Ebook: Healthcare of the Future 2025

This volume presents accepted papers from the third triennial Healthcare of the Future conference to be held since the series began in 2019. The theme of the inaugural conference [1] was ’Bridging the Information Gap’, and the conference showcased the innovative cross-institutional digital care pathways that connect patients at home with general practitioners, specialists, hospitals and rehabilitation centres [2]. It also envisioned a future where smart systems, wearable devices and telehealth services would enable more independent and empowered living at home. At the time, many of these technologies, applications and communication channels were in their early stages or existed only as proofs of concept.

This situation changed dramatically with the onset of the SARS-CoV-2 pandemic which began at the end of 2019, forcing the strained healthcare system to explore new and distance-based diagnostic and therapeutic measures to prevent overburdening healthcare systems. It also prompted the initiation of large data-collection applications, such as the COVID-19 Dashboard of Johns Hopkins University (https://gisanddata.maps.arcgis.com/apps/dashboards/bda7594740fd40299423467b48e9ecf6), which initially delivered faster and more accurate data than many national health authorities. In Switzerland, the way in which pandemic cases are reported and information is shared between physicians, laboratories and the government has significantly evolved since then, and new digital communication channels have been established. Digital teleconferencing tools have been introduced, not only in healthcare, but also in education and many other sectors, and have since become established tools for immediate and everyday use.

However, both the second and this third Healthcare of the Future conferences have recognised that telecommunication tools can only partially replace face-to-face contact and dialogue. The second edition of Healthcare of the Future in 2022 [3] was entitled ‘Digital health – from vision to best practices’. It covered key topics such as new approaches to interoperability, evaluation of IT solutions, better support for research in medicine and medical informatics, and applications for patients and healthcare professionals.

In this third edition of the conference, we acknowledge the emergence of new treatment pathways such as ‘hospital at home’, which is being promoted in Switzerland and other countries. This approach allows patients to either avoid inpatient admission or to be discharged earlier by providing intensive support and hospital-equivalent treatment in their home environment.

Several pilot projects have been established, e.g. in Arlesheim [4] or Zurich [4, 5], and the canton of Bern has established a Swiss Centre for Care@home at the Bern University of Applied Sciences [6] to support research and networking activities related to integrated home-based acute care.

In addition, artificial intelligence (AI) has made great advances with the advent of generative pre-trained transformers, which were made available to the public via platforms such as OpenAI’s ChatGPT 3.5 in 2022, quickly reaching 100 million active users each month [7] and creating a new hype in the field. Since then, numerous AI tools have followed, and a search for ‘AI in medicine’ in Pubmed on 17 March 2025 yielded 208,519 results, which represents a huge increase in the last few years. The historical development of AI in medicine can be seen in [8].

The keynotes of the 2025 conference also reflect these recent developments and the respective demands for IT:

∙ Revolutionizing Healthcare: Integrating AI for Enhanced Patient Care and Clinical Efficiency

by Maxim Topaz, Professor at Columbia University, New York, USA.

∙ Delivering Hospital at Home for Acute Medical Care – the Role of Digital Platforms

by Daniel Lasserson, Professor at University of Warwick, UK.

∙ From Ideas to Impact: How AI, Smartphones, and Wearables Are Revolutionizing Diabetes Self-Management

by Stavroula Mougiakakou, Professor at University of Bern, Switzerland.

The 2025 conference is made up of four sessions covering the topics:

∙ Next generation AI solutions in medicine

∙ Young researchers’ track

∙ Connected care – the key to a seamless patient journey

∙ AI and social media: benefits and harms.

We look forward to an interesting event and hope that you enjoy these proceedings.

Biel /Bienne May 19th 2025

The Organising Committee

References

[1] Bürkle T, Lehmann M, Denecke K, Sariyar M, Bignens S, Zetz E, Holm J (eds). Healthcare of the Future – Bridging the Information Gap. Studies in Health Technology and Informatics 259, 2019. ISBN 978-1-61499-960-7 (print) | 978-1-61499-961-4 (online), available under https://ebooks.iospress.nl/ volume/healthcare-of-the-future-bridging-the-information-gap-5-april-2019-biel-bienne-switzerland?_gl=1*177qpux*_up*MQ..*_ga*OTY2NjUyODI3LjE3NDIzOTc3MDM.*_ga_6N3Q014 1SM*MTc0MjM5NzcwMi4xLjEuMTc0MjM5ODI1NS4wLjAuMA..

