Good Morning MIE2021 Participants!
Sunday 30th of May was the second day of the 31st Medical Informatics Europe Conference, the first fully digital edition of the flagship Conference, of the European Federation for Medical Informatics. We started early and we had another full day, full of paper presentations, interactive workshops and panels.
As previously noted you can already consult the medline indexed papers in proceedings published by IOS press, and the video presentation earlier today and yesterday in the video library thanks to the generous sponsorship of SNOMED International.
I asked the chairs of the sessions and workshop hosts to share some quick impressions. Let’s see what they shared.
Catherine Chronaki, EFMI President (2020-2022)
Session 13- Health and prevention (May 30, 9am, room A), Chair Izet Masic
By Izet Masic, EFMI Honorary Fellow
In the sesion, «Health and Prevention» held on May 30th 2021, four very interesting papers were presented. Christina Gena Dascalu et al. „Ten-months Study Regarding COVID-19 Spreading in Romanian Counties“ analyzes the evolution of COVID-19 cases in Romanian counties over a period of 10 months, to highlight possible similarities that may contribute to a better understanding of the spreading pattern.
The study uses the number of active cases for each county in Romania, as well as Bucharest and the whole country, reported daily by the Romanian Ministry of Health. Authors compared the disease’s evolution in Suceava county with other counties in Romania in order to highlight the gaps between them and calculated the cross-correlations between counties interpreting as time series and found that lags varied between 1-15 days, the most counties having a lag of 6-7 days compared with Suceava and concluded that, on long term there are no important discrepancies between the regions in Romania regarding the evolution of the disease, which shows that the intervention efforts of the medical staff were uniform in efficiency.
Gabriela Varga et al. spoke about „Serious Gaming AI Supporting Treatment in Rheumatoid Arthritis“ and presented a complex application for rehabilitation of patients with first and second stage rheumatoid arthritis (RA). The application contains a module for the doctor, for the kinetotherapist, and a module as a gadme matching the symptoms for each stage of RA. This neural network application, as authors stated, could assist rehabilitation of the RA hand with support of digital technology and multimodal interaction: leap motion, serious gaming, and neuronal networks as support for patients to perform the exercises at home classifying the correct movement with accuracy of 95%. The application was tested by a group of students, and the degree of mental stress, fatigue in the fingers, wrists and physical exertion were insignificant in most cases.
The third paper entitled „Use and Users of the Webbased Omaolo Covid-9 Symptom Assessment Tool in Finland Since March 16, 2020“ presented by Vesa Jormanainen et al. intended to answer if it is possible to quickly produce medical symptom self-assessment tools within the existing infrastructure and experimented with the Finnish Omaolo Covid-19 web-based symptom self-assessment tool. Authors tested the web-based Omaolo Covid-19 symptom checker during the second wave of the epidemic on sample range with 1.72 million questionnaires were recorded, out of which 1.55 million from symptomatic persons. Some 15% of the responses (245,500) were directed to seek emergency medical care based on the online screening by respondent response profiles.
Lastly, authors Joao Bettencourt-Silva et al. presented paper titled: „Social Determinant Trends of COVID-19: An Analysis Using Knowledge Graphs from Published Evidence and Online Trends“ where he pointed that the social determinants of health should play an increasingly important part in complementing public health datasets and in strengthening population trend analyses. The paper presents the results of a new approach to discover related health and social factors during the COVID-19 pandemic, which leverages a knowledge graph of related concepts mined from a corpus of published evidence (PubMed) prior to the pandemic. Population trends from online searches were used to identify social determinants of health (SDoH) concepts that trended high at the outset of the pandemic from a list of SDoH topics from the World Health Organization (WHO). Authors suggest an approach to derive new related health and social factors that may have either played a role in, or been affected by, the onset of the global COVID-19 pandemic and their results show how, from a list of SDoH topics, Food Security, Unemployment trended the highest at the start of the pandemic.
Discussion opened a lot of questions and dilemmas about COVID-19 pandemic and its medical, social, ethical, economical and other consequences related to the WHO statements and global targets “Health for all”, regarding determinants for health, especially for WHO strategy about Healthy aging and providing prevention measures within risk people group, who has a few most frequent co-morbidities like: CVDs, Diabetes, Carcinomas, Pulmonary diseases, etc. Health scientific and academic experts, including medical informatics experts need to be active in this field.
