2. Use Case Methodology
3. Use Cases Descriptions and Expectations
3.1. Use Case 1: Coronary Artery Disease Diagnosis
3.2. Use Case 2: AI Based Automatic Arrhythmia Analysis
3.3. Use Case 3: Fetal State Assessment during Labour
3.4. Use Case 4: Diagnosis in Epidermolysis Bullosa, a Rare Genetic Disease
3.5. Use Case 5: AI Chronic Management and Decision Support Engine
3.6. Use Case 6: Chronic Resources Management Support Tool
3.7. Use Case 7: Adverse Events Identification and Prevention
3.8. Use Case 8: Monitoring of the Recovery Process
3.9. Use Case 9: Material Consumption Recognition and Prognosis
3.10. Use Case 10: Optimization of Human-Robot Teams in Hospital Logistics Operations
3.11. Use Case 11: Co-Development and Evaluation
4. Discussion: Benefits and Challenges for AI in Hospitals
Institutional Review Board Statement
Data Availability Statement
Conflicts of Interest
- Halawa, F.; Madathil, S.C.; Gittler, A.; Khasawneh, M.T. Advancing evidence-based healthcare facility design: A systematic literature review. Heal. Care Manag. Sci. 2020, 23, 453–480. [Google Scholar] [CrossRef] [PubMed]
- McKee, M.; Merkus, S.; Edwards, N.; Nolte, E. The Changing Role of the Hospital in European Health Systems; Cambridge University Press: Cambridge, England, 2020. [Google Scholar]
- Increase in the Share of the Population Aged 65 Years or Over Between 2009 and 2019. Available online: https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Population_structure_and_ageing (accessed on 15 December 2020).
- Michal, J.; Ecarnot, F. The shortage of skilled workers in Europe: Its impact on geriatric medicine. Eur. Geriatr. Med. 2020, 11, 345–347. [Google Scholar] [CrossRef]
- Moser, E.; Narayan, G. Improving breast cancer care coordination and symptom management by using AI driven predictive toolkits. Breast 2020, 50, 25–29. [Google Scholar] [CrossRef]
- Abràmoff, M.D.; Tobey, D.; Char, D.S. Lessons Learned About Autonomous AI: Finding a Safe, Efficacious, and Ethical Path Through the Development Process. Am. J. Ophthalmol. 2020, 214, 134–142. [Google Scholar] [CrossRef]
- Wood, D.A.; Mahmud, E.; Thourani, V.H.; Sathananthan, J.; Virani, A.; Poppas, A.; Harrington, R.A.; Dearani, J.A.; Swaminathan, M.; Russo, A.M.; et al. Safe Reintroduction of Cardiovascular Services During the COVID-19 Pandemic. J. Am. Coll. Cardiol. 2020, 75, 3177–3183. [Google Scholar] [CrossRef]
- McKinney, S.M.; Sieniek, M.; Godbole, V.; Godwin, J.; Antropova, N.; Ashrafian, H.; Back, T.; Chesus, M.; Corrado, G.S.; Darzi, A.; et al. International evaluation of an AI system for breast cancer screening. Nat. Cell Biol. 2020, 577, 89–94. [Google Scholar] [CrossRef]
- Arbabshirani, M.R.; Plis, S.; Sui, J.; Calhoun, V.D. Single subject prediction of brain disorders in neuroimaging: Promises and pitfalls. NeuroImage 2017, 145, 137–165. [Google Scholar] [CrossRef] [PubMed]
- Bzdok, D.; Meyer-Lindenberg, A. Machine Learning for Precision Psychiatry: Opportunities and Challenges. Biol. Psychiatry: Cogn. Neurosci. Neuroimaging 2018, 3, 223–230. [Google Scholar] [CrossRef] [PubMed]
- Lee, E.-J.; Kim, Y.-H.; Kim, N.; Kang, D.-W. Deep into the Brain: Artificial Intelligence in Stroke Imaging. J. Stroke 2017, 19, 277–285. [Google Scholar] [CrossRef] [PubMed]
- Awan, S.E.; Sohel, F.; Sanfilippo, F.M.; Bennamoun, M.; Dwivedi, G. Machine learning in heart failure: Ready for prime time. Curr. Opin. Cardiol. 2018, 33, 190–195. [Google Scholar] [CrossRef] [PubMed]