[2] Bürkle T, Denecke K, Lehmann M, Zetz E, Holm J (eds). Integrated Care Processes Designed for the Future Healthcare System. Stud Health Technol Inform 245 (2017), 20-24.

[3] Bürkle T, Denecke K, Holm J, Sariyar M, Lehmann M. Healthcare of the Future - Digital Health – From Vision to Best Practice! Studies in Health Technology and Informatics 292, 2022, ISBN: 978-1-64368-280-8 (print) | 978-1-64368-281-5 (online), available under https://ebooks.iospress.nl/ISBN/978-1-64368-280-8

[4] Koechlin S. Das Spital kommt nach Hause – neue Versorgungsmodelle. Schweizerische Ärztezeitung | 2024;105(1–2):16–19.

[5] Meyer M, Bobst S. So können Spitäler über 3 Milliarden Franken einsparen. Tages-Anzeiger Sonntagszeitung 26.5.2024, available under https://www.tagesanzeiger.ch/hospital-at-home-so-sparen-spitaeler-ueber-3-milliarden-franken-147966993603 last visited Mach 17th, 2025.

[6] Senn C. In Bern entsteht ein Zentrum für die Pflege in den eigenen vier Wänden. Berner Zeitung 31.1.2025, available under https://www.bernerzeitung.ch/gesundheit-im-kanton-bern-in-bern-entsteht-ein-zentrum-fuer-die-pflege-in-den-eigenen-vier-waenden-520533984620 last visites March 17th, 2025.

[7] Hu K. ChatGPT sets record for fastest-growing user base - analyst note. Reuters 2.2.2023, available under https://www.reuters.com/technology/chatgpt-sets-record-fastest-growing-user-base-analyst-note-2023-02-01/ last visited March 17th, 2025

[8] Enslin S, Kaul V. Past, Present, and Future: A History Lesson in Artificial Intelligence. Gastrointest Endosc Clin N Am. 2025 Apr;35(2):265-278. doi: 10.1016/j.giec.2024.09.003.

Challenges such as time constraints, distractions and multi-tasking can compromise patient safety in the demanding environment of surgery. To mitigate these risks, checklists have emerged as simple yet effective tools for ensuring critical aspects of patient care, such as verifying patient identity and planned interventions. However, their consistent and accurate implementation in daily practice remains a challenge. This paper presents an intelligent assistant, VoiceCheck, designed to enhance patient safety during surgical procedures by guiding the use of checklists. Seamlessly integrated into the surgical workflow, VoiceCheck uses advanced speech recognition technology to ensure compliance with safety protocols. Combining speech-to-text and text-to-speech capabilities, the assistant facilitates interactive communication with users and accurately captures approved information. Future work will study user acceptance and usability. An open issue is linking the system to a hospital information for retrieving relevant patient data.

Artificial intelligence (AI) holds potential to support persons with serious mental illness, but evidence remains limited. To explore opportunities and challenges there, we conducted qualitative interviews with multiple Norwegian mental health stakeholders. Data was analyzed using thematic analysis and sentiment analysis. Twenty-two informants shared their opinions, with half expressing moderately negative sentiments. Findings suggest AI can optimize mental health service delivery and encourage a flexible approach to recovery. However, AI’s ability to provide emotional support may inadvertently worsen isolation, highlighting the need for a more holistic approach to human oversight and personal adaptation. Enhancing AI literacy and gathering more evidence from both persons and service providers is essential for future development.