What do you think of Izet’s account of session 13? You can access the recording of session 13 in the video library. Send us your thoughts at email@example.com mentioning the “MIE2021 Daily May 30, Session 13” on the subject line.
Session 14 – Biomedical data, tools and methods, Chair: Patrick WEBER
By Patrick Weber, EFMI President (2013-2014)
This session was divided into engineer’s developments in the electric signals such as EEC EMG and terminologies on patient data. A better analyses and preparation of electric signals collections allow to better predict health related problems. The part of terminologies was highly discussed with questions. Finland minister representative is encouraging cross border data exchanges.
What do you think of Patrick’s account of session 14? You can access the recording of session 14 in the video library. Send us your thoughts at firstname.lastname@example.org mentioning the “MIE2021 Daily May 30, Session 14” on the subject line.
Session 16 – Human factors and citizen centred digital health – Social media, mobile apps, and patient portals, Chair Patrick Weber
By Patrick Weber, EFMI President (2013-2014)
This session was dedicated on the empowerment of the individual face health related problems. How to trust on available tools, how to understand the available data, how to help others in taking care of their own health situation.
What do you think of Patrick’s account of session 16? You can access the recording of session 16 in the video library. Send us your thoughts at email@example.com mentioning the “MIE2021 Daily May 30, Session 16” on the subject line.
Session 17 – Biomedical data, tools and methods – Artificial Intelligence, symbolic reasoning, machine learning, Chair Mihaela Marcella CRISAN-VIDA
By Mihael Marcella CRISAN-VIDA, co-chair EFMI HIIC WG
“Machine Learning Approach to Optimize Sedation Use in Endoscopic Procedures” was presented by Shorabuddin Syed. The objective of the paper presented was to build a Machine Learning (ML) model that predicts if a colonoscopy can be successfully completed with moderate sedation based on patients’ demographics, comorbidities, and prescribed medications. XGBoost model was trained and tested on 10,025 colonoscopies (70% – 30%) performed at University of Arkansas for Medical Sciences (UAMS). The proposed model can be employed as a decision support tool for physicians to bolster their confidence while choosing between moderate sedation and anesthesia for a colonoscopy procedure.
“Analyzing Association Rules for Graft Failure Following Deceased and Live Donor Kidney Transplantation” was presented by Syed Asil Ali Naqvi. This paper investigates the clinical attributes that contribute to kidney graft failure following live and deceased donor transplantation using an association rule mining approach. Analysis of a kidney transplantation dataset acquired from the Scientific Registry of Transplant Recipients that included over 95000 deceased and live donor recipients over 5-years was performed. The study presented in the paper identified that recipient peripheral vascular disease and prolonged cold ischemia time are often associated with graft loss. The novel finding presented of association rules comprising three or more donor-recipient predictors in combination may have direct clinical combination.
“On the efficiency of machine learning models in Malaria prediction” was presented by Ousseynou Mbaye. The paper presented proposes an extensive study of the efficiency of the most popular machine learning models for the task of Malaria occurrence prediction. The authors have considered patients from Senegal and have evaluated the overall accuracy of each considered algorithm based on sign and symptom information. The main result presented is that machine learning algorithms are promising, in particular Naive Bayesian presents a recall very close to that of a rapid diagnostic test while improving highly its precision by 9%
“Modeling of activity-induced changes in signal propagation speed of mechano-electrically stimulated neural fiber” was presented by Alina Troglio. The main purpose of the presented work was to test the combination of spike counting and predictive machine learning in modelling evoked speed changes of peripheral nerve fibers. The chosen approach showed promising results and additionally gave an important feedback for the further modelling process.
“Interpretable and Continuous Prediction of Acute Kidney Injury in the Intensive Care” was presented by Iacopo Vagliano. In this work, the authors proposed a machine learning model based on recurrent neural networks to continuously predict acute kidney injury. It was internally validated its predictive performance, both in terms of discrimination and calibration, and assessed its interpretability. This continuous model can support clinicians to promptly recognize and treat acute kidney injury patients and may improve their outcomes.
