HEALTHINF 2019 Abstracts


Full Papers
Paper Nr: 18
Title:

Integrating Physical Activity Data with Electronic Health Record

Authors:

Rishi Saripalle

Abstract: Wearables allow individuals to track, analyze, and visualize their physical activities and associated data such as vitals, activity information, etc. across time. But, none of this activity data is anywhere to be found in an electronic health record - the primary source of patient medical data for the healthcare providers. This inability doesn’t allow experts to view the complete health summary of an individual and also, activity data can play a key role in healthcare decisions. This problem is due to the lack of standards that can capture activity data from disparate sources (e.g., wearables, smart watches, trackers, etc.) and integrate it with an EHR. This research article identifies and provides a detailed analysis of the key factors contributing to the problem. Based on the detailed analysis, we design an interoperable model by leveraging HL7 FHIR standard to capture activity data from wearables and develop it using FHIR HAPI - an implementation of HL7 FHIR. This initial prototype is tested by capturing Fitbit data and integrating it with OpenEMR - an open source EHR.

Paper Nr: 22
Title:

How Anthropomorphism Affects User Acceptance of a Robot Trainer in Physical Rehabilitation

Authors:

Baisong Liu, Panos Markopoulos and Daniel Tetteroo

Abstract: Developments in social robotics raise the prospect of robots coaching and interacting with patient during rehabilitation training assuming a role of a trainer. This raises questions regarding the acceptance of robots in this role and more specifically, to what extent the robot should be anthropomorphic. This paper presents the results of an online experiment designed to evaluate the user acceptance of Socially Assistive Robots (SARs) as rehabilitation trainers, and the effect of anthropomorphism on this matter. User attitudes were surveyed with regards to three variations of a scenario where the robot with varying levels of anthropomorphism acts as a trainer. The results show that 1) participants are accepting towards SAR-assisted rehabilitation therapies, 2) anthropomorphism influences patient’s perceived self-efficacy and attitude towards the system. A second survey studied inventoried issues regarding patients’ acceptance of such systems, pertaining to technology acceptance, patient needs for rehabilitation training and the effect of anthropomorphism. Based on the above findings we propose user-informed design implications for improving user acceptance is rehabilitation settings.

Paper Nr: 23
Title:

Supporting Reuse of EHR Data in Healthcare Organizations: The CARED Research Infrastructure Framework

Authors:

Vincent Menger, Marco Spruit, Jonathan de Bruin, Thomas Kelder and Floor Scheepers

Abstract: Healthcare organizations have in recent years started assembling their Electronic Health Record (EHR) data in data repositories to unlock their value using data analysis techniques. There are however a number of technical, organizational and ethical challenges that should be considered when reusing EHR data, which infrastructure technology consisting of appropriate software and hardware components can address. In a case study in the University Medical Center Utrecht (UMCU) in the Netherlands, we identified nine requirements of a modern technical infrastructure for reusing EHR data: (1) integrate data sources, (2) preprocess data, (3) store data, (4) support collaboration and documentation, (5) support various software and tooling packages, (6) enhance repeatability, (7) enhance privacy and security, (8) automate data process and (9) support analysis applications. We propose the CApable Reuse of EHR Data (CARED) framework for infrastructure that addresses these requirements, which consists of five consecutive data processing layers, and a control layer that governs the data processing. We then evaluate the framework with respect to the requirements, and finally describe its successful implementation in the Psychiatry Department of the UMCU along with three analysis cases. Our CARED research infrastructure framework can support healthcare organizations that aim to successfully reuse their EHR data.

Paper Nr: 25
Title:

The Impact of Environmental Factors on Heart Failure Decompensations

Authors:

Garazi Artola, Nekane Larburu, Roberto Álvarez, Vanessa Escolar, Ainara Lozano, Benjamin Juez and Jon Kerexeta

Abstract: Heart failure (HF) is defined as the incapacity of the heart to pump sufficiently to maintain blood flow to meet the body's needs. Often, this causes sudden worsening of the signs and symptoms of heart failure (decompensations), which may lead on hospital admissions, deteriorating patients’ quality of life and causing an increment on the healthcare cost. Environmental exposure is an important but underappreciated risk factor contributing to the development and severity of cardiovascular diseases, such as HF. In this paper, we describe the development and results of a methodology to determine the effect of environmental factors on HF decompensations by means of hospital admissions. For that, a total number of 8338 hospitalizations of 5343 different patients, and weather and air quality information from open databases have been considered. The results demonstrate that several environmental factors, such as weather temperature, have an impact on the HF related hospital admissions rate, and hence, on HF decompensations and patientś quality of life. The next steps are first to predict the number of hospital admissions based on the presented study, and second, the inclusion of these environmental factors on predictive models to assess the risk of decompensation of an ambulatory patient in real time.

Paper Nr: 26
Title:

Rule-based and Machine Learning Hybrid System for Patient Cohort Selection

Authors:

Rui Antunes, João F. Silva, Arnaldo Pereira and Sérgio Matos

Abstract: Clinical trials play a critical role in medical studies. However, identifying and selecting cohorts for such trials can be a troublesome task since patients must match a set of complex pre-determined criteria. Patient selection requires a manual analysis of clinical narratives in patients’ records, which is a time-consuming task for medical researchers. In this work, natural language processing (NLP) techniques were used to perform automatic patient cohort selection. The approach herein presented was developed and tested on the 2018 n2c2 Track 1 Shared-Task dataset where each patient record is annotated with 13 selection criteria. The resulting hybrid approach is based on heuristics and machine learning and attained a micro-average and macro-average F1-score of 0.8844 and 0.7271, respectively, in the n2c2 test set. Part of the source code resultant from this work is available at https://github.com/ruiantunes/2018-n2c2-track-1/.

Paper Nr: 27
Title:

Mobile Apps for People with Dementia: Are They Compliant with the General Data Protection Regulation (GDPR)?

Authors:

Joana Muchagata and Ana Ferreira

Abstract: Mobile apps have the potential to improve the overall patients and caregivers’ quality of life and, particularly, of those with dementia. The ability to stimulate cognitive functions, keep the brain active and helping people to be as independent as possible in their daily lives are considered highly valued characteristics. But despite those advantages, there is a lack of security standards and guidelines focused on mobile apps and the general sense is that those provide low or no privacy/security and commonly do not comply with current regulations. We analysed eighteen apps with the ability to stimulate cognitive functions for people with dementia to verify if they were GDPR compliant. Results show that most analysed apps (78%) do not provide any information regarding how personal data are processed, and if they do, this is not clear. Also, users’ consent to allow that processing is rarely sought (11%). In conclusion, GDPR mandated requirements are still not implemented in most of the analysed mental health apps to ensure privacy and security in the interactions between users and mobile apps. This work intends to bring awareness to this issue to both researchers and developers, especially in the area of healthcare and mental health.

Paper Nr: 30
Title:

Text-based Medical Image Retrieval using Convolutional Neural Network and Specific Medical Features

Authors:

Nada Souissi, Hajer Ayadi and Mouna Torjmen-Khemakhem

Abstract: With the proliferation of digital imaging data in hospitals, the amount of medical images is increasing rapidly. Thus, the need for efficient retrieval systems, to find relevant information from large medical datasets, becomes high. The Convolutional Neural Network (CNN)-based models have been proved to be effective in several areas including, for example, medical image retrieval. Moreover, the Text-Based Image Retrieval (TBIR) was successful in retrieving images with textual description. However, in TBIR, all queries and documents are processed without taking into account the influence of certain medical terminologies (Specific Medical Features (SMF)) on the retrieval performance. In this paper, we propose a re-ranking method using the CNN and the SMF for text-medical image retrieval. First, images (documents) and queries are indexed to specific medical image features. Second, the Word2vec tool is used to construct feature vectors for both documents and queries. These vectors are then integrated into a neural network process and a matching function is used to re-rank documents obtained initially by a classical retrieval model. To evaluate our approach, several experiments are carried out with Medical ImageCLEF datasets from 2009 to 2012. Results show that our proposed approach significantly enhances image retrieval performance compared to several state of the art models.