The goal of segmentation in abdominal imaging for emergency medicine is to accurately identify and delineate organs, as well as to detect and localize pathological areas. This precision is critical for rapid, informed decision-making in acute care scenarios. Vision foundation models, such as Segment Anything Model (SAM), have demonstrated remarkable results on many different segmentation tasks, but they perform poorly on medical images because of the scarcity of medical datasets. They lack robust generalizability across diverse medical imaging modalities, and they need to be fine-tuned specifically for medical images, as these images considerably differ from natural images. This study aims to investigate the application of a foundation segmentation model to ultrasound (US) images of the abdomen. We employed SAMed to segment and classify all organs and free fluid present in each US image. A dataset comprising 286 US images, corresponding segmentation masks, and organ-level labels was collected from the Bern University Hospital Inselspital. Due to the relatively small size of our dataset, we pre-trained SAMed on a larger public US dataset to fine-tune it for US imaging. We then applied this fine-tuned SAMed on the Inselspital dataset to generate multi-class masks and assessed its performance against ground truth annotations using standard evaluation metrics. The results demonstrated that the fine-tuned SAMed can identify and classify multiple organs, though challenging cases, such as free fluid segmentation, reveal opportunities for improvement. Furthermore, transfer learning proved to be a reliable solution for managing small datasets, a key obstacle in the medical imaging realm.

Falls pose a substantial risk to elderly individuals, especially those over 65, often leading to severe consequences. This project investigates the potential of the tēmi robot for fall detection in care facilities and its integration into a simulated clinical workplace system. The prototype employs the YOLOv8 image recognition model to detect fallen individuals during patrols, transmitting incident data to a simulated clinical system via Fast Healthcare Interoperability Resources (FHIR). While initial tests delivered promising results, enhancements in image recognition accuracy are required for effective real-world deployment.

This study investigates the transition from analog to digital Huddle-Boards within the operational environment of a hospital. By analyzing the current analog system and implementing a digital prototype, the study aims to measure efficiency, usability, and overall system improvements. A combination of observational analyses, surveys, interviews, and usability studies were conducted to assess the potential benefits of digitization. The results highlight significant time savings, increased accessibility, and enhanced usability of the digital system compared to its analog counterpart.

Triage is used in emergency departments to ensure timely patient care according to urgency of treatment. However, triage accuracy and efficiency remain challenging due to time-constraints and high demand. This proof-of-concept study evaluates an AI-powered triage system that leverages speech recognition (STT) and large language models (LLMs) to process patient interactions in triage and to assign an Emergency Severity Index (ESI) triage level and a classification of the main presenting complaint according to the Canadian Emergency Department Information System (CEDIS). In Switzerland, different Swiss German dialects add to the complexity of the task. STT models achieved word error rates (WER) of 2.3% for High German and 17.66% for Swiss German. Despite the high WER, the AI’s classification accuracy reached 90–100% for ESI levels and CEDIS codes. These results highlight the potential of integrating AI into triage workflows, enhancing consistency and reducing the documentation burden for clinical staff. Future research should address multi-language adaptation and data security to ensure seamless implementation in real-world settings.

Background:

Despite the benefits, the practical implementation of shared decision-making (SDM) is challenged by the lack of decision aids for patients.

Objective:

This rapid review analyses how patient journey maps (PJMs) have highlighted the need for shared healthcare decisions between patients and professionals.

Methods:

The extension of the PRISMA methodology for scoping reviews was used as a guideline. The MEDLINE, APA PsycInfo, Embase, Emcare and Nursing Database digital databases were queried using a reduced selection of search expressions. Resulting articles from peer-reviewed journals since 2015 were analysed.

Results:

11 articles provide directions regarding the enhanced support of SDM throughout the patient journey. The utilisation of basic patient journey modelling along care trajectories has facilitated patient engagement with professionals, thereby enabling the disclosure of their needs and experiences. A generic model could be abstracted from each condition-specific PJMs, highlighting when decisions may impact patient experience.

Conclusions:

The development of a digital visual representation of a generic patient journey, from the onset of symptoms to the management of the condition, has the potential to serve as a valuable communication tool, assisting patients and healthcare professionals in better preparing, focusing and documenting SDM conversations.