What do you think of Mihaela’s account of session 17? You can access the recording of session 17 in the video library. Send us your thoughts at firstname.lastname@example.org mentioning the “MIE2021 Daily May 30, Session 17” on the subject line.
Session 18 – Supporting care delivery – Health information systems and evaluation Chair: Carlos Luis PARRA CALDERÓN
By Carlos Luis PARRA CALDERÓN, EFMI Treasurer
The session proceeded normally. Mr. Sotirios Anagnostopoulos, presented the first study in place of Professor Mantas. The presented papers reported interesting research related to the evaluation of complex clinical decision support tools and medical orders, the co-creation of user requirements for chronic older people, as well as exciting work on the assessment of the success of a hospital information system. Additionally, students’ perception of 3D digital printing in healthcare, tools to support self-management of hemophilia patients, as well as tools catering to the information needs of caregivers of ALS patients, were presented.
What do you think of Carlos’s account of session 18? You can access the recording of session 18 in the video library. Send us your thoughts at email@example.com mentioning the “MIE2021 Daily May 30, Session 18” on the subject line.
Keynote: AI and Earlier Medicine, Jack Li, Chaired by John Mantas
By John Mantas, EFMI President (2011-2012)
Jack Li, in his keynote speech on AI, presented the evolution of AI from early days 1960’s till today. He divided it into the age of reasoning, age representation, age of machine learning and the next age of super-intelligence. He suggested that AI in medicine should be applied mostly to the ‘early and earlier’ stages in health care, that is AI in disease prediction, disease screening and disease monitoring, as well as in disease progression, disease prognosis and mortality, and disease prediction and prevention. He mentioned that the best doctors focus on keep people health whereas the mundane doctors treat the ill. So, the message to take home was that AI makes Medicine more Preemptive. One app he has developed and used and practice in this direction is the MoleMe App. He also mentioned the use phenotypes for adverse events detection by AI procedures. In conclusion Jack Li mentioned that the current medicine focuses on treatment which is very costly, whereas with the evolution of big data and analytics with AI applications the trend is moving to ‘Earlier Medicine’ for a sustainable healthcare system.
What do you think of John’s account of Jack Li’s Keynote? You can access the recording of the keynote in the video library. Send us your thoughts at firstname.lastname@example.org mentioning the “MIE2021 Daily May 30, Keynote Jack Li” on the subject line.
WORKSHOP: Collaboration and cooperation as factors for mutual professional expertise growth in health informatics, Chair Arriel Benis
By Arriel Benis, Executive Officer (2021-2023), chair HIIC WG
During this workshop, the core team of the WG HIIC has successively presented:
- the challenges of cooperation and collaboration in the fields of health and medical informatics;
- the current flagship project of the WG, the EFMI Medical Informatics multilingual Thesaurus;
- the history of the WG, its origins, past, current, and near-future activities such as the working group involved in the HosmartAI European project wherein EFMI takes a leading part.
The core team of the workshop comprises Arriel Benis, WG HIIC Chair, Israeli representative at the EFMI Council, EFMI Executive Officer (2021-2023), Mihaela Crisan-Vida, WG HIIC Co-Chair,
Stefan Darmoni, WG HIIC member, and Lăcrămioara Stoicu-Tivadar, WG HIIC member, WG HIIC founder, EFMI Past President (2018-2020).
In the second part of the workshop, there was a dynamic and fruitful discussion between the audience and the team of the WG HIIC. This exchange mainly focused on the potential extension and uses of the EFMI Medical Informatics multilingual Thesaurus.
What do you think of Arriel’s account of the HIIC workshop? You can access the recording of HIIC workshop in the video library. Send us your thoughts at email@example.com mentioning the “MIE2021 Daily May 30, HIIC workshop” on the subject line.