Paper Nr: 31
Title:

Digital Picture Co-occurrence Texture Characteristics Discriminate between Patients with Early Dementia of Alzheimer’s Type and Cognitive Healthy Subjects

Authors:

Sibylle Robens, Thomas Ostermann, Sebastian Unger, Petra Heymann, Stephan Müller, Christoph Laske and Ulrich Elbing

Abstract: Gray level co-occurrence texture characteristics of digital drawings were compared between persons with early dementia of Alzheimer’s disease and healthy controls. It was hypothesized that texture characteristics contribute to the differentiation between these subject groups. The study population consisted of 67 healthy subjects and 56 patients with early dementia of Alzheimer’s type. Between subject groups comparisons of texture entropy, homogeneity, correlation and image size were conducted with Mann-Whitney-U tests. The diagnostic power of combining all texture features as explanatory variables was analysed with a logistic regression model and the area under curve (AUC) of the corresponding receiver operating control (ROC) curve was calculated. The gray level co-occurrence characteristics differed significantly between healthy and demented subjects and the logistic regression model resulted in an AUC of 0.86 (95% CI [0.80, 0.93], sensitivity=.80, specificity=.79).

Paper Nr: 53
Title:

Usability Study of a Tool for Patients’ Access Control to Their Health Data

Authors:

Sandra Reis, Ana Ferreira, Pedro Vieira-Marques and Ricardo Cruz-Correia

Abstract: Considering the need that is expected for the institutions to be compliant with the new legislation we intended to create a tool (MyRegister) that will allow patients to know which professionals accessed their health data in healthcare institutions where they have had previous contact. Before development we decided to create a functional prototype to validate the user interface of the tool with real users. We created an evaluation instrument consisting of 4 tasks and a SUS questionnaire that allowed us to evaluate the MyRegister interface of the prototype by the participants. The results of the evaluation of the prototype allowed us to identify some of the major usability problems of the interface, while the SUS score of 79.5 in 100 is a result that shows good usability. Regarding the performed tasks, all were completed by the participants but not all of them answered correctly to the questions asked. After correcting the problems found and implementing the suggestions of the participants that we consider permissible to include, we intend to continue this project with the development of the tool and test its usability as well as user experience in real environments with a wider and more heterogeneous sample.

Paper Nr: 67
Title:

Learning to Predict Autism Spectrum Disorder based on the Visual Patterns of Eye-tracking Scanpaths

Authors:

Romuald Carette, Mahmoud Elbattah, Federica Cilia, Gilles Dequen, Jean-Luc Guérin and Jérôme Bosche

Abstract: Autism spectrum disorder (ASD) is a lifelong condition generally characterized by social and communication impairments. The early diagnosis of ASD is highly desirable, and there is a need for developing assistive tools to support the diagnosis process in this regard. This paper presents an approach to help with the ASD diagnosis with a particular focus on children at early stages of development. Using Machine Learning, our approach aims to learn the eye-tracking patterns of ASD. The key idea is to transform eye-tracking scanpaths into a visual representation, and hence the diagnosis can be approached as an image classification task. Our experimental results evidently demonstrated that such visual representations could simplify the prediction problem, and attained a high accuracy as well. With simple neural network models and a relatively limited dataset, our approach could realize a quite promising accuracy of classification (AUC > 0.9).

Paper Nr: 69
Title:

Subjective Sleep Quality Monitoring with the Hypnos Digital Sleep Diary: Evaluation of Usability and User Experience

Authors:

Tudor Văcărețu, Nikolaos Batalas, Begum Erten-Uyumaz, Merel van Gilst, Sebastiaan Overeem and Panos Markopoulos

Abstract: Sleep diaries are records of individual’s sleep and wake times, extending over a period of several days up to several weeks. Sleep diaries are often used to support the diagnosis and treatment of sleep disorders though the emergence of self-tracking technologies also makes them interesting for intrinsically motivated individuals who wish to gain insight into their sleep patterns and related influences. This paper introduces Hypnos, a digital sleep diary, and a user study aimed to its usability and the resulting user experience. The study involved eighteen participants without a diagnosed sleep disorder for a period of ten days. Overall Hypnos was found useful and usable which supports its application in practice, but it is advisable to make the user experience more attractive, stimulating and innovative in order to also make the self-tracking more intrinsically motivating.

Paper Nr: 79
Title:

Clinical Caremap Development: How Can Caremaps Standardise Care When They Are Not Standardised?

Authors:

Scott McLachlan, Evangelia Kyrimi, Kudakwashe Dube and Norman Fenton

Abstract: Caremaps were developed to standardise care. They have evolved from text-based descriptions to flow-based diagrams. Standardising care is seen to improve patient safety and outcomes, and to reduce the costs of providing healthcare services, but contemporary caremaps are not standardised. This research investigates contemporary caremaps and proposes a standardised model for caremap content, structure and development. The proposed model is evaluated through two case studies to create caremaps for; 1) obstetric care during labour and birth, and; 2) management and for women with gestational diabetes mellitus, finding that it is an effective method for creating standardise caremaps.

Paper Nr: 84
Title:

Developing Enterprise-wide Provider Analytics

Authors:

James P. McGlothlin, Hari Srinivasan and Ilija Stojic

Abstract: In recent years, the growth of electronic medical records for hospitals has exponentially increased the quantity of healthcare data available for analysis and performance improvement. However, the general consumption of this data by providers has still been limited to analysts and power users. Most data is delivered via static reports which serve only a single purpose. This paper describes a project to deliver a vast quantity of data in a simple and secure manner to all hospital physicians and administrative leaders. This includes clinical, operational and cost information. The delivery is with versatile and intuitive interactive dashboards which are integrated into the EMR yet come from many different sources. This allows physicians to look at their performance and compare it to their peers. Executives are able to identify improvement opportunities across the system and directors are able to identify improvement opportunities within their service. Quality and performance improvement specialists can perform data analysis without having to generate report requests and wait for delivery. This allows them to target specific initiatives and patient populations, and to tailor improvement programs to the needs of the organization. These analytics and dashboards are designed to facilitate quality improvement, efficiency, treatment standardization and cost reduction.

Paper Nr: 86
Title:

Exploring Medical Data Classification with Three-Way Decision Trees

Authors:

Andrea Campagner, Federico Cabitza and Davide Ciucci

Abstract: Uncertainty is an intrinsic component of the clinical practice, which manifests itself in a variety of different forms. Despite the growing popularity of Machine Learning–based Decision Support Systems (ML-DSS) in the clinical domain, the effects of the uncertainty that is inherent in the medical data used to train and optimize these systems remain largely under–considered in the Machine Learning community, as well as in the health informatics one. A particularly common type of uncertainty arising in the clinical decision–making process is related to the ambiguity resulting from either lack of decisive information (lack of evidence) or excess of discordant information (lack of consensus). Both types of uncertainty create the opportunity for clinicians to abstain from making a clear–cut classification of the phenomenon under observation and consideration. In this work, we study a Machine Learning model endowed with the ability to directly work with both sources of imperfect information mentioned above. In order to investigate the possible trade–off between accuracy and uncertainty given by the possibility of abstention, we performed an evaluation of the considered model, against a variety of standard Machine Learning algorithms, on a real–world clinical classification problem. We report promising results in terms of commonly used performance metrics.

Paper Nr: 92
Title:

FAIRness in Biomedical Data Discovery

Authors:

Alina Trifan and José L. Oliveira

Abstract: The FAIR Guiding Principles are a recent, yet powerful set of recommendations for turning data Findable, Accessible, Interoperable and Reusable. They were designed with the purpose of improving data quality and reusability. Over the last couple of years they have been adopted more and more by both data owners and funders as key data management approaches. Despite their increasing popularity and endorsement by multiple research initiatives from some of the most diverse areas, there are still only a few examples on how these principles have been translated into practice. In this work we propose an open evaluation of their adoption by biomedical data discovery platforms. We first overview current biomedical data discovery platforms that introduce the FAIR guiding principles as requirements of their functioning. We then employ the more recent FAIR metrics for evaluating the degree to which these biomedical data discovery platforms follow the FAIR principles. Moreover, we assess their impact on enabling data interoperability and secondary reuse.