In a cross country comparison, we try to identify factors which may influence the degree of interaction between inpatient and ambulatory patient care. For three Scandinavian countries, the United states and Switzerland, the IT-systems in hospitals and healthcare regions as well as electronic health records are described and characterized and the results contrasted with the way healthcare is delivered and financed. As a result, the existence of a national patient identifier, a reduction in the number of hospital information systems and a common database for healthcare professionals in inpatient and outpatient care are identified as positive contributors towards seamless care pathways. In comparison, the existence of an Electronic Health Record in the hands of the patient, or the existence of a tax paid healthcare system or the amount of healthcare expenditure do not necessarily contribute to this effect, since they can be observed also in countries with intermediate or improvable linkage between inpatient and outpatient sector. Seamless patient care has no directly visible correlation to life expectancy or preventable mortality.

Inter-institutional and inter-professional communication is based on the prompt and complete transmission of relevant health data, whereas conventional paper-based data transmission is often delayed and incomplete. One main objective of the CAEHR project is to optimize inter-sectoral information provision and to establish structured health data transfer from hospitals to rehabilitation facilities for TAVI patients. After a thorough requirement analysis conducted through structured interviews with medical experts from different fields, web portal instances were deployed at two university medical centres, providing data access to rehabilitation facilities involved in the patient care. Data transfer of medical and nursing data is extracted from the primary hospital information system and transformed to FHIR and openEHR format, respectively, via data mapping. After informed consent has been obtained, data is sent to the web portal. This portal is implemented on a Kubernetes system and hosted at university medical centres, allowing restricted external access. Questionnaires and assessment outcomes are sent back as follow-ups from rehabilitation facilities. The ongoing clinical study has included 142 patients so far. Despite different system architectures at the university medical centres, a unified concept with comparable data flows could be widely applied.

Introduction:

Rapid advances in Artificial Intelligence (AI), especially with large language models, present both opportunities and challenges in healthcare. This article analyzes real-world AI-related harms in healthcare.

Methods:

We selected four recent AI-related incidents from the AIAAIC Repository.

Results:

The incidents discussed include: Whisper’s harmful hallucinations; UNOS’s algorithm delaying transplants for black patients; the WHO’s S.A.R.A.H. chatbot providing inaccurate health information; and Character AI’s chatbot promoting disordered eating among teens.

Discussion and conclusion:

These incidents highlight diverse risks, from misinformation to safety concerns, involving both industry and institutional providers. The article emphasizes the need for systematic reporting of AI-related harms, concerns about security, privacy, and ethics, and calls for a centralized health-specific database to enhance patient safety and understanding.

Background and objective:

ADHD affects 5–8% of children worldwide. Social media shows potential for ADHD interventions. This scoping review aims to assess the available literature on social media interventions for ADHD and their reported outcomes.

Methods:

A scoping review was conducted across four databases (ERIC, PubMed, Education Source, PsycINFO) using ADHD and social media keywords. Grey literature was searched via Google Scholar, conferences, and ADHD organizations. Data extracted covered study design, intervention, participants, platforms, outcomes, and quality (QualSyst, MMAT).

Results:

Eight studies were included, seven with strong methodological quality. The studies involved 386 participants (ages 4–18), some with parents/caregivers. Designs varied (feasibility studies, RCTs, mixed methods). Most interventions targeted physical activity or caregiver support, showing feasibility and mixed effects on health behaviors and social skills. One study reported mild adverse effects.

Conclusion:

While studies are limited, social media shows potential as an ADHD intervention, highlighting benefits, risks, and the need for informed choices.

Mental health challenges among university students are increasing, but stigma and limited access to professional support hinder help seeking. This study explored opportunities, requirements, and risks associated with developing a chatbot-based mental health application tailored to Swiss university students. Data were collected through semi-structured interviews with student counselors, administrators, and representatives, as well as a requirements engineering workshop involving key stakeholders. The results showed that a chatbot could reduce stigma, improve accessibility and support vulnerable groups, provided it included easy access, evidence-based content and emergency responses. However, concerns regarding data security, harmful advice, and over-reliance on the chatbot must be acknowledged. These findings highlight the need for ethical safeguards, robust design, and a complementary role for the chatbot within existing support systems to address student mental health effectively.