WORKSHOP: G-Lens to focus medication information – opportunities with IPS and structured ePI standard, Host Anne Moen
By Anne Moen, EFMI Honorary Fellow, EFMI President (2015-2016)
This afternoon the Gravitate-Health project shared ongoing work from strategy towards implementation and selected scenarios to support implementation of services for citizens to focus health information to improve access and understanding of the information. The forthcoming ePI proof of concept standard by EMA and the International Patient Summary will be important trusted sources of health information helping to focus information elements mindful of specific needs, concerns and goal that can be useful to meet specific contexts. Extensive use of European and national information resources and common interoperability standards, e.g., HL7 FHIR ®, will help make this important project feasible and sustainable.
What do you think of Anne’s account of the G-lens workshop? You can access the recording of G-lens workshop in the video library. Send us your thoughts at firstname.lastname@example.org mentioning the “MIE2021 Daily May 30, G-lens” on the subject line.
WORKSHOP: Interprofessional Health Informatics Education: A View from the Field, HOST: Ursula Huebner
By Ursula Huebner, IMIA Board Secretary Elect
This workshop was organised and presented by Ursula Hübner and Nicole Egbert from the University of Applied Sciences Osnabrück, Germany, Kaija Saranto and Ulla-Mari Kinnunen from the University of Eastern Finland, Kuopio, Finland and Pedro Marques from the University of Porto, Portugal. Building on existing health informatics educational recommendations (IMIA, TIGER etc.) this workshop presented results of the ERASMUS project eHealth4all@eu that focused on the verification of these recommendations. A verification performed through the lens of experts from the field, such as physicians, nurses, physiotherapists, managers and informaticians, is necessary to adjust the recommendations that were mainly developed in the academic arena. However, it is the healthcare workforce that is confronted with the digital tools and needs to be trained and upskilled regarding health informatics.
The developments are embedded in the rich tradition of EFMI’s and IMIA’s activities in the educational area which was presented by John Mantas from the University of Athens.
eHealth4all@eu, which aims to develop, implement and evaluate interprofessional courses for graduate students from all different health professions, utilizes an educational pipeline for the course development. This pipeline starts with recommendations rooted in the TIGER initiative (Technology Informatics Guiding Education Reform), updates these recommendations through a scoping review of relevant educational publications, verifies these findings in focus group interviews before defining the pedagogical methods, the learning platform and the syllabus.
This workshop concentrated on the finding from the focus group discussions that were conducted in the three partner countries, i.e. Finland (findings presented by Ulla-Mari Kinnunen), Germany (findings presented by Nicole Egbert) and Portugal (findings presented by Pedro Marques). The findings were presented and summarized by Kaija Saranto.
The findings showed that the initial plans of the consortium to develop, implement and evaluated courses in the fields of
- data protection and security,
- data analytics,
- innovation and entrepreneurship,
- leadership and governance and finally
- ethical and legal issues
met the requirements of the experts from the field. Some topics, e.g. data analytics, were highlighted only in one or two countries revealing some country-specific views.
There was an interesting discussion, which was moderated by Ursula Hübner und Kaija Saranto, about the need to include health informatics ethics in the curricula of health professionals due to the upcoming AI based solutions and the steadily increasing digital outreach to patients. Both fields raised ethical questions that go beyond the classic biomedical ethics discourse. Furthermore, the question was raised how to increase the medical and health informatics workload in the curricula of physicians and nurses that are already crowded.
As the workshop lasted only 60 minutes, 15 minutes for the discussion were not enough to share all opinions and viewpoints. The discussion also revealed that there is a great interest in educational topics which matches the current IMIA activities to update its educational recommendations in biomedical and health informatics.
What do you think of Ursula’s account of the eHealth4All@eu workshop? You can access the recording of eHealth4All@eu workshop in the video library. Send us your thoughts at email@example.com mentioning the “MIE2021 Daily May 30, eHealth4All@eu” on the subject line.
Session 19 – Supporting care delivery (May 30, Room A, 15:00) Chair: Anne MOEN
By Anne Moen, EFMI Honorary Fellow, EFMI President (2015-2016)
During the session we learned from 6 presentations focusing on different ways to Support Care Delivery. The support ranged from perspectives on digitally mediated consultations to participatory co-design and community of practice for health needs and health system literacy by refugees. The ASPIRE framework seen here helped engage and guide the participatory processes. Supporting Care Delivery also included presentation of required technology support or lack of appropriate tools according to nurses in German study, importance of cybersecurity and professionals’ attitudes, digital phenotyping to suggest eligibility for participation in clinical research studies and preparation of a virtual coach to support elderly to continue rehabilitation program after transfer to the home. The presentations in this session demonstrated in a nice way the many opportunities to support care delivery with health informatics tool.