Short Papers
Paper Nr: 11
Title:

ONKOVIS, Co-designed Oncology Follow-up Visualisation Tool

Authors:

Mikel Garmendia, Gorka Epelde, Beñat Zabala, Aurora Sucre, Arantxa Etxeberria, Gerardo Cajaraville, Jon Belloso, José F. Luengo and Soraya Camarero

Abstract: This paper presents a medical oncology follow-up visualisation tool developed following a co-design approach. The co-design and development of the solution has succeed involving a team of end-users (oncologists and nurses), UX and design experts and technology development experts. As a result, a medical oncology follow-up visualisation solution has been developed showing temporally aligned heterogeneous health data. The outcome of the co-design process has been to develop the solution in four different aggregation level views; (1) General timeline overview, (2) Medication treatment cycle view, (3) Detailed patient reported outcome / bloodwork’s’ view and (4) Medical image analysis focussed view.

Paper Nr: 13
Title:

Machine Learning based Predictions of Subjective Refractive Errors of the Human Eye

Authors:

Alexander Leube, Christian Leibig, Arne Ohlendorf and Siegfried Wahl

Abstract: The aim of this research was to demonstrate the suitability of a data-driven approach to identify the sphero-cylindrical subjective refraction. An artificial deep learning network with two hidden layers was trained to predict power vector refraction (M, J0 and J45) from 37 dimensional feature vectors (36 Zernike coefficients + pupil diameter) from a large database of 50,000 eyes. A smaller database of 460 eyes containing subjective and objective refraction from controlled experiment conditions was used to test for prediction power. Bland-Altmann analysis was performed, calculating the mean difference (eg ?M) and the 95% confidence interval (CI) between predictions and subjective refraction. Using the machine learning approach, the accuracy (?M = +0.08D) and precision (CI for ?M = ± 0.78D) for the prediction of refractive error corrections was comparable to a conventional metric (?M = +0.11D ± 0.89D) as well as the inter-examiner agreement between optometrists (?M = -0.05D ± 0.63D). To conclude, the proposed deep learning network for the prediction of refractive error corrections showed its suitability to reliably predict subjective power vectors of refraction from objective wavefront data.

Paper Nr: 15
Title:

MedClick: Last Minute Medical Appointments No-Show Management

Authors:

Inês Ferreira and André Vasconcelos

Abstract: A no-show is one of the phenomena that leads to an efficiency decrease in various sectors, including in the health care sector. When a scheduled patient misses an appointment without cancelling, it will not only waste the clinic’s resources, but it will also deny medical service to another patient who could have benefited from the respective time slot. This paper describes the research that is being developed in the context of MedClick, an online platform that aims to help medical service providers increase the efficiency of their practices. The solution supports the reduction of no-shows by predicting their occurrence and finding replacements to fulfill “last-minute” vacancy slots. A supervised learning algorithm (logistic regression) is being implemented and it will be used to predict the probability of no-show for each patient. The system will run this algorithm 48 hours before each appointment so that there is still enough time to find a replacement, if necessary. The prediction is based on features related to the respective clinic and patient, which requires access to the database.

Paper Nr: 29
Title:

Using Miniature Visualizations of Descriptive Statistics to Investigate the Quality of Electronic Health Records

Authors:

Roy A. Ruddle and Marlous S. Hall

Abstract: Descriptive statistics are typically presented as text, but that quickly becomes overwhelming when datasets contain many variables or analysts need to compare multiple datasets. Visualization offers a solution, but is rarely used apart from to show cardinalities (e.g., the % missing values) or distributions of a small set of variables. This paper describes dataset- and variable-centric designs for visualizing three categories of descriptive statistic (cardinalities, distributions and patterns), which scale to more than 100 variables, and use multiple channels to encode important semantic differences (e.g., zero vs. 1+ missing values). We evaluated our approach using large (multi-million record) primary and secondary care datasets. The miniature visualizations provided our users with a variety of important insights, including differences in character patterns that indicate data validation issues, missing values for a variable that should always be complete, and inconsistent encryption of patient identifiers. Finally, we highlight the need for research into methods of identifying anomalies in the distributions of dates in health data.

Paper Nr: 34
Title:

Successful Deployment of Cloud-hosted Services and Performance Management for Community Care

Authors:

Benjamin Eze, Craig Kuziemsky, Jamie Stevens, Paul Boissonneault and Liam Peyton

Abstract: Achieving systematic performance management of care processes across a health region requires an architecture that balances interoperability and data standardization with data governance and privacy compliance. This paper presents a case study of a successful pilot of cloud-hosted performance management for community care by a Regional Health Authority mandated with coordinating home care amongst 54 Community Support Services agencies. Cloud-hosted data services enabled data integration to a common data model. Formal data sharing agreements and privacy definition documents controlled aggregation and data masking to protect privacy while enabling accurate and comprehensive performance management services for all agencies.

Paper Nr: 37
Title:

Thermal Comfort and Stress Recognition in Office Environment

Authors:

Kizito Nkurikiyeyezu, Kana Shoji, Anna Yokokubo and Guillaume Lopez

Abstract: Work stress and thermal discomfort are some of the hurdles that office workers face every day. Office workers experience a periodic work stress because work is long and mentally challenging. At the same time, current thermal comfort provision technologies are inefficient and consume a large amount of energy. In our previous works, we proposed an efficient thermal comfort provision system that is based on a person's heart rate variability (HRV). However, because work stress can also affect the person's HRV, this paper investigates the possibility to distinguish HRV changes that are due to thermal discomfort from changes that emanate from work stress. We conducted experiments on subjects taking Advanced Trail Making Test (ATMT) and observed that stress alters HRV and that it is possible to distinguish stressed and non-stressed subjects with a 100% accuracy. We validated our method on the multimodal SWELL knowledge work (SWELL-KW) stress dataset and achieved similar results (99.25% accuracy and 99.75% average recall). Further analysis suggests that, although both thermal comfort and work stress affect HRV, their effect is perhaps non-overlapping, and that the two can be distinguished with a near-perfect accuracy. These results indicate that it could be possible to design an automatic and unobtrusive system that delivers thermal comfort and predicts work stress based on people's HRV

Paper Nr: 38
Title:

A Clustering-based Approach to Determine a Standardized Statistic for Daily Activities of Elderly Living Alone

Authors:

Alexander Gerka, Christian Lins, Max Pfingsthorn, Marco Eichelberg, Sebastian Müller, Christian Stolle and Andreas Hein

Abstract: The modeling of behavior by monitoring activities of daily living allows caregivers to recognize early stages of dementia. Therefore, many monitoring systems were presented in recent years. In this work, we present a behavior modeling system that is based only on two adjustable parameters and provides a single standardized output statistic. Therefore, this system enhances the comparison of recent and future activity monitoring systems. The approach is comprised of three parts: First, the clustering of power plug data to detect time windows in which appliances are used regularly. Second, the calculation of a comparison Matrix. Third the test of change using the χ2-statistic. We tested this approach successfully in a seven-month field study with two healthy subjects. We showed that the χ2-statistic reflected how regular activities were performed and that one to two months, depending on the regularity of the performed activities, provide the necessary amount of reference data for our approach to work.

Paper Nr: 42
Title:

Cesarean Section Avoidance based on Obstetric Hemorrhagic Risk: A Decision Support System

Authors:

Juliano S. Gaspar, Marcelo R. S. Junior, Regina P. Lopes and Zilma N. Reis

Abstract: Introduction: The junction of postpartum hemorrhage (PPH) and cesarean section (C-section) is a potential burden to take into account as a strategy to avoid unnecessary, and dangerous interventions. Despite most of the maternal death could have been prevented, rates are unacceptably high. According to the WHO, the rates of C-section are above recommended. The hypertension and PPH are the leading causes of maternal death worldwide. Aim: This study propose to analyze the association between C-section and PPH in a electronic health record (EHR) database and subsequently implementing an algorithm to assist health professionals in the avoidance of unnecessary C-section based on the estimation of obstetric hemorrhagic risk. Methods: Statistical analysis was performed using SISMater® database within 9,412 records about admissions to childbirth. The C-section rates associated with the occurrence of obstetric hemorrhage reported in the EHR was used to analysis. To implement the algorithm, the WHO and American College of Obstetricians and Gynecologists (ACOG) recommendations were used. The decision rules were developed to estimate the hemorrhagic risk score within the 10 groups proposed by the Robson classification. Discussion: It's expected that the system will help to reduce unnecessary C-section rates and prevent PPH, providing better conditions of prognosis for mother and her newborn.