- Hampe, N.; Wolterink, J.M.; Van Velzen, S.G.M.; Leiner, T.; Išgum, I. Machine Learning for Assessment of Coronary Artery Disease in Cardiac CT: A Survey. Front. Cardiovasc. Med. 2019, 6, 172. [Google Scholar] [CrossRef] [PubMed]
- Esteva, A.; Kuprel, B.; Novoa, R.A.; Ko, J.; Swetter, S.M.; Blau, H.M.; Thrun, S. Dermatologist-level classification of skin cancer with deep neural networks. Nature 2017, 542, 115–118. [Google Scholar] [CrossRef] [PubMed]
- Gulshan, V.; Peng, L.; Coram, M.; Stumpe, M.C.; Wu, D.; Narayanaswamy, A.; Venugopalan, S.; Widner, K.; Madams, T.; Cuadros, J.; et al. Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs. JAMA 2016, 316, 2402–2410. [Google Scholar] [CrossRef]
- Mishima, H.; Suzuki, H.; Doi, M.; Miyazaki, M.; Watanabe, S.; Matsumoto, T.; Morifuji, K.; Moriuchi, H.; Yoshiura, K.-I.; Kondoh, T.; et al. Evaluation of Face2Gene using facial images of patients with congenital dysmorphic syndromes re-cruited in Japan. J. Hum. Genet. 2019, 64, 789–794. [Google Scholar] [CrossRef]
- Global Strategy on Human Resources for Health: Workforce 2030. Available online: https://apps.who.int/iris/bitstream/handle/10665/250368/9789241511131-eng.pdf?sequence=1 (accessed on 18 December 2020).
- Artificial Intelligence in Medical Technology: Delivering on the Promise of Better Healthcare in Europe. Available online: https://www.medtecheurope.org/wp-content/uploads/2019/11/MTE_Nov19_AI-in-MedTech-Delivering-on-the-Promise-of-Better-Healthcare-in-Europe.pdf (accessed on 17 December 2020).
- Xu, S.; Hu, C.; Min, D. Preparing for the AI Era Under the Digital Health Framework. In Proceedings of the 2019 ITU Kaleidoscope: ICT for Health: Networks, Standards and Innovation (ITU K), Atlanta, GA, USA, 4–6 December 2019; pp. 4–6. [Google Scholar] [CrossRef]
- Tortorella, G.L.; Fogliatto, F.S.; Mac Cawley Vergara, A.; Vassolo, R.; Sawhney, R. Healthcare 4.0: Trends, challenges and research directions. Prod. Plan. Control. 2019, 31, 1245–1260. [Google Scholar] [CrossRef]
- Haluza, D.; Jungwirth, D. ICT and the future of health care: Aspects of health promotion. Int. J. Med. Inform. 2015, 84, 48–57. [Google Scholar] [CrossRef]
- Zijm, H.; Klumpp, M. Future Logistics: What to Expect, How to Adapt. In Dynamics in Logistics. Lecture Notes in Logistics; Freitag, M., Kotzab, H., Pannek, J., Eds.; Springer: Cham, Germany, 2017; pp. 365–379. [Google Scholar]
- Klumpp, M. Automation and artificial intelligence in business logistics systems: Human reactions and collaboration re-quirements. Int. J. Logist. Res. Appl. 2018, 21, 224–242. [Google Scholar] [CrossRef]
- Giusti, R.; Manerba, D.; Bruno, G.; Tadei, R. Synchromodal logistics: An overview of critical success factors, enabling tech-nologies, and open research issues. Transp. Res. Part E Logist. Transp. Rev. 2019, 129, 92–110. [Google Scholar] [CrossRef]
- Cardoen, B.; Demeulemeester, E.; Beliën, J. Operating room planning and scheduling: A literature review. Eur. J. Oper. Res. 2010, 201, 921–932. [Google Scholar] [CrossRef]
- Tuwatananurak, J.P.; Zadeh, S.; Xu, X.; Vacanti, J.A.; Fulton, W.R.; Ehrenfeld, J.M.; Urman, R.D. Machine Learning Can Improve Estimation of Surgical Case Duration: A Pilot Study. J. Med Syst. 2019, 43, 44. [Google Scholar] [CrossRef] [PubMed]
- Li, F.; Gupta, D.; Potthoff, S. Improving operating room schedules. Heal. Care Manag. Sci. 2015, 19, 261–278. [Google Scholar] [CrossRef] [PubMed]