What do you think of Anne’s account of session 19? You can access the recording of session 19 in the video library. Send us your thoughts at firstname.lastname@example.org mentioning the “MIE2021 Daily May 30, Session 19” on the subject line.
Session 20: Health and Prevention (May 30, Room B, 3:00 pm), Chair Mauro Giacomini
By Mauro Giacomini, Chair THI WG
In this session we had four preregistered presentations and one live presentation. Nathan Jeffreys (University of Surrey, UK) presented his research entitled “Using primary care data to report real-world pancreatic cancer survival and symptomatology.”. The authors extracted a pancreatic cancer cohort from the English primary care database of electronic health records, analyzing symptom and mortality data. This study provides clinicians with updated UK-based survival estimates and symptom prevalence data. Future analysis should clarify high risk ‘clusters’ of concurrent symptoms to aid earlier diagnosis and utilize primary and secondary care data linkage to improve information yield and granularity.
Artis Luguzis (University of Latvia) presented his research entitled “The impact of prescribed dose assumption in evaluation of adherence and persistence to medication in patients after acute myocardial infarction.” The authors evaluate adherence and persistence to medication in patients after myocardial infarction and show that corresponding estimates differ significantly, when using different dosage assumptions, namely, when using defined daily dose or tablet per day dosage regimens. Moreover, they demonstrate that observed differences between medications might be a result of inaccurate dosage assumptions. Moreover, they propose a comparison of distribution of days between dispensing events to that of days supplied as a relatively simple visual inspection to validate dosage assumptions.
Andrea Prunotto (University of Freiburg, Germany) presented his research entitled “Automatic Generation of German Translation Candidates for SNOMED CT Textual Descriptions.” The authors present an approach called MTP (multiple translation paths) aiming at assisting human translation in SNOMED CT localization projects based on free, web-based machine translation tools. For a chosen target language, MTP generates a scored output of translation candidates (TCs) for each input concept. This paper describes the basic idea of MTP, the distribution of its output TCs and discusses typical examples with German as target language. A first qualitative analysis was promising and supports their hypothesis that a majority voting applied to many translation candidates yields higher quality results than from one single engine and input language.
Patrick Essay (University of Arizona, Tucson, AZ, USA) presented his research entitled “Phenotyping COVID-19 Patients by Ventilation Therapy: Data Quality Challenges and Cohort Characterization.” In this study, the authors applied a rule-based phenotyping algorithm to classify COVID-19 patients requiring ventilatory support. They analyzed patient outcomes of the different phenotypes based on type and sequence of ventilation therapy. In addition to summary statistics for each phenotype, they highlight data quality challenges and importance of mapping to standard terminologies. This work illustrates potential impact of accurate phenotyping on patient-level and system-level outcomes including appropriate resource allocation under resource constrained circumstances.
Helen Chen (University of Waterloo, Canada) presented live her research entitled “Federated Deep Learning Architecture for Personalized Healthcare”. The authors used deep learning to advance personalized healthcare requires data about patients to be collected and aggregated from disparate sources that often span institutions and geographies. Researchers regularly come face-to-face with legitimate security and privacy policies that constrain access to these data. In this work, they present a vision for privacy-preserving federated neural network architectures that permit data to remain at a custodian’s institution while enabling the data to be discovered and used in neural network modeling. Using a diabetes dataset, the authors demonstrate that accuracy and processing efficiencies using federated deep learning architectures are equivalent to the models built on centralized datasets.
After these very interesting and technically optimal presentations the group had an interesting discussion enlightening specific aspects of the presented research.
What do you think of Mauro’s account of session 20? You can access the recording of session 20 in the video library. Send us your thoughts at email@example.com mentioning the “MIE2021 Daily May 30, Session 20” on the subject line.
