The current situation of the COVID-19 pandemic is unknown to all of us and is associated with various kinds of stress, especially for health care workers: overloading of intensive care beds by COVID-19 patients, vacancies in entire wards and practices, uncertainty about the course of the pandemic, anxiety about one’s own health, stress caused by the difficult reconciliation of work and family and many more.
Different studies show that during epidemics such as SARS or MERS employees in the health professions are particularly exposed to stress and are also at high risk in terms of mental health. At the same time, many of them also have structural, social and personal resources from which they draw strength and support.
Our multidisciplinary working group at the University Hospital Erlangen, Bonn and Ulm has developed an online survey to determine the current pressures and resources during the Covid-19 pandemic in a broad sample and to be able to offer targeted help to the colleagues affected in and after the crisis as quickly as possible.
The objective recording of subjectively multidimensional pain is a problem that has not been sufficiently solved so far (https://www.youtube.com/watch?v=JuWHpMR2a9Y). Especially in clinical pain measurement, verbal methods (pain scales, questionnaires) and visual analogue scales are common, but these are not very reliable and valid for mentally impaired people. Expressive pain and/or psychobiological parameters may offer a solution. Such coding systems exist, but they are either very expensive or have not been sufficiently evaluated in test theory. Subjects are exposed to painful stimuli under controlled conditions and mimic as well as psychobiological parameters are recorded (http://www.iikt.ovgu.de/BioVid.html). Based on psychobiological, video-based (mimic, gesture) and prosodic data, the aim is to identify pain-relevant characteristics and develop an automatic system to measure pain qualitatively and quantitatively (www.jove.com/t/59057).
- Eberhard Barth, Department of Anaesthesiology and Intensive Care Medicine, University Hospital Ulm
- Ayoub Al- Hamadi & Philipp Werner, IIKT, University of Magdeburg
- Oliver Wilhelm and Mattis Geiger, Differential Psychology and Psychological Diagnostics, Ulm University
- Adriano Andrade, Biomedical Engineering Laboratory (BioLab), Federal University of Uberlandia, Brazil
- Friedhelm Schwenker, Institute of Neuroinformatics, University of Ulm
- Magrit- Ann Geibel, oral, maxillofacial and facial surgery, University Hospital Ulm
VDI/ VDE; BMBF
The potential of cognitive-technical systems is increasingly recognized and used in medicine. Cognitive-technical systems interact autonomously and support their users through intelligent perception, recognition and action. The research and development of cognitive medical systems, as helpful and supportive assistants and experts, aims at optimizing patient-specific diagnoses and individual therapies in different medical disciplines. Within this project, funded by the Margarete von Wrangell Habilitation Program, individual components for an automated intelligent medical system for non-invasive cardiovascular prevention are designed and developed. The relationship between emotional stress and cardiovascular risks is investigated. Cognitive-technical systems that automatically process stressful emotions, detect them early and intelligently interpret them are a valuable diagnostic support for risk analysis, especially in medicine, and aim to improve the quality of care for patients. Processing, detection and interpretation of emotional and stressful states is sensor-based and is based on the processing and analysis of biosignals and, depending on the context, on multimodal sensor fusion and data analysis. Methods of machine learning are applied.
Margarete von Wrangell Habilitation Program, MWK Baden-Wuerttemberg
Beate Ditzen, Johannes Ehrenthal and Julia Mahal, Institute of Medical Psychology, Heidelberg University Hospital
In the planned project, we want to investigate how doctors and patients react to different AI mediated communication formats in realistic eHealth scenarios in order to find out which formats can actually be used responsibly and for the benefit of patients in doctor-patient interaction.
We will design interaction sequences for doctor-patient communication that contain varying levels of AI (use of a Wizard of Oz design) and compare them with real social communication in doctor-patient interaction. The interaction sequences describe different diagnostic situations and treatment planning under standard conditions, but also under aggravating circumstances (e.g. severe diagnosis, fear reaction or expression of suicidal thoughts by the patient). The response of the practitioner and the patient is documented at several levels (self-report, standardized survey of psychobiological measurements and video coding) and the different sequences are compared.
Based on these findings, we plan to derive recommendations for the use of AI-based techniques in the healthcare sector that take into account the psychological aspects of patients in particularly sensitive situations.