- Kohli, R.; Tan, S.S.-L. National University of Singapore Electronic Health Records: How Can IS Researchers Contribute to Transforming Healthcare? MIS Q. 2016, 40, 553–573. [Google Scholar] [CrossRef]
- Romanow, D.; Cho, S. Straub Editor’s Comments: Riding the Wave: Past Trends and Future Directions for Health IT Research. MIS Q. 2012, 36, 8. [Google Scholar] [CrossRef]
- Zhang, P.; Li, N. An assessment of human–computer interaction research in management information systems: Topics and methods. Comput. Hum. Behav. 2004, 20, 125–147. [Google Scholar] [CrossRef]
- Rzepka, C.; Berger, B. User Interaction with AI-enabled Systems: A Systematic Review of IS Research. In Proceedings of the International Conference on Information Systems, San Francisco, CA, USA, 13–16 December 2018. [Google Scholar]
- Preece, J.; Rogers, Y.; Sharp, H.; Benyon, D.; Holland, S.; Carey, T. Human-Computer Interaction; Addison-Wesley: Boston, MA, USA, 1994. [Google Scholar]
- Dix, A.; Finlay, J.; Abowd, G.; Beale, R. Human-Computer Interaction; Pearson/Prentice-Hall: Upper Saddle River, NJ, USA, 2003. [Google Scholar]
- Mattila, J.; Koikkalainen, J.; Virkki, A.; Simonsen, A.H.; Van Gils, M.; Waldemar, G.; Soininen, H.; Lötjönen, J.; Initiative, A.F.T.A.D.N. A Disease State Fingerprint for Evaluation of Alzheimer’s Disease. J. Alzheimer’s Dis. 2011, 27, 163–176. [Google Scholar] [CrossRef]
- Ayres-de-Campos, D.; Spong, C.Y.; Chandraharan, E. FIGO consensus guidelines on intrapartum fetal monitoring: Cardi-otocography. Int. J. Gynecol. Obstet. 2006, 131, 13–24. [Google Scholar] [CrossRef] [PubMed]
- American Academy of Pediatrics, Committee on Fetus and Newborn; American College of Obstetricians and Gynecologists, Committee on Obstetric Practice. The Apgar Score. Pediatrics 2006, 117, 1444–1447.
- Brasil, S.; Pascoal, C.; Francisco, R.; Ferreira, V.D.R.; Videira, P.A.; Valadão, A.G. Artificial Intelligence (AI) in Rare Diseases: Is the Future Brighter? Genes 2019, 10, 978. [Google Scholar] [CrossRef] [PubMed]
- Danial, C.; Adeduntan, R.; Gorell, E.; Lucky, A.W.; Paller, A.; Bruckner, A.; Pope, E.; Morel, K.D.; Levy, M.L.; Li, S.; et al. Prevalence and Characterization of Pruritus in Epidermolysis Bullosa. Pediatr. Dermatol. 2014, 32, 53–59. [Google Scholar] [CrossRef] [PubMed]
- Epstein, R.H.; Dexter, F. Economic analysis of linking operating room scheduling and hospital material management in-formation systems for just-in-time inventory control. Anesth. Analg. 2000, 91, 337–343. [Google Scholar] [PubMed]
- Reference Architecture Model. IOP Publishing Internationa Data Spaces. Available online: https://www.internationaldataspaces.org/wp-content/uploads/2019/03/IDS-Reference-Architecture-Model-3.0.pdf (accessed on 18 September 2020).
- Country & Technical Guidance—Coronavirus Disease. IOP Publishing WHO. Available online: https://www.who.int/emergencies/diseases/novel-coronavirus-2019/technical-guidance (accessed on 18 September 2020).
- OECD. World Health Organization Improving Healthcare Quality in Europe; OECD: Paris, France, 2019. [Google Scholar]
- Employment, Social Affairs & Inclusion. IOP Publishing Europa. Available online: https://ec.europa.eu/social/main.jsp?catId=1022&langId=en (accessed on 18 September 2020).
- EIT Health, McKinsey & Company (2020) Transforming Healthcare with AI: The Impact on the Workforce and Organisations. IOP Publishing EIT Health. Available online: https://eithealth.eu/wp-content/uploads/2020/03/EIT-Health-and-McKinsey_Transforming-Healthcare-with-AI.pdf (accessed on 15 December 2020).