Session 21 – Human factors and citizen centered digital health, Chair: Ivana OGNJANOVIĆ
By Ivana OGNJANOVIĆ, Chair yEFMI WG, yEFMI Officer EFMI Board
The session included eight papers addressing different aspects of human engagement and interaction with digital services, ranging from the roles of citizens and social media users, patients, to professionals and students as future health professionals and experts in the field.
Two presentations were related to Covid -19, one was focused on analysis of public online conversations by using WHO public health infodemic intelligence analysis methodology and has demonstrated the frequent resumption of information voids relevant for public health interventions and risk communication in an emergency response setting. The second paper was focused on tweets related to Covid -19 and use of fuzzy logic for the detection of symptoms from French Covid-19-related list. Reported results showed improved the detection by the addition of 42% more correct matches with an 81% precision.
Related to development and evaluation of digital health solutions we heard three presentations. One presentation made an overview of current e-health development in Burkina Faso, Africa, with focus on implementation of data warehouse for dangerous pathogen monitoring in a laboratories network. In another presentation focus was put on design of assistance services in child emergency which could help emergency physicians and make simplified and ensured communication between devices. One presentation was focused on the concept of patient-centredness from the perspective of bringing control to the patient and supporting decision making and care coordination for health, disease and social care issues.
In addition to already mentioned analysis of tweets related to Covid -19, there was additional presentation focused on online social media analysis from perspective of peer interactions and identification of behaviour switching patterns in e-cigarette-related transitions between electronic and combustible modes.
One presentation was related to potentials of machine learning approaches as promising tools to address disability advocacy data needs by using Wikibase for editing, integrating, storing structured disability related data.
Related to education of future health professionals, one presentation was focused on presenting evaluation procedure and achieved data, as a key instrument in quality assurance and control of health care related educational programs, continual efforts on their improvements and modernization.
What do you think of Ivana’s account of session 21? You can access the recording of session 21 in the video library. Send us your thoughts at firstname.lastname@example.org mentioning the “MIE2021 Daily May 30, Session 21” on the subject line.
Session 22 – Human factors and citizen centered digital health – Patient empowerment / engagement, Chair Jacob Hofdiijk
By Jacob Hofdijk, Honorary Fellow, EFMI President (2009-2010)
At the end of the second day, I was pleased to chair the Session on Human factors and citizen centered digital health – Patient empowerment / engagement (22). The session had 6 papers all dealing with the assessment of the use of digital tools in different health care settings. We started with a paper of Amalia Spataru from the Swiss Data Center reporting on the adherence of patients of an autoinjector device of the medication for children with growth disorders. The highly engaged patients had a more than double growth impact then the lower engaged patients. The regularity of the injection had the greatest impact. In the discussion it became clear that these findings were in line with the expectation of the providers, but it was proven in a real life situation within a population of around 18.000 children.
Then we went to Turkey were Mert Baskaya from the SRDC group from Ankara, presented a paper on the Interoperability Architecture of the ADLIFE Patient Empowerment Platform. The ADLIFE project is trying to provide an integrated care solution for the active management of patients with advanced chronic diseases. An interesting feature of this approach is the Patient Empowerment Platform (PEP, not Guardiola), which enables the patient and their informal carerers to manage the patient’s disease. The seamlessly shared care plan is made possible by the use of HL7 FHIR Repository for the exchange of data between the local EHR’s and the PEP. It might be seen as a breakthrough in killing the silo’s where many patients have been lost for many years. The care plan is also equipped with functions to gather PROM and PREM data. We look forward to the results of the 7 pilot sites with 75 hospitals and many primary care settings and around 900 patients. At the next MIE we hope to be surprised by the results. This is the recipe for killing the silos.
The next presentation by Helena Bornet came from Geneva and focused on the two-step process to offer COVID patients a Link with their families. The paper describes the very quick approach to support the patient-families with a Zoom based solution, not taking into account data protection and security. The result was overwhelming positive, but the approach limited the scalability of the solution. So, in the same agile and interactive way a second version was developed using Visio which met the data security requirements and was integrated in the patient app ecosystem.
The evaluation proved that the objective of the hospital to provide human centered care was met with a positive score of 94%. The anxiety of being alone in the (intensive)care far from home and alone was taken away by this way of communication. To be continued in the future.