Diseases of the musculoskeletal system are one of the main risks of occupational disease and often result in chronic pain, which significantly reduce the quality of life of those affected and lead to occupational disability. These illnesses accounted for 21.8 percent of sick leave and caused the most absenteeism with 326.9 days of incapacity for work per 100 insured persons. Because of this prevalence, these forms of illness also place a considerable burden on the health systems of solidarity in addition to the individual history of suffering. According to the Federal Statistical Office, the costs of diseases of the musculoskeletal system and connective tissue (M00-M99) in Germany amounted to around EUR 34.2 billion in 2015. The main causes include back disorders (M54), disc damage (M51), shoulder lesions (M75) and internal damage to the knee joint (M23). According to WHO, the most common cause of back pain is lack of exercise as well as persistent one-sided physical stress or improper stress.
In terms of the ecosystem thinking, diseases of the musculoskeletal system have an impact
- on the individual and his or her quality of life as well as his or her possibility for biographical self-realization.
- on the individual as an employee, their employment biography and the opportunity to identify and experience social appreciation from their employment.
- on the individual company and its planning security.
- on the health provider and its evaluation possibilities for the individual disease or therapy measure.
- on health insurance and its ability to assess the risk of the occurrence of the disease.
- on the solidarity system "health insurance" and its ability to distribute health resources fairly.
- on medical research and its ability to conduct prevention research and therapy evaluation using valid and anonymized large data sets.
All those involved in this ecosystem have a great interest in preventing the musculoskeletal system from becoming diseased. However, due to mutual lack of transparency and trust, there is no common approach to collective action in this complex socio-economic situation.
The collaborative project LOUISA addresses this problem by researching, developing and validating a learning model for multidimensional quantitative motion analysis. Innovations are:
- a quantitative movement score that is accepted as an intersubjective understandable and practicable measurand by all stakeholders in the ecosystem.
- the detection and prediction of risk factors by intuitive, human senses-appropriate 2D camera motion self-analysis and by the users self-determined extensible automatic muscle tone and pain detection.
- a defined adaptation process for future innovative sensor channels.
In line with the concept of prevention, LOUISA enables people to carry out medical prevention with regard to musculoskeletal diseases in an uncomplicated and low-threshold manner directly at the workplace (see § 20 SGB V Primary Prevention and Health Promotion). Through a combination of efficiency and transparency, LOUISA creates trust and acceptance among all stakeholders in the ecosystem. Within the scope of occupational health management, LOUISA supports companies, organizations and health insurance companies (see § 20b SGB V Occupational Health Promotion). Furthermore, LOUISA can be used in medicine or physiotherapy as a validated measurement method using pre- and post-test methods. The comparability of the measured movement score significantly supports the selection of suitable therapy measures or methods by the doctor or therapist. It is also possible to monitor the course of therapy with LOUISA.
Success of the project
The project is successful,
- if we can make a statement about the accuracy of the individual sensor channels in detecting and predicting risk factors.
- if we can make a statement about the effort and performance of the defined adaptation process for future innovative sensor channels.
- if the model can be applied safely, practicably and with confidence in the defined ecosystem (employees, employers, health providers, health insurers, reinsurers).
Possible ethical, legal and social implications have been researched and validated. There is nothing to prevent certification within the scope of occupational health promotion (BGF) according to § 20b SGB V and pay-off by health insurance companies.
- Frank Weber, AIMO GmbH
- Welf Löwe, Department of Computer Science and Media Technology, Linnaeus University, Sweden
- Wulf Loh, International Centre for Ethics in Science, University of Tübingen
The well-being of the individual is inextricably linked to his physical health. The aim of LOUISA is to motivate people at work to actively maintain, improve and safeguard their health and quality of life.
In the funding program „KMU Innovativ: Mensch-Technik-Interaktion“, the LOUISA joint project addresses the topic “Healthy life – prevention by technical accompaniment”, in particular “to motivate people by means of appealing visualization and comparability to exercise or generally healthy behavior.”
LOUISA as a learning model with a dynamic adaptation process is a technology for prevention. It combines human-technology interaction (smartphone with a mobile 2D camera scan, pain and muscle tone measurement, feedback and therapy suggestions) with machine learning and biosensor technology. Technology learns from and with humans without constantly monitoring them. LOUISA is easily accessible, user-friendly, motivated to use it intuitively, strengthens the patient’s autonomy and health responsibility and thus places the focus on people, their quality of life and their work biographical potential.
VDI/ VDE; BMBF