Paper Nr: 44
Title:

LOCATE: A Web Application to Link Open-source Clinical Software with Literature

Authors:

Zhengru Shen and Marco Spruit

Abstract: Nowadays, the effective utilization of open-source software could significantly boost both clinical research and practices, especially in resource-poor countries. However, the plethora of open-source clinical software has left many people unable to quickly locate the appropriate one for their needs. Commonly available software quality metrics and software documentation, such as downloads, forks, stars, and readme files, are useful selection criteria, but they only indicate the software quality from the perspective of IT experts. This paper proposes a method that offers additional insights on the performance and effectiveness of clinical software. It links open-source clinical software with relevant scientific literature, such as papers that use case studies of clinical software to reveal the strength and weakness of a given software from the clinical perspective. To interactively present the open-source clinical software and their related literature, we have developed the LOCATE web application that enables users to explore related literature for a given open-source clinical software. Moreover, the peer-review cycle of the application allows users to improve the application by confirming, adding or removing related literature. An evaluation experiment of the five most popular open-source clinical tools demonstrates the potential usefulness of LOCATE.

Paper Nr: 45
Title:

Evaluating the Process Capability Ratio of Patients’ Pathways by the Application of Process Mining, SPC and RTLS

Authors:

Sina Namaki Araghi, Franck Fontanili, Elyes Lamine, Nicolas Salatge, Julien Lesbegueries, Sebastien Rebiere Pouyade and Frederick Benaben

Abstract: Learning how patients receive their health treatments is a critical mission for hospitals. To fulfill this task, this paper defines patients’ pathways as business process models and tries to apply process mining, real-time location systems(RTLS), and statistical process control (SPC) as a set of techniques to monitor patients’ pathways. This approach has been evaluated by a case study in a hospital living lab. These techniques analyze patients’ pathways from two different perspectives: (1)control-flow and (2)performance perspectives. In order to do so, we gathered the location data from movements of patients and used a proof of concept framework known as R.IO-DIAG to discover the processes. To elevate the performance analyses, this paper introduces the process capability ratio of the patients’ pathways by measuring the walking distance. The results lead to the evaluation of the quality of business processes.

Paper Nr: 46
Title:

A Machine Learning-based Approach for Collaborative Non-Adherence Detection during Opioid Abuse Surveillance using a Wearable Biosensor

Authors:

Rohitpal Singh, Brittany Lewis, Brittany Chapman, Stephanie Carreiro and Krishna Venkatasubramanian

Abstract: Wearable biosensors can be used to monitor opioid use, a problem of dire societal consequence given the current opioid epidemic in the US. Such surveillance can prompt interventions that promote behavioral change. The effectiveness of biosensor-based monitoring is threatened by the potential of a patient’s collaborative non-adherence (CNA) to the monitoring. We define CNA as the process of giving one’s biosensor to someone else when surveillance is ongoing. The principal aim of this paper is to leverage accelerometer and blood volume pulse (BVP) measurements from a wearable biosensor and use machine-learning for the novel problem of CNA detection in opioid surveillance. We use accelerometer and BVP data collected from 11 patients who were brought to a hospital Emergency Department while undergoing naloxone treatment following an opioid overdose. We then used the data collected to build a personalized classifier for individual patients that capture the uniqueness of their blood volume pulse and triaxial accelerometer readings. In order to evaluate our detection approach, we simulate the presence (and absence) of CNA by replacing (or not replacing) snippets of the biosensor readings of one patient with another. Overall, we achieved an average detection accuracy of 90.96% when the collaborator was one of the other 10 patients in our dataset, and 86.78% when the collaborator was from a set of 14 users whose data had never been seen by our classifiers before.

Paper Nr: 47
Title:

Web Scraping Online Newspaper Death Notices for the Estimation of the Local Number of Deaths

Authors:

Rainer Schnell and Sarah Redlich

Abstract: Since access to real-world data is often tedious, web scraping has gained popularity. A health research example is the monitoring of mortality rates. We compare the results of local online death notices and print-media obituaries to administrative mortality data. The web scraping process and its problems are being described. The resulting estimates of death rates and demographic characteristics of the deceased are statistically different from known population values. Scaped data resulted in a sample that is more male, older and contains less foreign nationals. Therefore, using web scraped data instead of administrative data cannot be recommended for the estimation of death rates at this time for Germany.

Paper Nr: 61
Title:

Navigating the Jungle of Assistance Systems: A Comparison of Standards for Assistance Functionality

Authors:

Bastian Wollschlaeger and Klaus Kabitzsch

Abstract: Health Smart Homes offer assistance capabilities and facilitate the shift towards individual, precise health care. However, due to the variety of patient requirements and the enormous amount of existing solutions, the manual engineering of customized assistance systems becomes infeasible. By further automating this design approach, a customization of home-based assistance systems can be facilitated. In order to enable an automated design approach of assistance systems for home-based health care, a common functional vocabulary needs to be agreed upon. This paper proposes a literature-based categorization of established assistance functions and a literature comparison based on these categories to facilitate standardization of assistance functions. Therefore, we analyze standards and experience reports to identify and categorize the most common assistance use cases and functions. The results show that there is no single standard defining the most common assistance functions, which hampers communication and impedes the design process of health smart homes. To mitigate this effect, we envision a building block-based definition of a common vocabulary for assistance functions.

Paper Nr: 63
Title:

Accelerometer-based Sleep/Wake Detection in an Ambulatory Environment

Authors:

Jan Cornelis, Elena Smets and Chris Van Hoof

Abstract: It has been shown that poor sleep quality and stress are major causes for mental and physical health problems in developed countries. Thanks to advancements in wearable technology, remote patient monitoring has become possible, without the need of cumbersome and expensive equipment. A method for sleep/wake detection is proposed, using chest-worn accelerometer sensors. A total of 1727 nights from 580 individuals were analysed, resulting on the identification of an average sleep time of 463 min (SD=±80 min) per day. Our algorithm was able to automatically detect 483 min (SD=±97 min) of sleep. Results show that actigraphy with an accelerometer located at the chest has potential for sleep monitoring, though further research is required for further validation, preferably using polysomnography as a benchmark.

Paper Nr: 64
Title:

Prediction of Acute Kidney Injury in Cardiac Surgery Patients: Interpretation using Local Interpretable Model-agnostic Explanations

Authors:

Harry Freitas da Cruz, Frederic Schneider and Matthieu-P. Schapranow

Abstract: Acute kidney injury is a common complication of patients who undergo cardiac surgery and is associated with additional risk of mortality. Being able to predict its post-surgical onset may help clinicians to better target interventions and devise appropriate care plans in advance. Existing predictive models either target general intensive care populations and/or are based on traditional logistic regression approaches. In this paper, we apply decision trees and gradient-boosted decision trees to a cohort of surgical heart patients of the MIMIC-III critical care database and utilize the locally interpretable model agnostic approach to provide interpretability for the otherwise opaque machine learning algorithms employed. We find that while gradient-boosted decision trees performed better than baseline (logistic regression), the interpretability approach used sheds light on potential biases that may hinder adoption in practice. We highlight the importance of providing explanations of the predictions to allow scrutiny of the models by medical experts.