From Finland came the presentation on research done in a remote session. Finland is an ideal long stretched country for testing this approach. Paula Valkonen did the research on two eHealth interventions the DigiIN and Home Dialysis support. The study explains the approach taken and the lessons learned in using Remote Interviews and Visual Timelines with Patients. For a successful remote approach careful planning is important, for the use of the timeline approach required the pretask approach were the patients had to prepare the interview by drawing the timeline, which required additional training. As the interview is also done without a video connection, trust between patient and researcher is key. The study reached high quality data and the effect of a good contact at the first “visit” proved to be a key success factor.
Meghan Bradway from Norway introduced the results of her study into the pros and pitfalls of a pragmatic mixed-methods approach. The Norwegian Institute for e-Health has taken the bridge to combine quantitively and qualitative approaches for the evaluation of modern rapid developing mobile health applications. They developed a pragmatic approach to comprehensively understand a real-world situation. The ehealth app supports diabetes patients to self-manage, which has been introduced for a group of HCP’s. The focus was both on the administrative part of the issues related with the study design and on quantitively part on the result of the use of the intervention. The focus of the study was on the evaluation of the study design using Precis-2. It distinguishes 5 levels from explanatory to pragmatic and reflective of the real-world situations. The combination of the study design and the results of the eHealth intervention in the real world proved to be more than pragmatic, a great lesson to learn from the impact of new tools to support patients.
The last paper was the first for Alexandra Zingg at an international conference and was focused on the Digital Technology Needs in Maternal Mental Health: A Qualitative Inquiry. The study focused on the needs of pregnant women during pregnancy. Data were collected in open codes and then translated in Arial Coding, to facilitate the analysis on concepts, beliefs and barriers. Most woman relied on the internet to find answers to the issues they experienced. Over two third experienced episodes of baby blues, nearly half of them spoke with their obstetrician. The study learned that there are no perfect pregnancies and there is a need to include mental issues as part of the management of the pregnancy. There is still a world to win to better support women with mental needs during pregnancy. The results have been discussed with the clinicians and the future will show how digital tools can help support women during pregnancy and I am sure we will see Alexandra again presenting the results.
After the presentations we had a lively discussion and concluded that it would be of interest to find if the concepts presented in the papers could be integrated.
What do you think of Jacob’s account of session 22? You can access the recording of session 22 in the video library. Send us your thoughts at email@example.com mentioning the “MIE2021 Daily May 30, Session 22” on the subject line.
Session 23- Biomedical data, tools and methods (Room C, 4:30pm) Chair: Mira HERCIGONJA-SZEKERES
By Ivana OGNJANOVIĆ, Chair yEFMI WG, yEFMI Officer EFMI Board
The session included thirteen papers addressing different aspects of utilizing biomedical data, development of tools and methods. Three presentations were related to Covid-19, one was focused on emergency digital learning in a multilingual context, with preliminary findings on reduced complexity of learning contents in the most candidate languages while retaining the maximum amount of information. Another presentation highlighted the potential for creation of core outcome sets (COS) which could provide crucial means to standardize, aggregate, share, and analyze diverse research results related to Covid-19 pandemic. One presentation was aimed on enhancing COVID-19 diagnosis and monitoring by automated segment areas of COVID-19 infection over CT scans by using deep convolutional neural networks.
We heard two presentations about development of different prediction models. One presentation was focused on using eXtreme Gradient Boosting for prediction of high-risk cases of oral squamous cell carcinoma (OSCC) with salivary autoantibodies, while another presentation introduced GAN-based imputation and prediction approach for time series data.
We also heard three presentations related to patients’ enrolment and stakeholders/clients’ feedbacks. One paper introduced i2b2 (Informatics for Integrating Biology and the Bedside) with an extension developed specifically to help support the patient enrolment phase, reporting promising preliminary results. Another presentation gives key results of two online surveys feedback from stakeholders of German health research regarding the Portal of Medical Data Models which has been developed since 2011. It is important to highlight interest to reuse tried-and-tested forms as expressed by a large majority of the stakeholders. One presentation was focused on pseudonymization and privacy-preserving record linkage by using a web application hosted by a TTP, which was piloted and tested.