- Reuter-Oppermann, M.; Kühl, N. Artificial Intelligence for Healthcare Logistics: An Overview and Research Agenda, In Ar-tificial Intelligence and Data Mining in Healthcare; Masmoudi, M., Jarboui, B., Siarry, P., Eds.; Springer Nature: Cham, Germany, 2021. [Google Scholar]
- Davenport, T.; Kalakota, R. The potential for artificial intelligence in healthcare. Futur. Heal. J. 2019, 6, 94–98. [Google Scholar] [CrossRef]
- Noorbakhsh-Sabet, N.; Zand, R.; Zhang, Y.; Abedi, V. Artificial Intelligence Transforms the Future of Health Care. Am. J. Med. 2019, 132, 795–801. [Google Scholar] [CrossRef] [PubMed]
- Wynants, L.; van Calster, B.; Bonten, M.M.; Cllins, G.S.; Riley, R.D.; Heinze, G.; Schuit, E.; Dahly, D.L.; Damen, J.A.A.; Debray, T.P.A.; et al. Prediction models for diagnosis and prognosis of COVID-19 infection: Systematic review and critical appraisal. BMJ 2020, 369, m1328. [Google Scholar] [CrossRef] [PubMed]
- Niessner, H.; Rauner, M.S.; Gutjahr, W.J. A dynamic simulation–optimization approach for managing mass casualty incidents. Oper. Res. Heal. Care 2018, 17, 82–100. [Google Scholar] [CrossRef]
- Klumpp, M.; Zijm, H. Logistics Innovation and Social Sustainability: How to Prevent an Artificial Divide in Human–Computer Interaction. J. Bus. Logist. 2019, 40, 265–278. [Google Scholar] [CrossRef]
- Bertsimas, D.; Orfanoudaki, A.; Weiner, R.B. Personalized treatment for coronary artery disease patients: A machine learning approach. Heal. Care Manag. Sci. 2020, 23, 482–506. [Google Scholar] [CrossRef] [PubMed]
|Hospital||Type/Inpatient per Year/Outpatient per Year/Beds 1||Main Health Area(s)||Specific Application Areas (AA)/Focus on (F)/Expected Output (EO)|
|University Hospital of Bern (Switzerland)||Public/6000/737,830/64 (2018) 2||The University Clinic for Obstetrics and Gynecology of Inselspital||AA: Diagnosis|
F: Fetal state assessment during labor
EO: AI-based decision support system for fetal state assessment during labor. The solution can assist obstetricians in accurately assessing the fetal state in clinical practice during labor.
|Kuopio University Hospital (Finland)||Public/99,000/517,000/590 (2019) 3||All branches||AA: Diagnosis|
F: Finding new diagnostic and treatment methods, for coronary artery disease.
EO: An AI-based decision support system for selecting those patients among suspected CAD who benefit from further imaging.
|Hospital of Bozen (Italy)||Public/25,064/737,830/697 (2018) 4||All branches||AA: Care|
F: Rheumatological diseases and diabetes
EO: An intelligent tool able to support the definition and scheduling of the different laboratory tests, medical examinations and hospitalization.
|La Fe University Hospital (Spain)||Public/45,062/148,702/1004 (2019) 5||Management of |
Chronicity (Integrated Care) and Active and Healthy Aging
|AA: Care |
F: Strategic initiatives on integrated care for patients with complex chronic and/or oncological conditions
EO: An intelligent tool able to improve the management of chronic patients and to characterize the use of resources throughout chronic patients’ healthcare, reducing the economic burden for hospitals.
|Federico II University of Naples (Italy)||Public/n.a./365,000/1000 (2019) 6||Arterial hypertension on the cardiovascular system||AA: Care|
F: Arterial hypertension with particular reference to ischemia heart disease
EO: Development of diagnostic and therapeutic methodologies in the field of cardiac rehabilitation; development of remote monitoring systems (telemedicine) for patients with high cardiovascular risk.
|Orton Ltd., The Private Unit Helsinki Univ.Hospital (Finland)||Private/2000/22,000/40 7||Orthopedics, neurosurgery, cancer treatment, pain medicine and rehabilitation||AA: Care|
F: Ethical, rehabilitation and preventive care
EO: Developing new tools for the treatment and rehabilitation of musculoskeletal disorders and other conditions.
|Odense University Hospital (Denmark)||Public/104,229/1,104,229/1038 (2019) 8||All branches||AA: Logistics |
F: Future of health care in mind, incorporating innovative clinical and logistical technologies.