Paper Nr: 68
Title:

A MapReduce-like Deep Learning Model for the Depth Estimation of Periodontal Pockets

Authors:

Yusuke Moriyama, Chonho Lee, Susumu Date, Yoichiro Kashiwagi, Yuki Narukawa, Kazunori Nozaki and Shinya Murakami

Abstract: This paper explores the feasibility of diagnostic imaging using a deep learning-based model, applicable to periodontal disease, especially periodontal pocket screening. Having investigated conventional approaches, we find two difficulties to estimate the pocket depth of teeth from oral images. One is the feature extraction of Region of Interest (ROI), which is pocket region, caused by the small ROI, and another is tooth identification caused by the high heterogeneity of teeth (e.g., in size, shape, and color). We propose a MapReduce-like periodontal pocket depth estimation model that overcomes the difficulties. Specifically, a set of MapTasks is executed in parallel, each of which only focuses on one of the multiple views (e.g., front, left, right, etc.) of oral images and runs an object detection model to extract the high-resolution pocket region images. After a classifier estimates pocket depth from the extracted images, ReduceTasks aggregate the pocket depth with respect to each pocket. Experimental results show that the proposed model effectively works to achieve the estimation accuracy to 76.5 percent. Besides, we verify the practical feasibility of the proposed model with 91.7 percent accuracy under the condition that a screening test judges severe periodontitis (6 mm or more).

Paper Nr: 73
Title:

Predictive AI Models for the Personalized Medicine

Authors:

Luigi Lella, Ignazio Licata, Gianfranco Minati, Christian Pristipino, Antonio G. De Belvis and Roberta Pastorino

Abstract: Innovative information systems which enable personalized medicine are presented. The designed decision support systems are expected to infer with an excellent level of accuracy the outcome of a therapeutic intervention through the analysis of biometric, genetic and environmental data. They are also capable to motivate their predictions according to a dynamic knowledge base, which is kept updated with new analysed cases. These systems can be used by researchers to identify useful correlations between biometric, genetic and environmental data with potential risks and benefits of certain therapeutic choices. They can also be used by the patients to choose the most appropriate therapeutic intervention according to their needs and expectations. In other words the presented decision support tools can realize the vision of the predictive, preventive, personalized and participatory (P4) medicine pursued by the systemic medicine.

Paper Nr: 74
Title:

Towards an “Operational” Educational Model in Healthcare: Exploiting Computer-Interpretable Guidelines

Authors:

Alessio Bottrighi, Gianpaolo Molino, Luca Piovesan and Paolo Terenziani

Abstract: Clinical guidelines (GLs) encode the best medical practices. GLs have been widely exploited to enhance the quality of patient care, and to optimize it, and several computer-based approaches to manage computer-interpretable guidelines (CIGs) have been proposed in the literature. Quite surprisingly, however, the potentialities of CIG systems in medical education have not been considered yet. In this position paper we argue that, since CIG systems support the “simulation” of the application of GLs on specific patients, they can be used to show students how to apply medical knowledge and best practices on specific patients. Therefore, using CIG systems, students may learn an “operational methodology” that, otherwise, they could only learn from the medical practice. In this paper, we have taken GLARE (and its extension, META-GLARE) as an example of CIG system, and we have addressed the roadmap we intend to follow to fully exploit its potentialities in medical education.

Paper Nr: 87
Title:

Making Clinical Simulation Mannequins Talk

Authors:

Isac Cossa, Guilherme Campos, Pedro S. Couto and João Lindo

Abstract: This paper advocates the interest of applying speech synthesis to clinical simulation mannequins, by focussing on a particular case study of recognised practical interest (evaluation of consciousness level based on the Glasgow Coma Scale), chosen as a proof of concept. A response repository comprising 109 sentences was recorded and an application was developed in Microsoft® Visual Basic® to allow configuration of the simulation scenario, control of response generation on a low-fidelity mannequin equipped with a loudspeaker and assessment of trainee performance. The system received very positive assessment in initial user tests on a typical training setting.

Paper Nr: 94
Title:

ExPhobia: A mHealth Technology based on Augmented Reality for Support the Treatment of Arachnophobic Stimulus

Authors:

Fabiana V. Brasileiro, Tauily D. Taunay, José V. Filho, Manuela M. Santana and Ântonio F. Bastos

Abstract: Anxiety disorders attack a significant part of the population and have been treated, mainly, by pharmaceuticals or psychotherapies. Augmented Reality technologies have gotten some space in the treatment of these diseases, with promising results. The current paper presents the prototype ExPhobia that holds concepts and tools from the mHealth and Augmented Reality areas for the support to the facing of arachnophobia. We have as a main goal the development and the performing of the initial tests of ExPhobia for the systemic desensitization on levels of anxiety and escape/dodge in people with arachnophobia. The used methodology for the development of ExPhobia contemplates a laboratorial phase for the conception of this functional prototype, supported by the stages of the interaction design and, after the initial tests, the program is being planned in a way that a research in a method almost experimental can be done, with only one individual, applying the ABA designing. For measuring the levels of anxiety and phobia of the participants in scales, inventories and questionings, in addition to the heart frequency measuring. It is a hypothesis that can be used in the Augmented Reality as a technique of systematic desensitization, from an analytic-behavioral perspective, the positive and viable results for the treatment of specific phobias.

Paper Nr: 95
Title:

Inter-observer Reliability in Computer-aided Diagnosis of Diabetic Retinopathy

Authors:

João Gonçalves, Teresa Conceição and Filipe Soares

Abstract: The rapid growth of digital data in healthcare demands medical image analysis to be faster, precise and, at the same time, decentralized. Deep Learning (DL) fits well in this scenario, as there is an enormous data to sift through. Diabetic Retinopathy (DR) is one of the leading causes of blindness that can be avoided if detected in early stages. In this paper, we aim to compare the agreement of different machine learning models against the performance of highly trained ophthalmologists (human graders). Overall results show that transfer learning in the renowned CNNs has a strong agreement even in different datasets. This work also presents an objective comparison between classical feature-based approaches and DL for DR classification, specifically, the interpretability of these approaches. The results show that Inception-V3 CNN was indeed the best-tested model across all the performance metrics in distinct datasets, but with lack of interpretability. In particular, this model reaches the accuracy of 89% on the EyePACS dataset.

Paper Nr: 100
Title:

Run-time Support to Comorbidities in GLARE-SSCPM

Authors:

Alessio Bottrighi, Luca Piovesan and Paolo Terenziani

Abstract: Comorbidities play a relevant role in healthcare, so that, in the last years, several approaches Medical Informatics and Artificial Intelligence have developed software tools to support physicians in the treatment of comorbid patients. Computer Interpretable Guidelines (CIGs) are consolidated decision support tools to help physicians, but they are devoted to provide evidence-based recommendations for one specific disease. In order to support the treatment of patient affected by multiple diseases, challenging additional problems have to be addressed, such as (i) the detection of the interactions between CIG actions, (ii) their management, and, finally, (ii) the “merge” of CIGs. Several CIG approaches have been recently extended in order to face (at least one of) such challenging problems, and one of them is GLARE (GuideLine Acquisition Representation and Execution). However, such approaches have mostly focused on the “a-priori” treatment of such problems, while addressing them “run-time” (i.e., to support physicians during the execution of the CIGs on a specific patient) involves additional challenges, and requires additional methodologies. In this paper we take advantage of previous extensions of GLARE (to cope with issues (i), (ii), (iii)), and propose a new knowledge-based, “focused” and interactive management of comorbid patients.

Paper Nr: 101
Title:

Virtual Reality based Diagnosis System for Visuospatial Neglect

Authors:

Kais Siala, Mohamed Kharrat, Mohamed Belghith, Olfa Boubakri, Sameh Ghroubi, Mohamed H. Elleuch and Mohamed Abid

Abstract: The diagnosis of Visuospatial neglect is generally conducted using paper based manual methods. The obtained results could be confused with sensory inattention pathology. In this paper we are presenting a new Virtual Reality based diagnosis method for patients suffering from Visuospatial neglect. For this purpose a Virtual Reality simulation called Farm Parade has been developed where the patient, after wearing a Virtual Reality headset, will be guided inside a road crossing a farm like environment where animals at both sides of the road will slowly move and generate sounds to encourage patients to look at them. The patient head motion will be then tracked to generate a graph showing his capacity to move his head and judge if he is suffering from Visuospatial neglect. The strength of the proposed system is the generation of numerical values relative to the amount of head’s rotation which could be helpful for measuring precisely the degree of recovery.