The recognition problems were addressed by two presentations, while additional two presentations addressed automated classification approaches in different contexts. Classification of emotions was performed by using Short-Time Fourier Transformations of facial EMG signals, the peak frequency values extraction from the signal in intervals of one second and further classification of extracted features by support vector machine (SVM); which all resulted with an accuracy of 61,37%. Another presentation was focused on prediction of clinical relevance of pharmacist interventions (PIs) by identifying and labelling expressions from unstructured comments of PIs.
Automated detection of tuberculosis as abnormality detection in chest radiographs was analysed as classification problem over significant Speeded-Up Robust feature (SURF) and Scale Invariant Feature Transform (SIFT) features and resolved by SVM classifiers with precision greater than 82%.
One presentation addressed classification of Drug Sensitive (DS-TB) and Drug Resistant Tuberculosis (DR-TB) in chest X-rays, with proven statistical significance of mediastinal features and support vector machine approach with quadratic kernel which obtained classification performance values of greater than 93%.
What do you think of Ivana’s account of session 23? You can access the recording of session 23 in the video library. Send us your thoughts at firstname.lastname@example.org mentioning the “MIE2021 Daily May 30, Session 23” on the subject line.
Session 24 – Biomedical data, tools and methods, (May 30, Room C, 16:30) Chair: Kaija SARANTO
By Kaija Saranto, Chair EFMI Education WG
The session gave new insights through six interesting presentations. Image processing techniques were used to improve the quality of images in several patient cases e.g. to detect Alzheimer in early stage, to improve hearing loss by installing Cochlear implant and to localize vertebras in Adolescent Idiopathic Scoliosis.
Patient clustering has high potential to develop more in dept clinical pathways. However, the amount of data to create models for clustering is demanding. The quality of patient data is also challenging and need a lot of cleaning. In the analysis of fake news attached to Covid-19virus through Natural Language Processing the results made distinction of the role of titles and main text. Both seemed to have a meaning, but titles seem to be more unreliable.
What do you think of Kaija’s account of session 24? You can access the recording of session 24 in the video library. Send us your thoughts at email@example.com mentioning the “MIE2021 Daily May 30, Session 24” on the subject line.
Keynote: Digital phenotypes at home: New tools to engage patients and families, George Demiris, Chaired by John Mantas
By John Mantas, EFMI President (2011-2012)
The presentation of George Demiris starting with an introduction on the phenotypes in personalized medicine. The Advancement of current technology makes this possible on the condition that Ethics are closely followed according to regulations and law. The innovative approaches to Smart Home they can now be enhanced with the additional phenotype data and the AI applications. The system features on the smart home will not require retrofitting the home, it will work passively, replaceable technology, no cameras and face recognition, and will preserve privacy. The application will provide alerts when adverse events may have occurred. This is ideal environment avoiding patient falls for socially vulnerable older adults with mild cognitive impairment (www.sense4safety.org). One item of innovation is the capturing of gait changes and from that data one may extraction information on the patient’s condition and behaviour, activating possible alerts. The user may provide obtrusiveness feedback based on related framework. This environment is under further research that may have application and implication to the clinical workflow, with considerations to the users and to the consumers education.
What do you think of John’s account of George Demiris Keynote? You can access the recording of the Demiris keynote in the video library. Send us your thoughts at firstname.lastname@example.org mentioning the “MIE2021 Daily May 30, Keynote Demiris” on the subject line.
yEFMI community initiatives:
- All the Young Scientists in the field of Medical and Health Informatics across the Globe are invited to help EFMI to develop an innovative program to support the young scientists and professionals, by expressing their interests and expectations from yEFMI and completing the survey available at following link.
- All the National Societies in the field of Medical and Health Informatics are invited to contribute to yEFMI activities and help in aligning joint activities involving young researchers and professionals. The survey is open for interested representatives of the national societies and available at following link.
- All the EFMI Institutional members are invited to contribute in aligning joint activities involving young researchers and professionals, by expressing their interests and potentials for cooperation with yEFMI. The survey is available at following link
yEFMI page: https://efmi.org/workinggroups/yefmi-young-efmi/