EO: To serve as a test bed for new medical technology, including an extensive use of robotics and AI.
|Bayındır Hospital (Turkey)||Private/11,284/252,995/131 (2019) 9||All branches||AA: Logistics|
F: Materials management and scheduling
EO: Optimizing resource allocation and medical materials planning, reducing operational costs and patient waiting times
|University Hospital Essen|
|Public/50,000/195,000/1300 (2019) 10||Genetic medicine, immunology, oncology, |
cardiovascular medicine and transplants
|AA: Logistics//F: Care operations with materials management and supply//EO: Digitalized, patient- and employee-oriented organization. To minimize time spent for the nurses on documentation and administrative tasks to allow more time for direct patient care.|
|Heath Organization||Current Problems and Approach||Vision on Potential Application of AI||Expected Improvement—KPIs|
|University Hospital of Bern, Department of Obstetrics and Gynecology||Fetal assessment based on Cardiotocography (CTG) or electronic fetal monitoring (EFM) limitations.||Their vision is to develop a medical decision support system, which can assist obstetricians in accurately assessing the fetal state in clinical practice during labor.||Improvement of decision-making can improve fetal outcomes after delivery and avoid unnecessary medical interventions and their health implications for mother and fetus, as well as their economic implications. The KPIs of the AI application are:|
|Kuopio University Hospital||Currently, the diagnosis of coronary heart disease has changed towards the non-invasive imaging, which has led to increasing number of patients scheduled to CCTA. Interpretation of CCTA is affected by the image quality, experience of the doctor and by other issues, which can in terms lead to unnecessary repeated or additive diagnostic imaging.||The motivation is to develop an automatic AI-based analysis system for the coronary computed tomography angiography (CCTA): To enhance diagnostic accuracy of CCTA and to guide clinical decision making. Interpretation of CCTA will be systematically guided by the standard AI-based analysis system.||The patients need only one diagnostic method and the workflow of the interpretation of CCTA become more fluent. Relevant KPIs are: |
|Hospital of Bozen||Limitations on healthcare resources management and chronic care pathways definition||AI tools to support the definition and scheduling of the different laboratory tests, medical examinations and hospitalization which affect STHA patients, personnel, equipment and resources inside and outside the hospital and located in multiple areas of the geographical territory of its responsibility||Ease the management of healthcare resources with a particular focus on rheumatological diseases and diabetes as chronic diseases. Relevant KPIs are: |
Decrease waiting time to access to scheduled medical examinations and labor tests,
Average cost to provide the healthcare services to the chronic care population,
Quality of the medical treatment, e.g., percentage of re-hospitalized patients.
|La Fe University Hospital||Chronic diseases (CDs) represent the major cost of morbidity and mortality and lead to 86% of all deaths. In Europe, these account for more than 75% of the healthcare burden with a cost for the economy of €700 billion per year.||AI will help to: Improve the management of chronic conditions and multimorbidity in the face of aging population and its implication on public health; Contain the impact and global burden of chronic conditions, multimorbidity and frailty on individual quality of life and on healthcare systems; Strength the clinical management of complex chronic conditions and multimorbidity having a better understanding of the individual prognosis and disease evolution, and targeting personalized interventions.||Optimization of resources and the clinical flow of chronic patients at Hospital. Relevant KPIs are:|
Efficiency on the allocation and consumption of resources,
Right assignment of chronic patient to care pathway,
Decrease in turnaround time,
Selection of right pathway,
Avoidable episodes of care inadequate use.
|Federico II University of Naples||Today CVD is the leading cause of death in Europe; presently 47% of all deaths in Europe and 40% of all deaths in the European Union (EU) are attributable to CVD. This means that across Europe as a whole 4 million deaths per year currently occur due to CVD, of which 1.9 million are in the European Union||Use of AI may help clinicians in problem solving and patient’s management. AI process may be used to improve process of health care management with specific regards to resource allocation, patient management.||Rapid assessment of correct management strategy. Relevant KPIs are: |
Improvement of timeliness in critical event treatment,
Reduction of ambulatorial visits,
Forecasting of avoidable critical conditions.
|Odense University Hospital||Maintain high quality treatment for our patients in a demographic development scenario and increasing chronic conditions||Need to rely on AI and robots to ensure quality level and improve security in repetitive tasks, while alleviating staffing challenges.||Optimize handling of transports and logistics.|
Relevant KIPs are:
Improve timing for transportation of patients or samples.