Paper Nr: 1
Title:

Guidelines for Health IT Addressing the Quality of Data in EHR Information Systems

Authors:

Nabil G. Badr

Abstract: Quality of patient care is dependent on the quality of patient healthcare data. Electronic Health Record Information Systems (EHR-IS) capture patient health data for diagnosis, treatment, testing, medication and patient support. Issues in healthcare data quality comprise missing, incorrect, imprecise, and irrelevant data. Stakeholders of health data from practitioners, to patients, governments and lawmakers have long concerned themselves with these issues. Our paper looks at data quality in healthcare from the locus of ensuing risks, challenges and approaches in the literature. The paper proposes a reference for designing Electronic Health Record Information Systems and the evaluation of data quality in EHR-IS implementations.

Paper Nr: 9
Title:

Accessing and Sharing Electronic Personal Health Data

Authors:

Maria Karampela, Sofia Ouhbi and Minna Isomursu

Abstract: An increasing attention has been given to personal health data (PHD) research over the last years. The rise of researchers’ interest could be attributed to the increasing amount of PHD that are stored across various databases, as a result of individuals’ rapidly-evolving digital life. Accessing and sharing PHD is essential to create personalized health services and to involve patients in the design process of these services. This paper conducts a survey of literature to present an overview of literature about accessing and sharing of PHD. This study aims to identify limitations in research and propose future directions. Sixteen studies were selected from various bibliographic databases and were classified according to three criteria: research type, empirical type and contribution type. The results provide a preliminary review with respect to access and sharing of PHD, addressing a need for more research about PHD accessibility and for solution proposals for both topics.

Paper Nr: 16
Title:

Medication Literacy in a Cohort of Chinese Patients Discharged with Essential Hypertension

Authors:

Zhuqing Zhong, Siqing Ding, Shuangjiao Shi, Zehua Xu and Aijing Luo

Abstract: To assess medication literacy and important determinants of medication literacy in discharged patients with essential hypertension, we conducted a prospective cohort study in a tertiary hospital in Changsha, Hunan, China between March and June 2016.Patient’s demographic and clinical data were retrieved from hospital charts and medication literacy was measured by structured interview using the Chinese version of Medication Literacy Questionnaire on Discharged Patient between 7 and 30 days after discharged. The results showed that medication literacy was insufficient: > 20% did not have adequate knowledge on the type of drugs and frequency that they need to take the drugs, > 30% did not know the name and dosage of the drugs they are taking, and > 70% did not have adequate knowledge on the effects and side effects of the drugs they are taking. Medication literacy score decreased with age but increased with education level and length of hospital stay.

Paper Nr: 19
Title:

Automatic Detection and Recognition of Swallowing Sounds

Authors:

Hajer Khlaifi, Atta Badii, Dan Istrate and Jacques Demongeot

Abstract: This paper proposes a non-invasive, acoustic-based method to i) automatically detect sounds through a neck-worn microphone providing a stream of acoustic input comprising of a) swallowing-related, b) speech and c) other ambient sounds (noise); ii) classify and detect swallowing-related sounds, speech or ambient noise within the acoustic stream. The above three types of acoustic signals were recorded from subjects, without any clinical symptoms of dysphagia, with a microphone attached to the neck at a pre-studied position midway between the Laryngeal Prominence and the Jugular Notch. Frequency-based analysis detection algorithms were developed to distinguish the above three types of acoustic signals with an accuracy of 86.09%. Integrated automatic detection algorithms with classification based on Gaussian Mixture Model (GMM) using the Expectation Maximisation algorithm (EM), achieved an overall validated recognition rate of 87.60% which increased to 88.87 recognition accuracy if the validated false alarm classifications were also to be included. The proposed approach thus enables the recovery from ambient signals, detection and time-stamping of the acoustic footprints of the swallowing process chain and thus further analytics to characterise the swallowing process in terms of consistency, normality and possibly risk-assessing and localising the level of any swallowing abnormality i.e. the dysphagia. As such this helps reduce the need for invasive techniques for the examination and evaluation of patient’s swallowing process and enables diagnostic clinical evaluation based only on acoustic data analytics and non-invasive clinical observations.

Paper Nr: 32
Title:

Assessing the Need of Decision-making Frameworks to Guide the Adoption of Health Information Systems in Healthcare

Authors:

Raja M. Abbas, Noel Carroll and Ita Richardson

Abstract: Hospital Information System (HIS) is important in the healthcare industry as it supports a wide range of highly specialized health-care tasks, services and provide high-quality patient care. Adoption of HIS is one of the key decisions by hospital management, yet the function of hospital decision-makers within the area of new technology adoption, specifically the decision-making processes in the adoption of HIS remains unsupported. To investigate this phenomenon, this paper identifies HIS decision-making theories, their short-coming of adoption in healthcare organisations and decision-making facets that influence the adoption. These review will shed some light for future researchers to conceptualize, distinguish and comprehend the underlying HIS decision-making models and theories that may affect the future application of HIS adoption. A literature search was conducted to identify studies presenting HIS decision-making adoption theories/models in a healthcare environment. From synthesis of 26 studies, we identified five major facets that provides a structure to organize and capture information on the decision-making and adoption of HIS. The themes presented here provide a starting point in understanding the decision-making adoption theories, their major facets and their short-coming in adopting HIS. This will facilitate our future research on decision-making framework for the adoption of HIS.

Paper Nr: 39
Title:

An Annotation Model for Patient Record

Authors:

Khalil Chehab, Anis Kalboussi and Ahmed H. Kacem

Abstract: The efforts to computerize the medical record of a patient began in 1990. In the documents of this record, the healthcare professional practices the annotation activity. Most medical annotation systems are made to perform a specific task. As a result, we have dozens of medical annotation system that we sneak a fragmented image in the absence of generic classification for these. In this article, we try to present an annotation model to the patient record based on a unified classification of medical annotation systems. We have determined the existing limits and we try to override them in the proposed model.

Paper Nr: 43
Title:

A Framework for Performance Management of Clinical Practice

Authors:

Pilar Mata, Craig Kuziemsky and Liam Peyton

Abstract: In this paper, we present a framework for performance management of clinical practice. The framework defines a performance management participation model, which identifies the processes that need to be managed at the micro, meso, and macro levels for a clinical practice, and which identifies the key actors and tasks relevant to performance management. It defines a performance measurement model, which maps goals and indicators to the performance management participation model. In addition, it includes a methodology for implementation and evaluation of tools that can be integrated into care processes to support the data collection and report notification tasks identified in the performance management participation model. We revisit the case study of implementing a resident practice profile app, in light of the proposed framework, to support performance management of a family health practice. We demonstrate how the framework is useful for explaining why the use of the app was abandoned after two years of its introduction and, therefore, how the framework is an improvement to our previous methodology for development of performance management apps for clinical practice.

Paper Nr: 49
Title:

Hospital Bed Management Practices: A Review

Authors:

Flannagán Noonan, Jacinta O’Brien, Eilish Broderick, Ita Richardson and Joseph Walsh

Abstract: This paper reviews current literature on the bed management role seeking to highlight developments most likely to increase efficiency. A reduction in the number of in-patient beds due in part to innovative surgical techniques is causing increased pressure on a very finite resource. This requires a greater emphasis on the bed management role and the wider hospital team. A number of studies are presented describing initiatives implemented to support bed management both operationally, procedurally and from a decision support approach. Finally, literature on people, process technology approaches in healthcare is presented, which could support a sustainable improvement in the role.

Paper Nr: 55
Title:

Process Mining in Frail Elderly Care: A Literature Review

Authors:

Nik F. Farid, Marc De Kamps and Owen A. Johnson

Abstract: Process mining has proved to be a valuable technique for extracting process knowledge from data within information systems. Much work has been conducted in applying process mining to domains such as logistics, banking, transportation and many areas of the government, including healthcare. Frail elderly people who have an increased risk of adverse outcomes are amongst the main users of healthcare services and understanding healthcare processes for the frail elderly is challenging because of their diverse and complex needs combined with an often high number of co-morbidities. This paper aims to provide an overview of work applying process mining techniques to improving the care of frail elderly people. We conducted a literature search using broad criteria to identify 1,047 potential papers followed by a review of titles, abstract and content which identified eight papers where process mining techniques have been successfully applied to the care of frail elderly people. Our review shows that, to date, there has been limited application of process mining to support this important segment of the population. We summarise the results based on five themes that emerged: types of source data and process; geographical location; analysis methodology; medical domain; and challenges. Our paper concludes with a discussion on the issues and opportunities for process mining to improve the care pathways for frail elderly people.

