many women. However, few researchers have examined its effects for PMS symptoms, so few women use reflexology and most of them still rely on drugs such as oral contraceptives and chemical therapies. For this study, we examined reflexology effects on PMS symptoms and the body, devoting particular attention to bowel sounds. This paper presents results of experiments which treat reflexology for three women who have each PMS symptoms and other health difficulties. As results of bowel sounds, frequency spectrums of three women have been consolidated into a single curve. one after treatment. In the future, we increase the number of data and clarify the relation among reflexology, bowel sounds, PMS symptoms and other health difficulties.
Download
Paper Nr: 78
Title:

Improvement Potential of Business Models and Usability of Fitness Apps: Results of Expert Interviews and a User Study
Authors:

Leon Griesch and Kurt Sandkuhl
Abstract: Mobile applications (apps) supporting individual users in their personal fitness have received growing interest on the user side and in the research community. The value offered to fitness app users in general consists of support for changing attitude and behaviour towards a healthy lifestyle and maintaining this behaviour. Work presented in this paper aims at contributing to this field by focusing on the following main research question: what improvement potential exists for fitness apps that at the same time are contributing to the business model and desired from a user’s perspective? The contributions of this paper are (1) market and functionality analysis of fitness apps, (2) results of expert interviews on functionality and scenarios with business model improvement potential, and (3) results of a user study on desired improvements from a usability perspective.
Download
Paper Nr: 79
Title:

Contrast Set Mining for Actionable Insights into Associations Between Sleep and Glucose in a Normoglycemic Population
Authors:

Hoang H. Nhung and Zilu Liang
Abstract: Prior studies have suggested potential associations between poor sleep and glucose dysregulation among diabetic patients. However, little is known about the relationship between sleep and glucose regulation in healthy populations. In this study, we proposed a data mining pipeline based on contrast set mining to identify significant associations between sleep and glucose in a dataset collected from a normoglycemic population in free-living environments. Unlike traditional correlation analysis, our approach does not assume a linear relationship between sleep and glucose and can potentially discover associations when a pair of metrics fall within certain value ranges. The data mining result highlights the total sleep time as an important sleep metric associated with glucose regulation the next day, which is characterised by rules with high lift and confidence. Furthermore, the result suggests that having a higher time ratio in normal glucose range was associated with better sleep continuity at night. These results may provide insights that people can immediately act on for better sleep and better glucose control. Future research may leverage the proposed data mining protocol to develop healthy behaviour recommender systems.
Download
Paper Nr: 84
Title:

Deep Learning Assisted Plus Disease Screening of Retinal Image of Infants
Authors:

Vijay Kumar, Vatsal Agrawal, Shorya Azad and Kolin Paul
Abstract: Retinopathy of Prematurity (ROP) is the leading cause of blindness in premature infants in developing countries. The international classification of ROP (ICROP) classifies ROP based on location, severity, and stage. Plus disease is a crucial feature in ROP classification. Plus is the most severe form of vascular dilatation and tortuosity, and it causes severe ROP and visual loss if untreated. Despite decades of research, identifying and quantifying Plus diseases is challenging. Understanding and detecting Plus in ROP patients can help ophthalmologists provide better treatment, restoring vision to many infants with severe ROP. Hence, we have proposed a robust Deep Learning-assisted framework for Blood Vessels map generation and analysis that may effectively address the issue related to Plus disease screening and monitoring. We have extensively studied various methods for computing and locating different blood vessel map features such as vessel branch point, vessel width, vessel skeleton/centre-line, vessel segment tortuosity, etc. Additionally, we divided the branches into two levels based on the width of the branches. For our investigations, we have used both local and public databases. This work also includes a detailed analysis of these datasets’ vascular feature and their level. To the best of our knowledge, none of the publicly available models could independently classify branches and/or analyse the tortuousness based on the parent and child relationship of branches. For Plus, pre-Plus, and Healthy infants, the average tortuosity index is 1.959, 1.1530, and 1.126, and the percentage of vessels severely infected is 44%, 30%, and 20%, respectively. Moreover, our algorithm recognises and analyses many vessels. The precision of many parameters is remarkable.
Download
Paper Nr: 88
Title:

A Protocol to Assess Usability and Feasibility of e-SpeechT, a Web-based System Supporting Speech Therapies
Authors:

Vita S. Barletta, Miriana Calvano, Antonio Curci and Antonio Piccinno
Abstract: Speech disorders and impairments are a heterogeneous group of dysfunctions that concern linguistic abilities and, in the majority of cases, they affect people during their childhood from 4 to 6 years old. Appropriate treatment should be defined for each patient according to their problems, which can be physical, such as muscle weakness, brain damage, vocal cord damage, or paralysis, and psychological, such as autism, PTSD or Down Syndrome. Therefore, considering these aspects, this research work aims at supporting patients in carrying out therapies to solve speech impairments through gamification, game-based learning, e-health, and telemedicine. A web-based system is proposed to achieve the goals by supporting all stakeholders involved in speech therapy. The design and plan of a usability and a longitudinal study have been presented, with the goal of testing the system’s usability and medical efficacy by involving real end-users and domain experts.
Download
Paper Nr: 90
Title:

A Multi-Modality Approach to Medical Case Retrieval for Alzheimer's Disease
Authors:

Katarina T. Dineva, Ivan Kitanovski, Ivica Dimitrovski, Suzana Loshkovska and Alzheimer’s Disease Neuroimaging Initiative
Abstract: In this research, we evaluate medical case retrieval for AD on the bases of descriptors generated by combining different modalities (Magnetic Resonance Imaging (MRI) markers, Fluorodeoxy-glucose Positron Emission Tomography (FDG-PET) based measures, Cerebrospinal Fluid (CSF) protein levels, and Apolipoprotein-E (APOE) genotype and age as risk factors). We investigated whether they would provide complementary information aiming to improve medical case retrieval for AD. According to the obtained results, we concluded that this approach outperformed the retrieval results in the current reported research by gaining MAP value of 0.98 yet providing an efficient medical case retrieval for AD and keeping low dimensional feature vector.
Download
Paper Nr: 7
Title:

Artificial Intelligence Enabled Healthcare Ecosystem Model: AIEHEM Project
Authors:

Luigi Lella, Ignazio Licata and Christian Pristipino
Abstract: The AIEHEM project aims to analyze the data made available by the regional health system, using an unorganized Turing machine model (A-Type) trained with a swarm-evolutionary hybrid algorithm. The goal is to identify the main factors related to certain outcomes that the healthcare organization intends to achieve (which can be economic, organizational, social or environmental). The chosen AI model is used to enhance, not to replace the analytical capabilities of the healthcare system management. The insights of the AI model are in fact used not only to identify the main objects of study to be taken into consideration, but also to define the areas of intervention and consequently also the stakeholders to be involved in the organizational change project to be carried out through the Theory of Change methodology. AI is therefore used to identify the most suitable ecosystem for solving the considered problem.
Download
Paper Nr: 11
Title:

Launcher50+: An Android Launcher for Use by Older Adults
Authors:

Craig Leamy, Bilal Ahmad, Sarah Beecham, Ita Richardson and Katie Crowley
Abstract: While software is becoming increasingly ubiquitous, needed, and available, applications developed for use by older adults do not always take their specific requirements into account. Our thesis is that: implementing Usability and Accessibility smartphone application requirements which cater for Older Adults’ specific needs will improve their ability to engage with software. Our prior research developed 44 Recommendations for the Development of Smartphone Applications for the Ageing Population (ReDEAP). We assessed 5 existing launchers against relevant ReDEAP recommendations, finding that they had 48-64% compliance. To test the feasibility of ReDEAP, we implemented a subset of recommendations, developing a new launcher app, Launcher50+. The recommendations supported the implementation of a simple user interface, addressing a key concern raised in the wider ReDEAP study. In this paper, we also evaluate other smartphone launcher applications, assessing the level of ReDEAP recommendation compliance. This identified several weaknesses, suggesting that catering for the needs of older adults can be improved.
Download
Paper Nr: 12
Title:

Development and Application of Regional Level Complete Inspection Management Platform
Authors:

Jin Zhao, Guocheng Wang, Daguo Huang, Yue Teng, Yichu Bai, Xudong Gao and Yi Zhou
Abstract: Objective: To solve the problems associated with the multi-sectoral management of key populations in the regular fight against the epidemic, a platform for the management of due diligence was developed. Methods: an inspection due diligence management platform based on overall management, fine management, and process optimization was established to ensure the effective and scientific development of regional inspection due diligence. Results: The platform provided a good system to support the regional inspection work and solved the four major problems in management, including reporting process, data standard, data display, and data sharing. Conclusion: In normal epidemic prevention and control, the inspection management platform should be irreplaceable in department supervision, process optimization, data sharing, data accuracy, and other aspects.
Download
Paper Nr: 33
Title:

Technical Realization and First Insights of the Multicenter Integrative Breast Cancer Registry INTREST
Authors:

Thomas Ostermann, Sebastian Unger, Michaela Warzecha, Sebastian Appelbaum, Daniela R. Recchia, Holger Cramer and Heidemarie Haller
Abstract: Cancer is one of leading causes of mortality worldwide. According to GLOBOCAN database, 19.3 million new cancer cases and 10 million cancer deaths worldwide were counted in 2020. Thus, there is an absolute necessity for statistical data on cancer incidence and treatments. This is mainly done by cancer registries, which aim at collecting, managing, and analyzing health and demographic data on individuals diagnosed with cancer. As more and more patients make use of integrative oncology to optimize their health and quality of life during and after cancer treatment, it is important to gather clinical registry data of complementary as well as conventional cancer care. The INTREST registry is the first approach that aims to identify predictors of treatment-response in women undergoing individualized, integrative breast cancer treatment. This article reports on the technical realization and representativity of the registry based on 3,341 eligible women and 885 cases included in interim statistical analysis. The analyses show that the INTREST sample of women suffering from breast cancer does not significantly differ from population-based registries and pragmatic trial data of breast cancer patients in Germany with respect to main sociodemographic and clinical cancer data. However, completeness, particularly in tumor classification, currently is a major limitation.
Download
Paper Nr: 50
Title:

Pharmaceutical Audit Trail Blockchain-Based Microservice
Authors:

Stefano Loss, Lucas Cardoso, Nélio Cacho and Frederico Lopes
Abstract: Pharmaceutical manufacturing in Brazil requires that its processes are carried out by following rules defined by a supervisory body: the National Health Surveillance Agency (ANVISA, in Portuguese). ANVISA requires that all pharmaceutical systems guarantee all product information’s integrity, security, and traceability. These rules ensure that the manufactured products do not pose a risk to their consumers. One of the difficulties for pharmaceutical industries is to provide evidence that production procedures were carried out under internal regulations based on these rules. One way to do this is by using an audit trail. It can store this information automatically using a computer system to record all actions. However, only using audit trails does not guarantee data security; ensuring that all information is immutable is necessary. Therefore, in this paper, we propose an audit trail blockchain-based microservice. This technology stores all transactions in linked and encrypted blocks to avoid illegal modifications. It also guarantees data immutability, security, and traceability. In addition, we present a case study to evaluate the proposed approach using Nuplam’s (Nucleus for Research in Food and Medicines) Integrated Management Systems. A stress test was performed in this case study to evaluate the applicability of the proposed solution in pharmaceutical systems.
Download
Paper Nr: 53
Title:

Bed Management System Development
Authors:

Flannagán Noonan, Juncal Nogales, Ciarán Doyle, Eilish Broderick and Joseph Walsh
Abstract: The costs of supporting hospitals are rising, bed numbers are falling and a growing population living longer will require more hospital visits over their lifetime. Thus there is a global focus on increasing the efficiency of patient throughput in a hospital. Bed management systems are still commonly paper-based and are effectively memory-less from the hospital point of view. The hospital information systems are typically billing and ordering systems with minimal information on patient movement along the patient pathway. The literature suggests that technology and shared information allow for shared views to model and predict usage to better manage finite resources. Paper-based systems work against this. This paper presents the design considerations for a bed management application developed in conjunction with a local private hospital. The application developed, provides a hospital-wide view of patient and bed status by recording and capturing touchpoints, that is patient-hospital interactions. Furthermore, it captures data electronically such that the data can be used for analysing patient presentation and bed moving with a view to improve bed management and patient throughput.
Download
Paper Nr: 55
Title:

Intelligent Provision of Tailored, Easily Understood, and Trusted Health Information for Patient Empowerment
Authors:

Marco Alfano, John Kellett, Biagio Lenzitti and Markus Helfert
Abstract: Although digital transformation in healthcare is accelerating, there is still a disconnect between current healthcare, focused on disease management, and a more holistic approach that looks at the health and wellbeing of the whole person. The latter approach aims at empowering patients and other health information seekers by improving their comprehension of their health so that they can manage it better. Currently, few stand-alone applications for patient empowerment exist and they seldom help users to understand health information. Thus, health information seekers often interact with the Web through generic search engines, which often produce results that are overwhelming, too generic, and of poor quality. This paper shows how the use of Artificial Intelligence (AI) in a responsible way may provide patients and others with health higher quality information that empowers them to improve their health and wellbeing. It presents an AI engine that extracts health content from the Web and provides the user with health information that is relevant, trustworthy, and easy to understand. The AI engine has been used to create an Intelligent Empowering Agent (IEA) that dialogues with users in simple language to provide customised information on symptoms and diseases, which helps them form their own evidence-based opinion on health matters that concern them.
Download
Paper Nr: 57
Title:

The Use of Machine Learning to Predict Hospitalization of Covid-19: A Case Study in the State of Minas Gerais - Brazil
Authors:

Gerda G. Rodrigues de Oliveira and Cristiane N. Nobre
Abstract: This work aims to verify the applicability of using Machine Learning techniques to predict hospitalization in confirmed cases of Covid-19. The study also intends to discover which attributes have the most significant impacts on hospitalization. The machine learning (ML) algorithms used in this experiment were Decision Tree, Random Forest, Neural Networks, and Naive Bayes. The data used for this experiment were made available by the government of Minas Gerais - Brazil, through open data. The model based on Random Forest obtained the best results, presenting the following metrics: Precision, Recall and F1-Score of 0.85, 0.84 and 0.84, respectively. In this experiment, essential characteristics for classifying the patient’s hospitalization are Comorbidity, Age Group, and HDI. The results point to a good predictive ability, demonstrating the potential use of ML techniques to predict the hospitalization of people by COVID-19.
Download
Paper Nr: 59
Title:

Physiological Data Recording in VR Simulator for Sleepiness Detection During Driving
Authors:

Baptiste Chevallier, Dan Istrate, Vincent Zalc, Nicolas Vera and Christophe Labrousse
Abstract: Drowsy driving is a major issue in road safety. In this paper, we propose a description of an experimental data collection to develop a drowsiness detection model. The objective of this data collection was mainly to gather physiological data of individuals in simulated driving situations. We designed a realistically annoying scenario to induce fatigue while staying close to real driving conditions. The experiment was run on an augmented reality platform called CAVE. The need for contextualization came early in the design of the experiment. Therefore, in addition to physiological data, we added much more data sources, from driving habits to driving behaviour in addition to self-assessment of fatigue levels and the gold standard (EEG). As a result, this experience helped us create a data set of physiological data completed by elements of context and driving behaviour. Thus allowing us to perform a very rich analysis of these physiological data.
Download
Paper Nr: 62
Title:

GASTon: A Graph-Exploration System for Indexing, Annotating and Visualizing PubMed Articles to Enhance the Analysis of Social deTerminants of Health
Authors:

Simone Bottoni, Alberto Trombetta, Flavio Bertini, Danilo Montesi, Francesca Bonin, Alessandra Pascale, Martin Gleize and Pierpaolo Tommasi
Abstract: Many works have shown associations between social determinants of health (SDoH) –the social circumstances in which people live– and health-related outcomes. However, the lack of SDoH data increases the challenges in measuring and understanding their effect on people’s health and health systems. In this paper, we present GASTon, a system for the indexing, annotation, and graph-based rendering of PubMed information to enable the search and retrieval of SDoHs in scientific literature. Our work provides a way to associate specific concepts with peer-reviewed articles to simplify the search for social factors. It builds a knowledge graph based on PubMed publications and associates them with concepts extracted from the Unified Medical Language System (UMLS) Metathesaurus. GASTon allows a full-text search and graph-based navigation and supports an overview of the concepts and related publications. Moreover, the architecture allows scale-up thanks to its containerized nature and parallelization capabilities. The system is open-source under the Apache V2 license.
Download
Paper Nr: 68
Title:

LifeSeniorProfile: A Multisensor Dataset for Elderly Real-time Activity Track
Authors:

Maicon D. Much, Julio A. Sieg, Ayalon D. Moraes Filho, Vanessa M. Bartoski, Guilherme Schreiber and César Marcon
Abstract: Real-time tracking and detection of risky situations in the elderly, such as falls and sudden changes in vital signs, requires reliable, continuous, and automated monitoring systems based on relevant information. Wireless biosensors provide a great opportunity to remotely detect and monitor hazardous situations, allowing for a fast response in an emergency. Motion data is widely used to track daily activities. Physiological data can also be used for this exact purpose. However, there is yet to be a database available in the field of research in which the patient’s physiological and movement information were collected simultaneously, considering daily activities and simulation of falls. This work presents a multisensor dataset for developing real-time tracking systems for the daily activities of older people. The data sensed refer to movement, using a triaxial accelerometer, and physiology, considering blood volume pulse, electrodermal activity, heart rate, inter-beat interval, and skin temperature. We collected these data from ten volunteers while performing 36 daily activities in a simulated environment.
Download
Paper Nr: 72
Title:

Camera-Based Tracking and Evaluation of the Performance of a Fitness Exercise
Authors:

Linda Büker, Dennis Bussenius, Eva Schobert, Andreas Hein and Sandra Hellmers
Abstract: Back pain is a significant condition worldwide that is becoming more common over the years. Although individually adapted exercise therapy would be very successful, the problem is - due to lack of capacity it is rarely prescribed by physicians. We analyse if it is feasible to automatically track and evaluate the performance of an exercise used in a fitness check to support physicians in adapting exercise therapy and thereby providing that kind of treatment to more patients. A depth camera with body tracking is used to detect the pose of the subjects. We have developed a system that evaluates the execution of an exercise in terms of correct performance based on the recognized joint positions. The feasibility study conducted shows, that it is important to avoid as many occlusions of any kind as possible to get the best achievable body tracking. Then, however, the evaluation of the performance of the tested finger-floor-distance exercise seems feasible.
Download
Paper Nr: 77
Title:

A Profile Recognition System Based on Emotions for Children with ASD in an Interactive Museum Visit
Authors:

Nicolás Araya, Javier Gomez and Germán Montoro
Abstract: Children with Autism Spectrum Disorder (ASD) may experience difficulties in doing some activities on their own as they are likely to be more sensitive to visual or auditory stimuli. This may limit their selection of activities, and these must be adapted for them and their companions. One of the most important issues for this group is to be able to manage and recognise their emotions, which leads to a better understanding of themselves and their surroundings. In recent years, information technology has helped to develop assistance tools for education and daily habits. However, research in emotional management in children with ASD has not been fully explored for leisure and cultural activities. In this paper we present a proposal for a user model and a mobile application intended to assist children with ASD when visiting a leisure space and assess the emotional impact as they go through the different attractions. Users will respond to a questionnaire based on the basic emotions in an itinerary suitable for their general behaviour. The methodology is validated by a non-profit organisation, who helped to create a case study, intended to provide guidance for recommendations on leisure activities for these children and their caregivers.
Download
Paper Nr: 80
Title:

Automated Identification of Yellow Flags and Their Signal Terms in Physiotherapeutic Consultation Transcripts
Authors:

Joep Wegstapel, Thymen den Hartog, Mick Sneekes, Bart Staal, Ellis van der Scheer-Horst, Sandra van Dulmen and Sjaak Brinkkemper
Abstract: This paper investigates the possibility of automating the process of identifying yellow flags and their signal terms in physiotherapeutic consultation transcripts from patients with low back pain, using Automated Text Identification. It is part of the Automated Medical Reporting research domain. In physiotherapy focused on low back pain, yellow flags are considered psycho-social predictors of poor recovery and risk factors for chronic disability development. This paper uses a 6-step mixed method approach. Consultation transcripts and yellow flag assessment guidelines were collected, an automated identification tool was built and the OSPRO assessment guideline was used to test the tool for accuracy. It was found that it is possible to identify Yellow Flags and their Signal Terms automatically with the tool developed in this experiment. However, this is just the beginning, and much more research must be done in the future to further enhance the tool, mainly to improve precision.
Download