Release of staffing resources to other tasks/areas. As well as an improved working environment for staff.
|Use Case||Objectives||AI Method||Data Available||Defined Outcomes||Contributions against Pandemic Situations|
(1) MDS for Coronary Artery Disease (CAD) diagnosis
|The aim of this study is to train a ML algorithm to distinguish patients with suspected CAD to those who benefit from further imaging studies and to those who don’t. In other words, to evaluate the risk of the patient to have prognostic CAD for customized clinical decision-making.||Disease State Index (DSI), which is a method to quantify the probability to belonging to a certain disease population, originally developed to support clinicians in diagnosing Alzheimer’s Disease . It is designed to be ‘disease-agnostic’, so that it can be used equally well for other diseases, provided that data are available.||For the study, a number of contemporary CCTA studies imaged in Kuopio University Hospital (KUH) as well as ECG, myocardial perfusion, invasive coronary angiography imaging and essential clinical data (age, sex and other demographic data, medical history, cardiovascular risk factors and medication) are gathered from existing clinical databases in KUH.||Algorithms and AI solutions for doctors supporting clinical decision making in CAD diagnosis.||Reduction of visits to the hospital, which increases the patient and personnel safety.|
(2) AI based automatic arrhythmia analysis
|In this use case, state-of-the-art artificial intelligence (AI) based arrhythmia analysis algorithms are developed and integrated into wearable sensors. |
Development of novel AI-based arrhythmia monitoring system aims to improve arrhythmia detection:
Enable longer non-invasive monitoring time.
|State-of-the-art AI based arrhythmia analysis algorithms are developed and utilized to atrial fibrillation (AF) screening in patients with transient ischemic attack (TIA) or cryptogenic stroke (CS) and detection of post-operative atrial fibrillation in cardiac surgery patients. Used methods: neural networks, deep learning, ML.||6000 24 h Holter recordings with arrhythmia annotations.|
Wearable sensor database: 700 patients (300 patients with AF episodes) with wearable sensors.
New: TIA/CS database is collected: 48h home monitoring of simultaneous wearable PPG and ECG-recordings from 100 TIA/CS patients.
|Developed AF-screening solution will enable long arrythmia monitoring time and increased rate of AF diagnosis. Wearable sensors offer improved patient usability and AI assisted arrythmia diagnosis requires minimal effort from healthcare professionals; AF diagnosis has important impact to patient itself, since anticoagulation may save the patient’s life (prevent cardioembolic stroke). Cost saving potential: one prevented stroke can save 120,000€ to society.||Reduction of visits to the hospital, which increases the patient safety.|
Possibility to assess arrythmia of corona patients remotely. Increases patient and personnel safety.
(3) Medical decision support system for fetal assessment during labor
|Improving fetal assessment with accurate prediction of fetal hypoxia and reduction of caesarean and instrumental delivery rates. Develop an AI-powered clinical decision support system.||Ensemble methods (e.g., stacking and blending) combining Explainable AI (aka XAI), neural networks (e.g., CNN and RNN), and gradient boosting techniques (e.g., XGBoost)||The maternity ward of the Department for Obstetrics and Gynecology in the University Hospital of Bern will provide a dataset of cardiotocographic (CTG) recordings. It includes physiological data such as maternal heart rate, fetal heart rate, contraction strength. The dataset is labelled by MDs.||The AI will focus on removing the existing great inter- and intra-observer variability while achieving at least the same average accuracy as medical doctors following the “Updated 2015 FIGO Intrapartum Fetal Monitoring Guidelines”. The integration of our AI-powered system should lead to a direct and positive impact on effectiveness and efficiency.||Assisting personnel in diagnosis with AI in a situation where there are not enough experienced personnel available due to the pandemic.|
(4) Diagnosis in Epidermolysis bullosa, a rare genetic disease
|To support disease prediction and diagnosis through the integration of extensive biological data (images, genomics, molecular) and epidemiological (immunological, clinical, demographic, lifestyles) to identify genomic lesions, proteins and immune-logical states associated (biomarkers).||ML algorithms will be trained to integrate phenotypic and clinical data to improve accurate prediction of progress of Epidermolysis bullosa. AI-based methods will also be used for disease comprehension and therapeutic target selection by unravelling the affected genetic and molecular players and pathways.||This use case will exploit data, competencies, and facilities of the Modena EB-Hub, the center for diagnosis, research, assistance and development of innovative therapies created in January 2020 at the General Hospital of Modena.||Definition of AI-based decision support systems to expedite diagnosis, correct misdiagnosis, diagnose previously undiagnosed, and stratify EB patients for advance therapeutic intervention through the integrative analysis of clinical phenotypes and patient health records, genetic information, molecular levels, biochemical fingerprints and patient images.||Assisting doctors’ in the diagnostic process during the pandemic, when the resources to be used for diagnosis is limited.|
Maintaining normal procedures of diagnosing other health problems during the pandemic.