Paper Nr: 58
Title:

A Graphical User Interface for an Automatic Rest-activity Cycle Detection and Dichotomy Index Computation

Authors:

Racha Soubra, Aly Chkeir, Mohammad O. Diab, Majd Abdallah and Jacques Duchene

Abstract: The success of chemotherapy treatment is achieved based on “Chronotherapy”: the concept of administering the correct drug at a precise time based on the circadian rhythm study. This paper aims to detect the rest/activity cycle and automatically calculate the dichotomy index (I<O), as both parameters have been proved to be reliable indices of the circadian rhythm. First, the DARC “Détection Automatique du Rythme Circadien” algorithm is used to segment the rest-activity phases automatically. Then, a Graphical User Interface (GUI) is used to calculate easily the I<O across several days of records and smooth the analysis. The outcome of this study provides an easy-to-use GUI that minimizes patients’ intervention, facilitates user involvement, and reduces the time required for analysis.

Paper Nr: 59
Title:

A Novel Approach to Gene Analysis: Gene Panels and Cluster Definition to Assist Genotyping Patients with Congenital Myopathies

Authors:

Marco Calderisi, Ilaria Ceppa, Denise Cassandrini, Rosanna Trovato, Giulia Bertocci, Alessandro Tonacci, Guja Astrea, Raffaele Conte and Filippo M. Santorelli

Abstract: The boundaries between congenital myopathies and muscular dystrophies and other neuromuscular disorders are becoming blurred because of the significant overlap in disease genes, clinical presentations, and histopathological features. Using a MotorPlex7.0 gene panel in massive sequencing, we define disease causative mutations in 76% of our sample. We then analysed the extent of gene information in the data using non metric multidimensional scaling (nMDS), a well-known algorithm for multivariate analysis, and clustering techniques. To perform this analysis, we developed a software that allows for an interactive exploration of the variants dataset and of the results of the nMDS model. Using these techniques, we were able to quickly study a dataset consisting of thousands of variants, identifying groupings of patients based on the presence or absence of specific sets of mutations.

Paper Nr: 60
Title:

Modeling (Multi-)Morbidity and (Poly-)Pharmacy in Outpatient Treatment with Gamma Distributions

Authors:

Reinhard Schuster, Thomas Ostermann and Timo Emcke

Abstract: Polypharmacy is often direcly causes by age and gender dependent multimorbidity. Todays treatment concepts, the individual decisions taken by physicians and the administration have to adress the complex needs of multimorbid patients. For modeling those phenomena on a collective level of an entire federal state a sufficiently large data repository is essential. The administrative bodies of the statutory health insurance in Germany have the data necessary and built up an extensive skill-set and inexpensive free-software tool-set for those evaluations. This study analyses the complete patient data of all outpatient treatments and drug prescriptions in Schleswig-Holstein (Northern German federal state) in the first quarter of 2017. Well adopted probability distributions for the frequency of diseases and drug groups decreasingly ordered within the classification system for all patients and age/gender partitions are estimated. Subsequently the levels of multimorbidity and polypharmacy (level of ICD-10/ATC-codes per quarter) are analysed in the same way. As a main result gamma distributions provide a well-adjusted model class for ICD and ATC code frequencies in the present very large routine dataset. The goodness-of-fit (full range of magnitudes of measurements) is much better than using mean values and variances.

Paper Nr: 62
Title:

An Experimental Investigation Comparing Age-Specific and Mixed-Age Models for Wearable Assisted Activity Recognition in Women

Authors:

Pratool Bharti, Arup K. Dey, Sriram Chellappan and Theresa Beckie

Abstract: In this paper, we investigate the impact of age diversity on accuracy for activity recognition among women with wrist-worn wearables. Using a sample of 10 elder women and 10 younger women, and by monitoring five activities related to cardiac care (Running, Brisk Walking, Walking, Standing and Sitting), we show that while personalized models are best, activities classification based on age specific models are definitely superior in terms of accuracy compared to classification using mixed age models. We do so by a) extracting 11 features from inertial sensing data; b) reducing dimensionality using Linear Discriminant Analysis methods; c) quantifying variance among features using Principal Component Analysis; d) clustering activities; and finally e) comparing classification accuracies of all activities for personalized, age-specific and mixed-age models. We believe that our study is unique, and potentially important for superior healthcare for women, a demographic that is largely underserved today across the world.

Paper Nr: 78
Title:

Quantitative Gait Measurement with Wave Doppler-radar for Elderly Walking Speed Recognition

Authors:

Maha Reda, Aly Chkeir, Racha Soubra and Mohamad Nassereddine

Abstract: This paper studies the use of a device based on a Doppler sensor in estimating the gait velocity in a non-controlled environment. It provides signals of the instantaneous velocity with an irregular time sampling. A high accuracy motion capture system, Vicon, was employed to provide the reference data for device evaluation. The gait parameters have been validated with a Vicon motion capture system in our lab. A proper algorithm based on the Lomb-Scargle periodogram was proposed to extract features from the radar signals such as the dominant frequency and the number of steps performed. These features were then used to calculate gait parameters such as the gait velocity and the step duration on a 5m walking sequence at a normal pace. The results showed the reliability of the Doppler device in estimating the gait velocity (mean error = 5.1%).

Paper Nr: 80
Title:

Comparison of Spatial Interpolation Methods based on Exposure Assessments of Air Pollutants: A Case Study on Nuclear Substances in Fukushima

Authors:

Takahiro Otani, Kunihiko Takahashi, Ayano Takeuchi and Mari Asami

Abstract: In response to accidents and disasters involving the proliferation of pollutants to the environment, performing exposure assessments across a region of impact is important for evaluating health effects. Owing to the typical unavailability of the spatially continuous data of pollutant concentrations immediately after accidents, various spatial interpolation methods have been studied to assess exposures using limited available data. In this study, we compared representative spatial interpolation methods based on the estimation of the distributions of exposures through a case study of the Fukushima Daiichi nuclear disaster initiated by the Great East Japan earthquake and subsequent tsunamis. The nearest neighbour method, inverse distance weighted method, and ordinary kriging method were compared in the context of exposure assessments. Even though estimated air dose rates were slightly different depending on the method used, different interpolation methods produced significantly equivalent estimates of the distribution of cumulative exposure over one year.

Paper Nr: 81
Title:

Ensuring Secure Health Data Exchange across Europe. SHIELD Project

Authors:

Borja López-Moreno, David Martín-Barrios, Ivan Revuelta-Antizar, Santiago Rodríguez-Tejedor, M. d. Valle and Eunate Arana-Arri

Abstract: Nowadays, many people move from one country to another for various reasons: tourism, work, studies, etc.; even with chronic or multi-pathological diseases. The main objective of SHIELD project is to create open and extendable security architecture with supported privacy mechanisms and trust of citizens, to provide systematic protection for the storage and exchange of health data across European borders. epSOS is a European project funded and finished dealing with security and interoperability of eHealth data is, that result in an OpenNCP (National Contact Point) architecture. In SHIELD project for the initial validation framework two OpenNCP virtual nodes would simulate the real nodes between Italy and Spain. Validation scenarios (realistic use cases) have been developed in three different member states (Italy, United Kingdom and Spain). The first scenario is an Italian citizen traveling to Spain that has an acute emergency episode (e.g. stroke) and loses consciousness. Spanish emergency department suddenly assists that patient and doctor wishes to check patientś health record. Results of the first round of validation frameworks of SHiELD project have been made successfully and presented to the European Commission. Security challenges need to be addressed when assessing eHealth solutions. Among others, the challenges are: interoperability, confidentiality, availability, integrity, privacy, ethics, regulations and eHealth data. Which data are going to be shared and by which mean? The first validations will be useful as the basis for both the “in depth” requirements analysis as well as setting the main pillars for the SHIELD architecture detailed design.