(5) Chronic care pathway and resources characterization, simulation of demand and prognosis.
|AI techniques applied to analyze the pathways of chronic care patients providing simulation and prediction capacities about the demand of use of hospital services and resources||ML techniques (neuronal networks; LSTM; statistics predictions modeling; random forest; decision trees). AI adjustment to chronic care attention, prototype testing, application evaluation (KPI).||Historical clinical records for patients with chronic diseases. Data about care plans and use of hospital services and resources (pathways) made by this group of patients based on degree of frailty. Macro parameters from population (estimate demand/prognosis)||AI agent and tool for dimensioning demand of resources, including prognosis and simulation, both at individual and population level. Intelligent assistant for redefinition/optimization of care plans||Reduction of the transmission risks by being able to re-organize the pathways according to pandemic context.|
(6) Critical Conditions identification and prevention
|Identification and prevention of critical conditions: Analysis of vital signs, automatic recognition of symptoms (e.g., skin rash, mood change) and direct interaction with patients.||Machine Learning Techniques such as DNN, Reinforcement Learning, Natural Language Processing and Statistical Methods.Adjustment chronic care, prototype, evaluation (KPI).||Test of algorithms in hospital of Bozen with either live settings or retrospective data. Retrospective data as heart rate, respiration rate, oxygen saturation and blood pressure. Moreover, general data such as age, sex, weight, height and other diseases.||AI tool for critical conditions identification and prevention along the chronic care pathway||Control of patients with COVID-19 confined to their homes, before variations in their critical conditions.|
Increase in patient and family safety, especially in patients with COVID-19 who live alone.
(7) Intelligent resources management
|An intelligent algorithm is developed to efficiently manage the scheduling of hospital resources.||Evolutive, self-learning and auto-adaptive techniques focused on chronic care, prototype testing, validation through KPI.||Hospital models of processes for resource utilization. Information: processes, cost, service level, delivery time, resource utilization, medical personnel qualification.||Scheduling planning tool for optimal management of hospital care resources for patients with chronic diseases.||Reduction the transmission risks.|
Better planning of resources in compatibility with pandemic demand.
(8) Monitoring of the recovery process
|Remote determination of vital parameters such as heart rate and respiration rate for an improved recovery monitoring.||Methods in the Area of computer vision and ML i.e., CNN, BNN, adaptive optical flow, SVM etc.||Recordings from lab situations available; more data will be generated within the Fraunhofer InHaus-Centre, Test of algorithms in hospital of Bozen||Software for vital parameters. Transfer to hospital environment; continuous monitoring; fast obstacle identification; safe solution; contactless||Reduction the transmission risks in professionals by reducing contact with monitored admitted patients with COVID-19.|
(9) Material consumption recognition and prognosis
|Develop an automatic material documentation on the care wagon or the material store in the nursing ward based on computer vision. A material consumption prognoses is developed with the derived data.||ML (computer vision, CNN): Used materials are matched with patient cases and their diagnoses and treatments. Thus, it is known which and how many materials are needed by the individual patient cases.|
(10) Optimizing logistic operations
|Optimize the internal logistics operations of the hospital by considering both manual and automatic transport in a resource management and scheduling framework. Generate recommendations for how to improve manual and robotic logistics, based on gathered data.||Reinforcement learning (multi-agent motion and path planning)|
(11) Co-development and evaluation
|Integration of optimization of internal logistics operations and material consumption||Predictive analytics and cognitive automation|
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.
© 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).