Paper Nr: 82
Title:

Subgroup Anomaly Detection using High-confidence Rules: Application to Healthcare Data

Authors:

Juan L. Domínguez-Olmedo, Jacinto Mata, Victoria Pachón and Manuel Maña

Abstract: In real datasets it often occurs that some cases behave differently from the majority. Such outliers may be caused by errors, or may have differential characteristics. It is very important to detect anomalous cases, which may negatively affect the analysis from the data, or bring valuable information. This paper describes an algorithm to address the task of automatically detect subgroups and the possible anomalies with respect to those subgroups. By the use of high-confidence rules, the algorithm determines those cases that satisfy a rule, and the cases discordant with that rule. We have applied this method to a dataset regarding information about breast cancer patients. The resulting subgroups and the corresponding outliers have been presented in detail.

Paper Nr: 83
Title:

Animal Health Informatics: Towards a Generic Framework for Automatic Behavior Analysis Position Paper

Authors:

Dmitry Kaplun, Aleksandr Sinitca, Anna Zamansky, Stephane Blauer-Elsner, Michael Plazner, Asaf Fux and Dirk van der Linden

Abstract: The field of veterinary healthcare informatics is still in its infancy, and state of the art solutions from human healthcare are not easily adapted. IoT and wearable technologies may be bringing a wind of change, as large amounts of health data of animals are now being produced. It makes this a timely moment to initiate a discussion on the possibilities for cross-fertilization between the worlds of human and animal health informatics. In this position paper we report on an ongoing project developing a framework for automatic video-based analysis of animal behavior and describe its concrete application for decision support of behavioral veterinarians. The framework is generic, allowing for reuse across species and different analytical tasks. We further discuss the possibilities for cross-fertilization between human and animal behavior analysis in the context of health informatics.

Paper Nr: 85
Title:

Effect of Feedback Medium for Real-time Mastication Awareness Increase using Wearable Sensors

Authors:

Guillaume Lopez, Hideto Mitsui, Joe Ohara and Anna Yokokubo

Abstract: Increasing the number of mastications can help suppress obesity, but it is difficult to keep constant awareness of it in everyday life. Besides, the conventional mastication number measurement apparatus is large and non-portable such it is difficult to use it in daily life. This research proposes a system that supports chewing and utterance consciousness improvement in real-time. It is composed of a cheap and small bone conduction microphone to collect sound intra-body sound signal, and a smartphone that processes sound and provides feedback in real-time so that it can be used conveniently in daily life. First, the accuracy of mastication counting and utterance length estimation has been evaluated, confirming to be sufficient to provide real-time feedback. Second, the evaluation of the effect on chewing and utterance consciousness of different ways and medium of real-time feedback during a meal was carried out. It was possible to clarify the impact of real-time feedback, as well as to determine the factors that affect more efficiently the improvement of mastication number and utterance.

Paper Nr: 88
Title:

Evaluating Health-care-related Active Learning Class Lectures using Class Achievement and Text Mining of Free Descriptions

Authors:

Kazuma Mihara, Takahito Tamai and Yukie Majima

Abstract: Active learning is defined as "a general term for professors and learning methods that incorporate participation in active learning of the students, unlike teachers' unilateral lecture style education." Universities that improve classes from the viewpoint of active learning are increasing in recent years. A class evaluation questionnaire has been established to improve the understanding and satisfaction of students' classes. In many cases, the Likert scale is used for the class evaluation questionnaire. There are also aspects for which statistical processing is easy to do. However, it is difficult to ascertain the students ' specific opinions and ideas alone. Therefore, we attempted to evaluate health-care-related subjects from the two viewpoints of ‘free description’ and ‘degree of accomplishment of class goal’ for active learning classes aimed at students' subjective learning.

Paper Nr: 90
Title:

Non-Invasive Blood Pressure Monitoring with Positionable Three-chamber Pneumatic Sensor

Authors:

V. E. Antsiperov, G. K. Mansurov, M. V. Danilychev and D. V. Churikov

Abstract: The main goal of the paper is to present a new type of a sensor for non-invasive continuous blood pressure measurements. The principle of such a sensor operation, based on local pressure compensation, is in the centre of discussion. The sensor presented has very small sensing pads (1 mm2 or less) which permits accurate sensor positioning directly on the elastic surfaces such as the human skin. Thanks to that it is possible to provide a high quality of the blood pressure measurement, when keeping the continuity of the measurement parameters and minimizing the level of disturbances. For this reason, the paper focuses on a detailed discussion of the positioning problem and the results of developing approaches to its solving. As a promising method, a positioning based on the pulse wave controlling by a three-chamber pneumatic sensor is proposed.

Paper Nr: 91
Title:

Strategies to Access Patient Clinical Data from Distributed Databases

Authors:

João R. Almeida, Olga Fajarda, Arnaldo Pereira and José L. Oliveira

Abstract: Over the last twenty years, the use of electronic health record systems has become widespread worldwide, leading to the creation of an extensive collection of health databases. These databases can be used to speed up and reduce the cost of health research studies, which are essential for the advance of health science and the improvement of health services. However, despite the recognised gain of data sharing, database owners remain reluctant to grant access to the contents of their databases because of privacy and security issues, and because of the lack of a common strategy for data sharing. Two main approaches have been used to perform distributed queries while maintaining all data control in the hands of the data custodians: applying a common data model, or using Semantic Web principles. This paper presents a comparison of these two approaches by evaluating them according to parameters relevant to data integration, such as cost, data quality, interoperability, extendibility, consistency, and efficiency.

Paper Nr: 96
Title:

The Role of Technology in Protecting the Environment: Evaluating the Liquidity of Interactive Waste Bins

Authors:

Fabiana V. Brasileiro, Jully E. Melo, José V. Filho and Lucas A. Marques

Abstract: Issues related to the pollution of ecosystems and possible alternatives for their preservation are present in various debates. Behavior analysis has been devoted to the study of sustainable social behaviors and cultural practices involved in such problems. In this perspective, the project presented is an evaluation of the effectiveness of the "Interactive Waste Bin", a prototype developed by the Nucleus of Information Technology Application (NATI) that emits a sound after trash disposal, developed as an alternative to traditional waste bins models. The hypothesis is that the use of "interactive waste bins" is more efficient for the acquisition and maintenance of the behavioral patterns sought, reducing pollution levels and their effects for future generations through the use of immediate social reinforcement. To achieve the proposed goal in the project, there will be experiments aimed at exposing people that walk inside UNIFOR to the equipment, called the "interactive waste bin". The volumetric monitoring tool contained in the bin will count the waste dumped. The study will contain a Baseline Session (SA1); Intervention Sessions (B) and Baseline Return Session (SA2), seeking to evaluate the maintenance of the obtained effect.

Paper Nr: 103
Title:

Development of a Speech Recognition System Specialized in Applications in a Hospital Setting

Authors:

Alejandro Renato, Oscar Jáuregui, Mariana Daus, Hernán Berinsky, María L. Gambarte, Daniel Luna and Miguel F. Dachery

Abstract: In this paper we describe the development of an automatic speech recognition system for medical applications within a hospital setting. The system was developed using available resources in the field of the hospital, both documents and audios. It is based on the Kaldi (Povey et al, 2011) toolkit, open and available code, using training techniques appropriate to the state of the art. The acoustic training was done with 800 hours of audio from Spanish speakers in Argentina. The language model was built with 12 million medical notes, which are part of a 5-year history of the hospital and 80,000 words of vocabulary. The development of the system is oriented to be representative of the domain, from the acoustic aspects as well as the vocabulary. The effort in the development was concentrated in the quality of the corpus of text, dictionary and acoustic models over the quantitative aspects. Although initially oriented to serve as an alternative to the entry of information through the keyboard in Electronic Health Record (EHR) systems, this article evaluates the possibility of generalizing the system for various needs to record information within the hospital setting. To this end, an evaluation was made for the transcription of 10 hours of medical reports from different specialties randomly selected from a set of 620 hours available on the web. The results of a test performed in the hospital of clinical notes read shows a performance of 94% accuracy, while for the evaluation of fragments of papers randomly chosen, in uncontrolled acoustic conditions and spontaneous oral speech, had a performance of 84,8% in audios with at least 20 db of signal to noise ratio . Finally, we reflect on the advantages of the development of a project of these characteristics within a hospital environment as well as its projections.